<%BANNER%>

Bone Marrow Dosimetry via MicroCT Imaging and Stem Cell Spatial Mapping

Permanent Link: http://ufdc.ufl.edu/UFE0024755/00001

Material Information

Title: Bone Marrow Dosimetry via MicroCT Imaging and Stem Cell Spatial Mapping
Physical Description: 1 online resource (245 p.)
Language: english
Creator: Kielar, Kayla
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2009

Subjects

Subjects / Keywords: cd34, dosimetry, hematopoietic, lymphoma, marrow, radioimmunotherapy, skeletal, toxicity
Nuclear and Radiological Engineering -- Dissertations, Academic -- UF
Genre: Nuclear Engineering Sciences thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: In order to make predictions of radiation dose in patients undergoing targeted radionuclide therapy of cancer, an accurate model of skeletal tissues is necessary. Concerning these tissues, the dose-limiting factor in these therapies is the toxicity of the hematopoietically active bone marrow. In addition to acute effects, one must be concerned as well with long-term stochastic effects such as radiation-induced leukemia. Particular cells of interest for both toxicity and cancer risk are the hematopoietic stem cells (HSC), found within the active marrow regions of the skeleton. At present, cellular-level dosimetry models are complex, and thus we cannot model individual stem cells in an anatomic model of the patient. As a result, one reverts to looking at larger tissue regions where these cell populations may reside. To provide a more accurate marrow dose assessment, the skeletal dosimetry model must also be patient-specific. That is, it should be designed to match as closely as possible to the patient undergoing treatment. Absorbed dose estimates then can be tailored based on the skeletal size and trabecular microstructure of an individual for an accurate prediction of marrow toxicity. Thus, not only is it important to accurately model the target tissues of interest in a normal patient, it is important to do so for differing levels of marrow health. A skeletal dosimetry model for the adult female was provided for better predictions of marrow toxicity in patients undergoing radionuclide therapy. This work is the first fully established gender specific model for these applications, and supersedes previous models in scalability of the skeleton and radiation transport methods. Furthermore, the applicability of using bone marrow biopsies was deemed sufficient in prediction of bone marrow health, specifically for the hematopoietic stem cell population. The location and concentration of the HSC in bone marrow was found to follow a spatial gradient from the bone trabeculae in lymphoma patients. Interestingly, chemotherapy was not found to effect the HSC population in concentration or gradient. Together, this work will provide more realistic and accurate dosimetry in internal radiation therapy of cancer patients.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Kayla Kielar.
Thesis: Thesis (Ph.D.)--University of Florida, 2009.
Local: Adviser: Bolch, Wesley E.

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2009
System ID: UFE0024755:00001

Permanent Link: http://ufdc.ufl.edu/UFE0024755/00001

Material Information

Title: Bone Marrow Dosimetry via MicroCT Imaging and Stem Cell Spatial Mapping
Physical Description: 1 online resource (245 p.)
Language: english
Creator: Kielar, Kayla
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2009

Subjects

Subjects / Keywords: cd34, dosimetry, hematopoietic, lymphoma, marrow, radioimmunotherapy, skeletal, toxicity
Nuclear and Radiological Engineering -- Dissertations, Academic -- UF
Genre: Nuclear Engineering Sciences thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: In order to make predictions of radiation dose in patients undergoing targeted radionuclide therapy of cancer, an accurate model of skeletal tissues is necessary. Concerning these tissues, the dose-limiting factor in these therapies is the toxicity of the hematopoietically active bone marrow. In addition to acute effects, one must be concerned as well with long-term stochastic effects such as radiation-induced leukemia. Particular cells of interest for both toxicity and cancer risk are the hematopoietic stem cells (HSC), found within the active marrow regions of the skeleton. At present, cellular-level dosimetry models are complex, and thus we cannot model individual stem cells in an anatomic model of the patient. As a result, one reverts to looking at larger tissue regions where these cell populations may reside. To provide a more accurate marrow dose assessment, the skeletal dosimetry model must also be patient-specific. That is, it should be designed to match as closely as possible to the patient undergoing treatment. Absorbed dose estimates then can be tailored based on the skeletal size and trabecular microstructure of an individual for an accurate prediction of marrow toxicity. Thus, not only is it important to accurately model the target tissues of interest in a normal patient, it is important to do so for differing levels of marrow health. A skeletal dosimetry model for the adult female was provided for better predictions of marrow toxicity in patients undergoing radionuclide therapy. This work is the first fully established gender specific model for these applications, and supersedes previous models in scalability of the skeleton and radiation transport methods. Furthermore, the applicability of using bone marrow biopsies was deemed sufficient in prediction of bone marrow health, specifically for the hematopoietic stem cell population. The location and concentration of the HSC in bone marrow was found to follow a spatial gradient from the bone trabeculae in lymphoma patients. Interestingly, chemotherapy was not found to effect the HSC population in concentration or gradient. Together, this work will provide more realistic and accurate dosimetry in internal radiation therapy of cancer patients.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Kayla Kielar.
Thesis: Thesis (Ph.D.)--University of Florida, 2009.
Local: Adviser: Bolch, Wesley E.

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2009
System ID: UFE0024755:00001


This item has the following downloads:


Full Text

PAGE 1

1 BONE MARROW DOSIMETY VIA MICROCT IMAGING AND STEM CELL SPATIAL MAPPING By KAYLA N. KIELAR A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLOR IDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2009

PAGE 2

2 2009 Kayla N. Kielar

PAGE 3

3 ACKNOWLEDGMENTS I would like to express my gr atitude first and foremost to Dr. Wesley Bolch for his guidance throughout my eight-year acad emic career at the University of Florida. His leadership, respect, and concern for both his students and the publications of the Bone Imaging Dosimetry group is unprecedented. The entire Nuclear a nd Radiological Engineering department has become like family to me, and I appreciate the gu idance of all of the facu lty and staff. I would also like to recognize the Warringt on College of Business faculty a nd staff for their assistance in shaping me into a more professional being. Several colleagues contributed to this work a nd I would like to thank them for their time, specifically Deanna Pafundi, Di dier Rajon, Amish Shah, Chris Watchman, and Vince Bourke. Without your help this would not have been possi ble. I would also like to thank Dr. Edward Dugan and Dr. Sanford Berg for their wonderful recommendation letters and Dr. Amir Shahlaee, Dr. Parker Gibbs, and Dr. Bill Slayton for being part of my committee. Thank you to the United States Department of Energy for funding me under the Health Physics Fellowship, as well as the American Nuclear Society for several need based scholarships. Without this, I truly would not ha ve been able to achieve this dream, and I am forever grateful for your selflessness. I am especially indebted to the previous students in the Bone Imaging and Dosimetry group for setting high standards for our work and the current students for their efforts in making this work possible. Finally, I am much obliged to my family for their support and my friends for their patience.

PAGE 4

4 TABLE OF CONTENTS page ACKNOWLEDGMENTS...............................................................................................................3 LIST OF TABLES................................................................................................................. ..........7 LIST OF FIGURES................................................................................................................ .......10 ABSTRACT....................................................................................................................... ............16 CHAPTER 1 INTRODUCTION..................................................................................................................18 2 BACKGROUND....................................................................................................................26 Bone Structure and Physiology..............................................................................................26 Hematopoiesis and Target Cells.............................................................................................28 Lymphoma....................................................................................................................... .......30 Chemotherapy and Marrow Toxicity......................................................................................31 Radionuclide Therapies......................................................................................................... .32 Internal Dosimetry Calculations.............................................................................................34 Previous Methods Detailed.....................................................................................................35 Radiation Transport using Chord-based Methods...........................................................35 Radiation Transport using Voxel-based Methods...........................................................37 Stem Cell Distributions in Normal Marrow....................................................................38 Stem Cell Distribution in Diseased Marrow...................................................................40 3 REFERENCE SKELETAL DOSIMETRY MODEL FOR THE ADULT FEMALE............48 Introduction................................................................................................................... ..........48 Materials and Methods.......................................................................................................... .52 Female Cadaver Selection...............................................................................................52 In-vivo Macrostructural Image Database........................................................................52 Ex-vivo Macrostructural Image Database.......................................................................53 MicroCT Microstructural Image Database......................................................................54 Mass Calculations............................................................................................................56 Bone sites containing active marrow.......................................................................56 Bone sites not containing active marrow.................................................................58 Transport Modeling.........................................................................................................59 Absorbed Fractions..........................................................................................................60 Results........................................................................................................................ .............60 Discussion..................................................................................................................... ..........62 Spongiosa Volume and Percentages................................................................................62 Volume Fractions............................................................................................................63 Tissue Masses..................................................................................................................63

PAGE 5

5 Surface to Volume (S/V) Ratios......................................................................................65 Surface Areas.................................................................................................................. .65 Weighting Factors...........................................................................................................66 Specific Absorbed Fractions (SAF).................................................................................68 SAF for individual bone site s containing active marrow.........................................68 SAF for all bone sites containing active marrow.....................................................69 SAF for all bone sites not containing active marrow...............................................70 Skeletal Averaged Absorbed Fractions (AF)..................................................................71 Effect of Update in Shallow Active Marrow Size...........................................................73 Conclusion..................................................................................................................... .........74 4 COMPARISON OF HUMAN STEM CELL MEASUREMENTS IN BIOPSIES VERSUS AUTOPSY SPECIMENS.....................................................................................109 Introduction................................................................................................................... ........109 Materials and Methods.........................................................................................................111 Autopsy Collection and Preparation..............................................................................112 Digital Imaging and Processing.....................................................................................112 Measurement of Hematopoietic CD34+ Cells..............................................................113 Cellularity Measurement...............................................................................................114 Marrow Area Measurement...........................................................................................114 Results........................................................................................................................ ...........116 Discussion..................................................................................................................... ........116 Full field Autopsy Measurements.................................................................................116 Total Simulated Bi opsy Measurements.........................................................................117 Vertical Small Field Measurements..............................................................................117 Horizontal Small Fi eld Measurements..........................................................................117 Statistical Analysis........................................................................................................118 Anisotropy.....................................................................................................................119 Dose Implications..........................................................................................................120 Conclusion..................................................................................................................... .......121 5 EFFECT OF CHEMOTHERAPY ON TH E SPATIAL DISTRIBTUION OF STEM CELLS IN HUMAN BONE MARROW.............................................................................129 Introduction................................................................................................................... ........129 Materials and Methods.........................................................................................................129 Patient Selection and Slide Preparation.........................................................................132 Digital Imaging and Processing.....................................................................................132 Measurement of Hematopoietic CD34+ Cells..............................................................133 Cellularity Measurement...............................................................................................133 Marrow Area Measurement...........................................................................................134 Results........................................................................................................................ ...........134 Mean Cellular Concentration........................................................................................136 Statistical Analysis........................................................................................................137 Spatial Gradient.............................................................................................................137 Discussion..................................................................................................................... ........137

PAGE 6

6 Comparison to Normal Marrow....................................................................................138 Effect of Chemotherapy................................................................................................138 Timing Effect.................................................................................................................140 Effect on Cellularity......................................................................................................140 Dose Implications..........................................................................................................141 Conclusion..................................................................................................................... .......141 6 CONCLUSIONS AND FUTURE WORK...........................................................................150 Conclusions.................................................................................................................... .......150 Future Work.................................................................................................................... ......152 Improvements in Skeletal Database..............................................................................152 Reference S Values for UFRF.......................................................................................153 Improvements in Cellular Measurements......................................................................153 Scalability and Clinical Applications............................................................................154 APPENDIX A TABLES OF SPECIFIC ABSORBED FRACTION............................................................155 B GRAPHS OF SPECIFIC ABORBED FRACTION.............................................................168 C SPECIFIC ABSORBED FRACTIONS FOR ALL BONE SITES.......................................220 D SPECIFIC ABSORBED FRACTIONS FOR LONG BONES.............................................225 E SKELETAL AVERAGED AB SORBED FRACTIONS......................................................230 LIST OF REFERENCES.............................................................................................................237 BIOGRAPHICAL SKETCH.......................................................................................................245

PAGE 7

7 LIST OF TABLES Table page 3-1 Ex-vivo CT data for all skeletal sites.................................................................................75 3-2 MicroCT data for the skelet al sites with active marrow....................................................76 3-3 Tissue compositions (% by mass) and mass densities used in skeletal mass calculations................................................................................................................... .....77 3-4 Percentage of spongiosa of reference female compared to a database of 20 cadavers......77 3-5 Spongiosa volume and percentage of total for UF reference male and female.................78 3-6 Averaged marrow volume fractions, bone volume fractions and shallow marrow volume fractions by skeletal site for UFRF.......................................................................79 3-7 University of Florida Reference Female masses for the trabecular spongiosa, cortical bone, and mineral bone regions for all skeletal sites.........................................................80 3-8 University of Florida Reference Female active marrow masses.......................................81 3-9 Shallow marrow masses for UFRF....................................................................................82 3-10 S/V Ratios................................................................................................................ ..........83 3-11 Surface Areas for UFRF in both tr abecular and cortical bone regions..............................84 3-12 Weighting factors for marrow sources for each bone site for UFRF.................................85 3-13 Weighting factors for bone surface and volume sources for UFRF..................................86 4-1 Pertinent values for each patient in the simulated biopsy study......................................122 4-2 Specimen fractional weight in total simula ted biopsies or hori zontal and vertical biopsies....................................................................................................................... .....122 4-3 Cellular concentration for simulated biopsies versus autopsies in active and total marrow......................................................................................................................... ....123 4-4 Cellular concentration for horizontal vers us vertical biopsies in active and total marrow......................................................................................................................... ....123 5-1 Pertinent data for all 12 patients, in both pre and post-chemotherapy biopsies..............143 5-2 Mean cellular concentrati on and error for both normoce llular and diseased marrow.....144 A-1 UFRF absorbed fraction for all sources ir radiating the active marrow in the cranium...155

PAGE 8

8 A-2 UFRF absorbed fraction for all sources irradiating the shallow active marrow in the cranium........................................................................................................................ ....155 A-3 UFRF absorbed fraction for all sources ir radiating the active marrow in the clavicles..156 A-4 UFRF absorbed fraction for all sources irradiating the shallow active marrow in the clavicles...................................................................................................................... ......156 A-5 UFRF absorbed fraction for all sour ces irradiating the active marrow in the mandible....................................................................................................................... ....157 A-6 UFRF absorbed fraction for all sources irradiating the shallow active marrow in the mandible....................................................................................................................... ....157 A-7 UFRF absorbed fraction for all sources ir radiating the active marrow in the scapulae...158 A-8 UFRF absorbed fraction for all sources irradiating the shallow active marrow in the scapulae....................................................................................................................... .....158 A-9 UFRF absorbed fraction for all sources ir radiating the active marrow in the sternum....159 A-10 UFRF absorbed fraction for all sources irradiating the shallow active marrow in the sternum........................................................................................................................ .....159 A-11 UFRF absorbed fraction for all sources irradiating the active marrow in the ribs..........160 A-12 UFRF absorbed fraction for all sources irradiating the shallow active marrow in the ribs........................................................................................................................... .........160 A-13 UFRF absorbed fraction for all sources irradiating the active marrow in the cervical vertebrae...................................................................................................................... .....161 A-14 UFRF absorbed fraction for all sources irradiating the shallow active marrow in the cervical vertebrae.............................................................................................................161 A-15 UFRF absorbed fraction for all sources ir radiating the active marrow in the thoracic vertebrae...................................................................................................................... .....162 A-16 UFRF absorbed fraction for all sources irradiating the shallow active marrow in the thoracic vertebrae.............................................................................................................162 A-17 UFRF absorbed fraction for all sources irradiating the active marrow in the lumbar vertebrae...................................................................................................................... .....163 A-18 UFRF absorbed fraction for all sources irradiating the shallow active marrow in the lumbar vertebrae..............................................................................................................163 A-19 UFRF absorbed fraction for all sources ir radiating the active marrow in the sacrum.....164

PAGE 9

9 A-20 UFRF absorbed fraction for all sources irradiating the shallow active marrow in the sacrum......................................................................................................................... .....164 A-21 UFRF absorbed fraction for all sources irradiating the active marrow in the ossa coxae.......................................................................................................................... ......165 A-22 UFRF absorbed fraction for all sources irradiating the shallow active marrow in the ossa coxae..................................................................................................................... ...165 A-23 UFRF absorbed fraction for all sources irradiating the active marrow in the proximal humeri......................................................................................................................... .....166 A-24 UFRF absorbed fraction for all sources irradiating the shallow active marrow in the proximal humeri...............................................................................................................166 A-25 UFRF absorbed fraction for all sources irradiating the active marrow in the proximal femora......................................................................................................................... .....167 A-26 UFRF absorbed fraction for all sources irradiating the shallow active marrow in the proximal femora...............................................................................................................167 D-1 Specific absorbed fractions in the shafts of the leg bones...............................................225 D-2 Specific absorbed fractions in the shafts of the arm bones..............................................226 E-1 UFRF skeletal averaged absorbed fracti on (AF) for various s ources irradiating the trabecular active marrow..................................................................................................230 E-2 UFRF skeletal averaged absorbed fracti on (AF) for various s ources irradiating the shallow trabecular active marrow....................................................................................231 E-3 UFRF skeletal averaged specific abso rbed fraction (SAF) for various sources irradiating the trabecular active marrow..........................................................................231 E-4 UFRF skeletal specific averaged ab sorbed fraction (SAF) for various sources irradiating the shallow trabecular active marrow.............................................................232

PAGE 10

10 LIST OF FIGURES Figure page 1-1 Sites of active marrow (red) and in active marrow (yellow) in the adult...........................24 1-2 Hematopoietic stem cell pathway.....................................................................................25 1-3 Marrow cavities showing bone surface sources.................................................................25 2-1 Human compact and spongy bone showi ng location of osteons and trabeculae...............41 2-2 Bone remodeling and resoprtion are po ssible by osteoblasts and osteoclasts, respectively................................................................................................................... .....41 2-3 Osteoporotic versus normal bone.......................................................................................42 2-4 Bone Mineral Density (BMD) changes with age and gender............................................42 2-5 Stem cells may become either other stem cells (expansion) or further specialized cells (proliferation).......................................................................................................... ...43 2-6 Hematopoietic stem cell differentiation pathways.............................................................44 2-7 Lymphatic system and lymph circulation..........................................................................45 2-8 Chord lengths across bone trabeculae (black) and marrow cavities (white) at scanning angle ................................................................................................................46 2-9 Paired Image Radiat ion Transport (PIRT) ........................................................................47 2-10 Schema of future reference mode ls at the University of Florida.......................................47 3-1 Top in-vivo scan of UF Reference Woman.......................................................................87 3-2 In-vivo contoured image of the ribs...................................................................................88 3-3 In-vivo contoured image of the cranium............................................................................89 3-4 In-vivo contoured image of the cranium showing the separated lobes:.............................90 3-5 CT images showing the abnor mality of the right humerus. ..............................................90 3-6 Contoured ex-vivo image of the femur,.............................................................................91 3-7 Contoured ex-vivo image of the lumbar vertebrae............................................................91 3-8 Active marrow distribution as a function of age used as reference values in ICRP 70.....92

PAGE 11

11 3-9 Cellularity varies by bone si te and decreases with age......................................................93 3-10 Current ICRP 70 reference cellula rities based on bone site and age.................................93 3-11 Microstructure of the fr ontal lobe (40% marrow)..............................................................94 3-12 Microstructure of the oc cipital lobe (11% marrow)..........................................................94 3-13 Microstructure of the femoral neck (87% marrow)...........................................................94 3-14 Microstructures of the femoral head (71% marrow...........................................................95 3-15 Microstructure of the sternum (99% marrow)...................................................................95 3-16 Microstructure of the 3rd cervical vertebrae (76% marrow)..............................................95 3-17 Microstructure of the 5th lumbar vertebrae (91% marrow)................................................96 3-18 Specific absorbed fraction for the TAM irradiating the TAM for the lumbar vertebrae...................................................................................................................... .......97 3-19 Specific absorbed fraction for the TAM irradiating the TAM50 for the lumbar vertebrae...................................................................................................................... .......98 3-20 Specific absorbed fraction for all sour ces irradiating the TAM for the lumbar vertebrae...................................................................................................................... .......99 3-21 Specific absorbed fraction for all sources irradiating the TAM50 for the lumbar vertebrae...................................................................................................................... .....100 3-22 Specific absorbed fraction for the TAM irradiating the TAM for all bone sites.............101 3-23 Specific absorbed fraction for the TAM irradiating the TAM50 for all bone sites..........101 3-24 Specific absorbed fraction for the CIM irradiating the CIM50 in the shafts of all long bones.......................................................................................................................... ......102 3-25 Specific absorbed fraction for the CBSMC irradiating the CIM50 in the shafts of all long bones..................................................................................................................... ...103 3-26 Specific absorbed fraction for the CBV irradiating the CIM50 in the shafts of all long bones.......................................................................................................................... ......104 3-27 Skeletal averaged AF for marrow sour ces irradiating the TAM for both UFRF and Stabin and Siegel (2003)..................................................................................................105 3-28 Skeletal averaged AF for bone trabec ulae sources irradiating the TAM for both UFRF and Stabin and Siegel (2003)................................................................................106

PAGE 12

12 3-29 Skeletal averaged AF for ma rrow sources irradiating the TAM50 for both UFRF and Stabin and Siegel (2003)..................................................................................................107 3-30 Skeletal averaged AF for bone tr abeculae sources irradiating the TAM50 for both UFRF and Stabin and Siegel (2003)................................................................................108 4-1 Limitations of cell count in biopsies................................................................................124 4-2 The lumbar vertebrae was chosen due to planar symmetry.............................................124 4-3 Artificial biopsy sections.................................................................................................125 4-4 Independent cellularity measurements were made for each biopsy.................................125 4-5 Stained CD34+ cells in human bone marrow .................................................................125 4-6 Image processing to determine 50m areal contours from surfaces of bone trabeculae ............................................................................................................................... ...........126 4-7 Spatial gradient of HSCs in th e active marrow in simulated biopsies.............................127 4-8 Spatial gradient of HSCs in the active marrow for anisotropy........................................128 5-1 Co-stained biopsy.......................................................................................................... ...145 5-2 Binary digital images of total marrow and trabecula.......................................................146 5-3 CD34+ cell concentration as a func tion of depth into marrow cavities...........................147 5-4 Change in HSC concentration as a function of time between biopsies...........................148 B-1 Specific absorbed fraction for the TAM irradiating the TAM for the cranium...............168 B-2 Specific absorbed fraction for all sour ces irradiating the TAM for the cranium.............169 B-3 Specific absorbed fraction for the TAM irradiating the TAM50 for the cranium............170 B-4 Specific absorbed fraction for all sources irradiating the TAM50 for the cranium..........171 B-5 Specific absorbed fraction for the TAM irradiating the TAM for the mandible.............172 B-6 Specific absorbed fraction for all sour ces irradiating the TAM for the mandible...........173 B-7 Specific absorbed fraction for the TAM irradiating the TAM50 for the mandible...........174 B-8 Specific absorbed fraction for all sources irradiating the TAM50 for the mandible........175 B-9 Specific absorbed fraction for the TAM irradiating the TAM for the scapulae..............176

PAGE 13

13 B-10 Specific absorbed fraction for all sour ces irradiating the TAM for the scapulae............177 B-11 Specific absorbed fraction for the TAM irradiating the TAM50 for the scapulae............178 B-12 Specific absorbed fraction for all sources irradiating the TAM50 for the scapulae.........179 B-13 Specific absorbed fraction for the TAM irradiating the TAM for the clavicles..............180 B-14 Specific absorbed fraction for all sources irradiating the TAM for the clavicles............181 B-15 Specific absorbed fraction for the TAM irradiating the TAM50 for the clavicles...........182 B-16 Specific absorbed fraction for all sources irradiating the TAM50 for the clavicles.........183 B-17 Specific absorbed fraction for the TA M irradiating the TAM for the sternum...............184 B-18 Specific absorbed fraction for all sour ces irradiating the TAM for the sternum.............185 B-19 Specific absorbed fraction for the TAM irradiating the TAM50 for the sternum.............186 B-20 Specific absorbed fraction for all sources irradiating the TAM50 for the sternum..........187 B-21 Specific absorbed fraction for the TAM irradiating the TAM for the ribs......................188 B-22 Specific absorbed fraction for all sour ces irradiating the TAM for the ribs....................189 B-23 Specific absorbed fraction for the TAM irradiating the TAM50 for the ribs...................190 B-24 Specific absorbed fraction for all sources irradiating the TAM50 for the ribs.................191 B-25 Specific absorbed fraction for the TAM irradiating the TAM for the cervical vertebrae...................................................................................................................... .....192 B-26 Specific absorbed fraction for all sources irradiating the TAM for the cervical vertebrae...................................................................................................................... .....193 B-27 Specific absorbed fraction for the TAM irradiating the TAM50 for the cervical vertebrae...................................................................................................................... .....194 B-28 Specific absorbed fraction for all sources irradiating the TAM50 for the cervical vertebrae...................................................................................................................... .....195 B-29 Specific absorbed fraction for the TAM irradiating the TAM for the thoracic vertebrae...................................................................................................................... .....196 B-30 Specific absorbed fraction for all sources irradiating the TAM for the thoracic vertebrae...................................................................................................................... .....197

PAGE 14

14 B-31 Specific absorbed fraction for the TAM irradiating the TAM50 for the thoracic vertebrae...................................................................................................................... .....198 B-32 Specific absorbed fraction for all sources irradiating the TAM50 for the thoracic vertebrae...................................................................................................................... .....199 B-33 Specific absorbed fraction for the TAM irradiating the TAM for the lumbar vertebrae...................................................................................................................... .....200 B-34 Specific absorbed fraction for all sources irradiating the TAM for the lumbar vertebrae...................................................................................................................... .....201 B-35 Specific absorbed fraction for the TAM irradiating the TAM50 for the lumbar vertebrae...................................................................................................................... .....202 B-36 Specific absorbed fraction for all sources irradiating the TAM50 for the lumbar vertebrae...................................................................................................................... .....203 B-37 Specific absorbed fraction for the TAM irradiating the TAM for the sacrum.................204 B-38 Specific absorbed fraction for all sour ces irradiating the TAM for the sacrum..............205 B-39 Specific absorbed fraction for the TAM irradiating the TAM50 for the sacrum..............206 B-40 Specific absorbed fraction for all sources irradiating the TAM50 for the sacrum............207 B-41 Specific absorbed fraction for the TAM irradiating the TAM for the os coxae..............208 B-42 Specific absorbed fraction for all sources irradiating the TAM for the os coxae............209 B-43 Specific absorbed fraction for the TAM irradiating the TAM50 for the os coxae............210 B-44 Specific absorbed fraction for all sources irradiating the TAM50 for the os coxae.........211 B-45 Specific absorbed fraction for the TAM irradiating the TAM for the proximal humeri......................................................................................................................... .....212 B-46 Specific absorbed fraction for all sour ces irradiating the TAM for the proximal humeri......................................................................................................................... .....213 B-47 Specific absorbed fraction for the TAM irradiating the TAM50 for the proximal humeri......................................................................................................................... .....214 B-48 Specific absorbed fraction for all sources irradiating the TAM50 for the proximal humeri......................................................................................................................... .....215 B-49 Specific absorbed fraction for the TAM irradiating the TAM for the proximal femora......................................................................................................................... .....216

PAGE 15

15 B-50 Specific absorbed fraction for all sour ces irradiating the TAM for the proximal femora......................................................................................................................... .....217 B-51 Specific absorbed fraction for the TAM irradiating the TAM50 for the proximal femora......................................................................................................................... .....218 B-52 Specific absorbed fraction for all sources irradiating the TAM50 for the proximal femora......................................................................................................................... .....219 C-1 Specific absorbed fraction for the TAM irradiating the TAM for all bone sites.............220 C-2 Specific absorbed fraction for the TAM irradiating the TAM50 for all bone sites..........220 C-3 Specific absorbed fraction for the TBS irradiating the TAM for all bone sites...............221 C-4 Specific absorbed fraction for the TBS irradiating the TAM50 for all bone sites............221 C-5 Specific absorbed fraction for the TBV irradiating the TAM for all bone sites..............222 C-6 Specific absorbed fraction for the TBV irradiating the TAM50 for all bone sites...........222 C-7 Specific absorbed fraction for the CBV irradiating the TAM for all bone sites..............223 C-8 Specific absorbed fraction for the CBV irradiating the TAM50 for all bone sites...........223 C-9 Specific absorbed fraction for the TIM irradiating the TAM for all bone sites...............224 C-10 Specific absorbed fraction for the TIM irradiating the TAM50 for all bone sites............224 D-1 Specific absorbed fraction for the CIM irradiating the CIM50 in the shafts of all long bones.......................................................................................................................... ......227 D-2 Specific absorbed fraction for the CBSMC irradiating the CIM50 in the shafts of all long bones..................................................................................................................... ...228 D-3 Specific absorbed fraction for the CBV irradiating the CIM50 in the shafts of all long bones.......................................................................................................................... ......229 E-1 Skeletal averaged AF for marrow sour ces irradiating the TAM for both UFRF and Stabin and Siegel (2003)..................................................................................................233 E-2 Skeletal averaged AF for trabecular bone sources irradiating the TAM for both UFRF and Stabin and Siegel (2003)................................................................................234 E-3 Skeletal averaged AF for marrow sources irradiating the TAM50 for both UFRF and Stabin and Siegel (2003)..................................................................................................235 E-4 Skeletal averaged AF for trabecu lar bone sources irradiating the TAM50 for both UFRF and Stabin and Siegel (2003)................................................................................236

PAGE 16

16 Abstract of Dissertation Pres ented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy BONE MARROW DOSIMETRY VIA MICROCT IMAGING AND STEM CELL SPATIAL MAPPING By Kayla N. Kielar August 2009 Chair: Wesley E. Bolch Major: Nuclear Engineering Sciences In order to make predictions of radiation dose in patients undergoing targeted radionuclide therapy of cancer, an accurate model of skeletal tissues is ne cessary. Concerning these tissues, the dose-limiting factor in these therapies is th e toxicity of the hema topoietically active bone marrow. In addition to acute effects, one must be concerned as well with long-term stochastic effects such as radiation-induced leukemia. Part icular cells of interest for both toxicity and cancer risk are the hematopoietic stem cells (H SC), found within the active marrow regions of the skeleton. At present, cellular-level dosimet ry models are complex, and thus we cannot model individual stem cells in an anatomic model of the patient. As a result, one reverts to looking at larger tissue regions where thes e cell populations may reside. To provide a more accurate marrow dose assessment, the skeletal dosimetry model must also be patient-specific. That is, it should be designed to matc h as closely as possible to the patient undergoing treatment. Absorbed dose estimat es then can be tailored based on the skeletal size and trabecular microstructure of an individual for an accurate prediction of marrow toxicity. Thus, not only is it important to accurately m odel the target tissues of interest in a normal patient, it is important to do so fo r differing levels of marrow health.

PAGE 17

17 A skeletal dosimetry model for the adult fe male was provided for better predictions of marrow toxicity in patients undergoing radionucli de therapy. This work is the first fully established gender specific model for these app lications, and supersedes previous models in scalability of the skeleton and radiation transp ort methods. Furthermore, the applicability of using bone marrow biopsies was deemed suffi cient in prediction of bone marrow health, specifically for the hematopoietic stem cell popul ation. The location and concentration of the HSC in bone marrow was found to follow a spat ial gradient from th e bone trabeculae in lymphoma patients. Interestingly, chemotherapy was not found to effect the HSC population in concentration or gradient. Together, this work will provide more realistic and accurate dosimetry in internal radiati on therapy of cancer patients.

PAGE 18

18 CHAPTER 1 INTRODUCTION The importance of assessing radiation dose was noted within one month of Wilhelm Roentgens discovery of x-rays, when he repor ted radiation burns following x-ray exposure. Marie Curie also reported similar burns while working with radium, and as early as 1900, the need for precautions when working with radiation were well understood (Lombardi, 2006 ). Since then, dose assessment has become an important fact or for not only health concerns, but for use in accurately treating medical conditio ns. All types of radiation therapy with energies high enough to pass through the outer layer of bone will pr ovide some complication to the bone marrow, especially to those tissues that ar e hypersensitive to radiation (Lim et al. 1997a). In particular, internal radiation therapy require s special dosimetry concerns sin ce radiation is now completely encompassed within bone marrow itself. In order to predict the risk of bone cancer, an accurate assessment of absorbed dose to skeletal tissues is necessary. Concerning these ti ssues, the limiting factor in radiation protection is the toxicity of hemat opoietically active bone marrow ( Sgouros et al., 1993 ; Siegel et al., 1990b ). Bone marrow is made up of red (active) marrow including stem cells, progenitors, precursors, and mature blood cells, and yellow (in active) fatty marrow. Active marrow, fat, and trabecular bone comprise the spongiosa, which is encased by dense, cortical bone (Marieb, 2006). In adults, hematopoiesis occurs only within the bone marrow of the axial skeleton consisting of the skull, vertebra l column, sternum, ribs and the proximal ends of the femur and humerus, shown in Figure 1-1 (Vogler, 1988). All blood cells develop from pluripotentia l stem cells, which are stimulated by hematopoietic growth factors into committed pathways depicted in Figure 1-2 (DOH, 2006) It is these hematopoietic stem cells that are the radiosensitive cell population and the basis for

PAGE 19

19 acute toxicity and cancer risk. In particular, hematopoietic stem cells and osteoprogenitor cells are the primary cells of interest for the risk of inducing leukemia or bone cancer, respectively. The endosteum is defined as the connective ti ssue lining of all trabecu lar and cortical bone surfaces within bone, and is a radiosensitive or gan because it is the site of OPC production (Gossner et al ., 2000; Gossner, 2003). In all forms of ra diation therapy, it is important to know the concentration and loca tion of these cells, as well as their response to radiation in order to optimize cancer treatment with th e risk of inducing new cancers. To assess radiation dose, the Medical Inte rnal Radiation Dose (MIRD) Committee is responsible for providing fundamental quantities and standard models fo r radiation protection, risk assessment, and trea tment planning (Loevinger et al. 1991; Sgouros et al. 2009). MIRD methodologies of absorbed dose requires separa te knowledge of the cumu lated activity within the source and the radionuclide S value (absor bed dose per unit cumulated activity). Much research has been done to asses the cumulated ac tivity within the source, while little has been done to make patient-specific assessments of absorbed dose per unit cumulated activity (Juweid et al. 1995; Macey et al. 1995; Sgouros, 1993; Sgouros et al. 1996b; Siegel et al. 1990a). Furthermore, these dose calculati ons require known reference valu es such as the fraction of energy deposited in a target tissue and the mass of those target tissues. Th is inherently requires the proper modeling of source and target tissues, including the ra diosensitive cell populations within bone marrow. In internal radioimmunotherapy, radiolabeled antibodies guide the radioactive element to a specific site of interest. Since these radionuc lides localize in bone, th ey will cause a greater damage to endosteal tissues than would conventional radiation therapy (Gossner et al. 2000). Bone marrow is continuously in contact with ra diation while these partic les circulate through the

PAGE 20

20 patients blood, causing dose to the hematopoiet ic stem cell population. Furthermore, the endosteum and trabecular bone regions r eceive a high dose, possibly damaging the osteoprogenitor cell populat ion. This increases the risk fo r marrow toxicity, leukemia, and bone cancer in radioimmunotherapy patients (Lim et al. 1997b). The spatial dist ribution of the stem cells is now an important factor in yielding a dos e estimate that is predic tive of marrow toxicity, especially if the antibody is a bone surface seeker, shown in Figure 1-3. Current radioimmunotherapy procedures used fo r the treatment of bl ood cancers such as lymphoma may result in a large dos e to the bone marrow, especially if there is some amount of marrow involvement. Lymphoma is a cancer of the white blood cells, which circulate through the bloodstream and lymphatic system via grav ity and muscle compression. Lymphomas may arise in lymph nodes, bone marrow, or other organs such as the spleen, tonsils, appendix, thymus and the intestinal Peyers patch, and may continue to circulate or may lodge themselves within that organ. Like other cancer s, lymphomas divide continuousl y and do not undergo normal cell death. Consequently, nonfunctiona l white blood cells build up with in affected organs and crowd out other functioning normal cells that are meant to nurture the bone marrow environment. This provides the basis for immunosuppresion in marrow-involved lymphomas (Harris et al. 1994; Kuppers et al. 1999). Radionuclide therapies currently used to treat non-Hodgkins lymphoma include Tositumomab and Iodine 131 (Bexxar) and Ibritumo mab tiuxetan and Yttrium 90 (Zevalin). Both of these therapies rely on man-made antibodies designed to bind to the CD20 antigen on both malignant and nonmalignant mature B-lymphocytes while carrying a radiat ion component. Thus, this targeted therapy delivers radiation to a sp ecific cell for cell killing. Apoptosis is induced through gamma and beta radiation for Bexxar (Iodine 13 1) and beta radiation only with Zevalin

PAGE 21

21 (Yttrium 90). Many of these patients will al so undergo chemotherapy be fore, during, or after radiation treatment. The stem cell population may be disturbed to some degree by the radiation and chemotherapy and measurement of marrow dosi metry should be adjusted in these cases (Siegel et al. 1993). Accordingly, not only is it importan t to accurately model the target tissues of interest in a normal patient, it is important to do so for differing leve ls of marrow health. Previous analysis has shown that absorbed dose is an accurate measure of hematological toxicity, but only if the absorbed dose estimates is specific to the particular patient. At present, the vast majority of skeletal reference models (SRMs) used for these purposes are based on studies in the late 1960s and ear ly 1970s at the University of Leeds in which a novel optical scanning method was used to obtain linear chord-leng th distributions across several skeletal sites of a single 44-year male subject. These data form an essential component of the ICRPs SRM published in ICRP Publications 30, 70, and 89. Recent work has been done at the University of Florida Bone Imaging and Dosimetry Group to more accurately determine the fractiona l energy deposition in bone for alpha emitters and beta emitters as a function of bone site. Also, complete sets of skeletal macrostructural and microstructural data, in a format sufficient fo r radiation transport si mulations, have been established for an image-based skeletal refe rence model for the adult male at an age representative of cancer patient s undergoing radionuclide therapy (66year). Relevant tissues of the microstructure, such as th e bone trabeculae, bone endoste um, and marrow cavities were accounted for through image-based techniques. Skel etal tissue masses, which are important in determining radionuclide S values, have been re ported to update those summarized in ICRP Publications 70 and 89 (ICRP, 1995, 2002).

PAGE 22

22 At present, we cannot model the actual cell populations on the scale of the entire human skeleton, and so we revert to using the larger tissue regions in stead. Along with being patient specific, dose estimates should properly model the tissue of interest. Radiosensitive tissues, particularly the stem cells within the red bone marrow, should be accurately known and modeled. The concentration and distribution of stem and progenitor cells with in the hematopoietically active marrow thus needs to be character ized. This should be known for both normal and diseased states as the stem cell population can change drastically. A patient specific model is of no use if target or source tissues are poorly modeled. When incor porated into radiation transport techniques, an accurate, patient specific, absorbed fraction can be found and used in the S value calculation. It is currently assumed by the Intern ational Commission on Ra diological Protection (ICRP) that target cells of interest for bone marrow dosimetry are uniformly distributed throughout the marrow cavities, based on previous studies (Charlton et al. 1996). Absorbed dose is averaged through all regions of bone marrow accordingly. Also, the number of target cells is considered to be proportional to the vo lume of red bone marrow as proposed by the ICRP due to a lack of data. However, recent studies have shown that a spatial gradient in the mouse femur may exist (Cui et al. 1996; Frassoni et al. 1982; Lord and Hendry, 1972). Moreover, studies have also shown a simila r spatial gradient in hematopoiet ic stem and progenitor cells in the human iliac crest (Watchman et al. 2007). It is proposed that the he matopoietic stem and progenitor cells are preferentially located closer to the bone trabecular surfaces, and decrease in concentration further into the bone marrow caviti es. This could cause significant error in calculating the dose estimates using the current dose-response models fo r internal radionuclide therapy since the tissue surrogate (the entire red bone marrow) no longer matches the radiosensitive cells of interest.

PAGE 23

23 In order to provide an accurate dose asse ssment in radiation pr otection, the reference skeletal dosimetry model must be patient-specif ic. That is, it should be designed to match as closely as possible to the indi vidual for both radiation protecti on and medical treatment. The skeletal tissue doses are then to be used only to establish dose limits based upon an acceptable risk or radiation induced eff ects (stochastic or non-stochast ic). Also, the location of radiosensitive cells must be characterized, as we ll as their relationship in diseased marrow. Background information regarding these as pects is provided in chapter two. Previous work has been done to give mass es timated for the male model, however, work had not been done to assess differences based on ge nder. To date, there is no skeletal reference female model and the third chapter serves to pres ent a companion skeletal reference model to the adult male. This reference model will be fully scalable for both skeletal size (macrostructure) and marrow health (microstructure) due to os teoporotic concerns in the current female population. The fourth and fifth chapters serve to be tter understand the concen tration and spatial gradient of the hematopoietic stem cell in both normal and diseased patients. Furthermore, current methods rely on using patient biopsies as a surrogate for the entire patient marrow, and the usefulness of small samples will be discussed. Finally, the sixth chapter provides insight into improvements needed in patient specif ic models for internal dosimetry.

PAGE 24

24 Figure 1-1. Sites of active marrow (red) and in active marrow (yellow) in the adult (Vogler, 1988).

PAGE 25

25 Figure 1-2. Hematopoietic st em cell pathway (DOH, 2006). Figure 1-3. The marrow cavity encompasses both the orange (deep) and blue (shallow) marrow. Bone surface seekers (yellow) provide a higher dose to the radiosensitive stem cell population.

PAGE 26

26 CHAPTER 2 BACKGROUND Bone Structure and Physiology Bone is a living, growing organ made up of tw o types of tissue: compact (cortical) and spongy (Spongiosa) bone, which differ by density a nd cellular function. Th e entire skeletal system consists mainly of bones, but also include s joints, ligaments and cartilage which together account for 20% of total body mass. Cortical bone consists of a centra l osteonic (haversian) canal surrounded by concentric rings (lamell ae) of bone matrix, which make up the entire Haversian system. It is within the lamellae th at mature bone cells cal led osteocytes reside, although they travel throughout the skeletal syst em through blood vessels in the canaliculi or small channels between the bone matrix (Marie b, 2006). Cortical bone makes up 80% of total bone mass. Within the cortical bone and contained within the lining of the peri osteum, is the spongy, cancellous bone whose lattice structure resemble s that of a honeycomb. Thin bone structures called trabeculae are interwoven throughout the spongiosa, housing small, irregular cavities for bone marrow. Although the honeycomb structure may seem random, it is purposefully created based on the weight bearing requirements of the pa rticular bone site. Because it contains the bone marrow, the spongy bone is the site of hema topoiesis or development of blood cells from hematopoietic stem cells. The compact and spongy bone regions are shown in Figure 2-1. Both the lining of the trabeculae and the lining of the Haversian canals make up the endosteum. Along with the periosteum, this delica te tissue contains osteob lasts and osteoclasts, which are the cells responsible fo r creating and destroying bone, re spectively (Figure 2-2). This is the also the locati on of the osteoprogenitor cell population which mature into osteoblasts. Traditional definitions of the endosteum are pr oposed in ICRP Publication 26, and assume a

PAGE 27

27 layer of osteoprogenitor cells located 10 m from all bone surfaces (ICRP, 1977). Thus, all new bone production starts from the linin g of the cortical (periosteu m) and trabecular (endosteum) bone. It is well known that the trabecular b one varies by age (Atkinson, 1965; Snyder et al. 1993) and bone site (Eckerman et al. 1985; Patton, 2000). There have also been studies indicating that trabecular struct ure may vary with gender (Patt on, 2000; Mosekilde, 1989). Since trabecular structure varies, the location and amou nt of endosteal tissue does as well. These studies have led to revisions in the current ICRP definition of th e endosteum due to its relevance in bone cancer induction. It has been proposed that the cells of concern may reach beyond 10 m from both trabecular and cort ical bone, and as deep as 50 m into the marrow cavity. Bone is constantly remodeled in orde r to maintain constant levels of Ca2+ and PO4 3-, and as a result of mechanical stress the particular bon e site endures. At birt h, osteoclasts are highly dominated by osteoblasts, leading to more bone growth than resop tion. However, well into adult life, osteoclasts begin to outnumber osteoblas ts, leading to less bone production (Calvi et al. 2003; Zhang et al. 2003). Osteoporosis is the most comm on bone disease in the United States and is a disorder that causes bone resorption to become much greater than bone formation. Thus, marrow cavities seem to get larger as trab eculae have unfilled cavities (Figure 2-3). Women have a higher predisposition to oste oporosis, measured through bone mineral density (BMD). A low BMD corresponds to a low mineral content (grams of calcium and other bone minerals), which leads to less dense, weak er bones. Individuals with a t-score (bone mineral density compared to average 30 year ol d) less than 1.0 are considered to have normal bone health. A BMD of 1.0 to 2.5 would indicate an oteopenic individual, and a BMD of greater than 2.5 would show one with osteoporosis.

PAGE 28

28 Osteoporosis affects women, and its onset and progression depends on both menopausal status and skeletal site among other things. Studies have sh own that women have bone changes that vary depending on whether th at bone site was axial or appe ndicular. It seems that appendicular bone mass always decreased, whereas axial bones go through a five phase change of (1) buildup of BMD from 20-29 years old, (2) maintenance from 35 until 44 years old, (3) mild bone decrease from 40 until 49 years old, (4 ) rapid bone decrease at 50-54 years old, and (5) decelerated bone d ecrease after 55-59 (Yao et al. 2001). This five-phase change is shown in Figure 2-4. It was also found that BMD values are consis tently higher for men in appendicular bones, but the same for both sexes in axial bones until about 50 years old. After 50 years old, BMD for women decreases greatly. This rapid bone decrea se around 50 years old seems to be due to a decreased estrogen level postmenopausal and is also shown in Figure 2-4 (Yao et al. 2001). Clearly, bone mass and bone mass changes are different by gender, and thus reference individuals of both male and female ge nder need to be established. The mass of a skeletal site is strongly a ffected by Marrow Volume Fraction (MVF), or amount of marrow compared to the total skeletal mass. Osteoporosis changes the MVF, thus changing bone mass. Thus, for patient specific estimates of absorbed dose, marrow masses must be made as specific as possible and must include osteoporotic state. Cu rrently, work is being done at the University of Florida to relate the measured BMD in a bone site of a patient to an estimate of the MVF of the bone site. It is hoped to build a dataset of reference individuals with varying microstructure that is s calable to a particular patient. Hematopoiesis and Target Cells All types of blood cells are generated through he matopoiesis in a self-regulated system set by demand. When the need for a particular type of cell increases, cytokines are released that

PAGE 29

29 stimulate hematopoietic stem cells to generate new mature blood cells (Sugiyama et al. 2006). Stem cells are unspecialized cells that can differe ntiate (turn into other cells) or self-regenerate (produce more stem cells) as shown in Figure 25. Therefore, under normal proliferation, some daughter cells remain allowing for an abundant su pply of stem cells. Ot her daughter cells are myeloid or lymphoid progenitor cells, which comm it to becoming a specific type of blood cell through various differentiation pathways (Figure 2-6). The only cells not generated by demand are lymphocytes, which are continually generate d and destroyed. Therefore, lymphopoiesis is inefficient when compared to all ot her forms of hematopoiesis (DOH, 2001). Blood cells are divided into th ree categories: erythroid, my eloid, and lymphoid. Erythroid cells are the oxygen carrying red blo od cells, myeloid cells are varied cells such as granulocytes, megakaryocytes and macrophages, and lymphoid cells are T-cells and B-cells responsible for the bodys immune system. Prenatally, hematopoiesis occurs in the yolk sack, liver, spleen, and bone marrow, but only occurs in the bone marrow in normal adults. Stem cells are rare in adults (<1% of all blood cells); howev er, a single HSC is capable of regenerating the entire hematopoietic system (Zhang et al. 2003). The entire bone marrow microenvironment and hematopoiesis is shown in Figure 2-6. Because they are the most primitive and least differentiated cells, hematopoietic stem cells are the most radiosensitive (Lim et al. 1997a). Radiation sensitiv ity is characterized by DNA mutations (as opposed to cell death) that occur during the more vulnerable forms of the molecule seen during DNA replication in the cell cycle as i nduced by ionizing radiati on or other mutagens. The more differentiated a cell is, the less sensitive it is to radia tion because the cell is less likely to undergo mitosis during periods of potential radiation exposure. Currently it is assumed that HSCs are uniformly distributed throughout the ac tive marrow, which resides in the axial skeleton

PAGE 30

30 (skull, vertebral column, sternum, ribs, and the proximal ends of the femur and humerus) of adults (ICRP, 1977). HSCs are found in bone marrow through marker s present on their cell surface that are recognized by specific sets of antibodies. Methods to identify ce ll type and stage of differentiation were first proposed by the 1st International Workshop and Conference on Human Leukocyte Differentiation Antigens (HLDA) (H eddy and Swart, 2005). Through that body, the cluster of differentiation (CD) was established in 1982 and is the current pr otocol used to allow cells to be defined based on the molecules present on their surface. To date, CD34+/CD31(indicating expression of CD34 antigen and s uppression of CD31 antigen ) is the most well known method to immunohistoche mically stain CD34 antigen s ites present on hematopoietic stem and progenitor cells, as well as endothelia cells, while excluding CD31 antigen sites present only on the vascular endothelium (Foucar, 2001; Rafii et al. 1994; Kuznetsov et al. 2001). Lymphoma Lymphoma is a disease of the bone marrow, whereby non-functional lymphocytes (a type of white blood cell), originally meant to prot ect from illness, build up and crowd out other nurturing cells. Many lymphomas ar ise within the lymph node, but some may originate in the spleen, thymus, Peyers patches of the intestine, or from within the bone marrow itself. These cancerous cells then circulate in the blood and lymphatic system or reside in those organs, as shown in Figure 2-7. It is beli eved that these specific leukocytes undergo a malignant change sometime during cell division or maturation during circulation. Because of this, the separation between lymphoma and leukemia is becoming an out moded idea, as both diseases are different manifestations of the same malignant cell as reflected in the Revised European American Lymphoma (REAL) classifi cation system (Swerdlow et al. 2008; Armitage, 2005).

PAGE 31

31 There are two types of lymphomas, differentiated by the presence of Reed-Sternberg cells found only in Hodgkins lymphoma. All other lymphomas are termed non-Hodgkins lymphoma (NHL) and are separated into either B-cell or T-cell lymphoma, dependent upon whether the Blymphocytes or T lymphocytes, respectively are in fected. These cells were named via the organ responsible for their maturation, which is the bone marrow for B-cells and the thymus for Tcells. Lymphomas are then classified using the World Health Orga nization (WHO) system, which includes categories based on the appear ance and chromosomal features of the cells (Swerdlow et al. 2008). B-cell lymphomas make up 85% of NHL in the United States and are commonly Diffuse Large B-cell lymphoma, Follicu lar Lymphoma or Mantle Cell Lymphoma. Bone marrow involvement is assessed by the amount of malignant lymphocytes found in the bone marrow and is usually higher for leukemias than lymphomas (Muller et al. 2005). Chemotherapy and Marrow Toxicity Chemotherapy is often used in conjunction wi th radiation therapy to provide systemic treatment of cancer cells (Forer o and Lobuglio, 2003). Chemothera py targets cells that divide rapidly, however, normal cells may divide just as fast as cancer cells, leading to less healthy tissue sparing. A common from of chemothe rapy for non-Hodgkins lymphoma is CHOP-R, a mixture of the drugs Cyclophosphamide, Doxorubi cin, Vincristine, Pred nisone, and an anti CD20 monoclonal antibody Rituxi mab. While the drugs help disrupt abnormal growth by causing apoptosis, Rituximab targets B cell ly mphocytes, through the CD20 antigen which is present on their cell surface. However, thes e proteins are also found on both normal and malignant cells, leading to ill side effect s resulting from healthy tissue damage. In all therapeutic regimes for hematological malignancies, some amount of depression in the immune system is guaranteed. Myelosupre ssion is a well-known side effect of cancer treatment, whereby the bone marrow activity is seve rely decreased. This re sults in a reduction in

PAGE 32

32 the number of red and white blood cells, as well as platelets. As radiat ion or chemicals produce lesions in DNA such as single/ double strand breaks or base alterations, mutagenesis or cell killing can occur. Strictly speak ing, cell killing is preferred to mu tagenesis, as these aberrations cannot be passed on to progeny. In radioimmunotherapy, as the antibody travel s to the tumor, it will come into contact with all marrow and cause radiat ion damage to normal cells. Th ere is also some amount of radiation that will cause hematopoietic toxi city in a specific patient. High marrow involvement, that is, marrow that has a large amount of abnormal cells within, greatly affects radiation dosimetry levels (Martinez-Jaramillo et al. 2004; Madrigal-Velazquez et al. 2006). As the antibody travels to the tumor, a higher concentration will stay in the marrow and the crossfire from cancerous cells to normal cells will be high. Treatments are performed in cycles to allo w for increases cell sensitivity with reoxygenation, and also to all for recovery from the side effects of marrow toxicity including fatigue and infection. In severe cases, a bone marrow cell transpla nt may be required to restore the depleted blood cell population. Some diseases that affect the bone marrow itself can also result in myelosupression. Bone marrow failure s in hematological diseases including leukemia and lymphoma may be the result of inherent de fects of the hematopoie tic stem and progenitor cells or due to damage caused by ionizing radiation, viruses or ch emical toxins. The increasing importance of HSCs in marrow health during th erapies has brought attention to the stem cell niche, specifically the spa tial codependence between stem cells and marrow vasculature (Taichman, 2005). Radionuclide Therapies Radiation may be delivered either externally or internally to treat malignant tumors or bone marrow cancers. DNA damage is induced by directly or indirectly ionizing i ndividual atoms. In

PAGE 33

33 external beam radiotherapy, the bone marrow as a whole is at risk and thus it is important to consider the entire tissue as a radiosensitive targ et. Radiation beams are shaped to conform as best as possible to the tumor in order to keep radiation dose extremely low in the healthy tissue. In internal radiotherapy, or radioimmunothera py (RIT) radiolabeled antibodies guide the radioactive element to a specific s ite of interest. If the antibody is a bone surface seeker, such as radiopharmaceuticals used for bone pain palliatio n, the distribution of the stem cells is now important in estimating the radi ation toxicity level (Kvinnsland et al. 2001). For treatment of non-Hodgkins lymphoma, both Bexxar and Zevalin use beta sources to deliver radiation to abnormal cells throughout the bone marrow and tumor site. Monoclonal antibodies intended to bind to the CD20 antigen on lymphocytes are conjugated to radioisotopes and injected into the bloodstream Tumor cells are then killed e ither by the antibody itself or by the crossfire from the radiati on. In Bexxar, Tositumomab is tagged to Iodine 131, which emits gamma and beta radiation at approximately 0.6 Me V. Zevalin uses Ibritumomab Tiuxetan with Yttrium 90 which only emits beta radiation at a higher (maximum) energy of 2.3 MeV. Beta particle emission should eliminate tumor cells within 1000-5000 m from their deposition. Gamma particles, if present, allow for more accurate dosimetry through monitoring radiation uptake in targeted sites. Since February 2002, Zevalin has been approve d by the FDA for treatment of relapsed, low grade, follicular B-cell NHL that is refracto ry to Rituximab. Similarly, Bexxar has been approved by the FDA for the treatment of CD20 positive, follicular B-cell, non-Hodgkins lymphoma (NHL) patients whose disease in unman ageable by Rituximab and who have relapsed following chemotherapy since June 2003. Ho wever, Zevalin has been approved for use concurrently with therapeutic regimes involving Rituximab.

PAGE 34

34 There has been much debate over whether these RIT procedures should be used in conjunction with chemotherapy, as well as their use as frontline treatment (Wiseman et al. 2001). Compared to chemothera py, RIT has the ability to indu ce apoptosis through radiation damage, as well as drugs aimed at targeting the CD20 antigen present on lymphocytes. This may provide a less toxic marrow environment than the untargeted concoction administered in chemotherapy. Clinical trials have shown good pr ognosis in the use of RI T as primary treatment (Kaminski et al. 2005). The low use of Bexxar and Zevalin may be due to lack of awareness, insurance coverage, and unintende d conflicts of interest among th e range of therapies in local practice. Furthermore, lack of knowledge by referring oncologist unfamiliar with radionuclide therapy lends to it lack as frontline therapy. Internal Dosimetry Calculations Internal radiation dos e is calculated accord ing to methods determined by the Medical Internal Radiation Dose (MIRD) Committee. Dose to an organ k is defined as: h h k h kS A D) { ~ (2-1) where h is the source organ, Ah is the cumulated activity or total disintegrations per unit time of the radionuclide in the source or gan during irradiation, and S(k h) is the radionuclide S value as defined below. Much work has been done to quan tify the cumulated activity but little effort has been afforded correct calculation of the radionuclide S value. Radionuclide S values are defined for sp ecific source rs and ta rget rt region by: i T S T i i r rm r r SS T (2-2) where mT is the mass of the target region, i is the mean energy emitted per nuclear transition, and is the absorbed fraction (AF) of energy in th e target region for the ith radiation type that

PAGE 35

35 originated in the source region. AFs are the frac tion of energy that reaches the target organ from which originated in the source organ. Dividing the AF by the mass of the target yi elds the Specific Absorbed Fraction (SAF), and now the size of the organ is reflected in the fraction of energy re tained. Regarding the radionuclide S value, the SAF is the critical component as it depends on the geometry and composition of the two organs, as well as any or gans between the two in concern. The purpose of our modeling is to provide be tter approximation to the AF a nd mass of the target for more accurate clinical dosimetry. Previous Methods Detailed Radiation Transport using Chord-based Methods Current absorbed fraction values used in clin ical dosimetry, such as those implemented in the MIRDOSE program (Stabin, 1994), are acquire d from AFs reported in ICRP Publication 30 (1979). These values are outdate d for given current technological enhancements, as they are energy independent and based on ch ord length distributions from the Spiers and Whitwell studies (Spiers et al. 1978; Whitwell, 1973; Whitwell and Spie rs, 1976). As early as 1960 at the University of Leeds in England, Spiers and his students used an op tical scanning device to measure chord-length distributions of bone marrow spaces from 5 to 7 skeletal sites from 1.7, 9 and 44 year old males (Beddoe et al. 1976). Incorporating the pa th length through anisotropic bone or marrow and the frequency of occurrence a llows for calculation of the fraction of the particles energy deposited in each tissue. Coup ling these chord length distribution frequencies with range-energy relationships and providing dos e conversion values, tr abecular bone dosimetry models such as Reference Man were born. Radionuclides studied in the Spiers era we re 14C, 18F, 22Na, 32P, 45Ca, 90Sr, and 90Y, because of their application to radiation protectio n as the role of radiat ion for medical use was

PAGE 36

36 not fully understood. Thus, the Spiers data incl uded limited values of absorbed fraction for beta particles both for energy and loca tion. For example, any beta pa rticle originating in the bone volume would receive a single value regardless of energy, whereas one originating on the bone surface would receive either a high or low AF de pending on whether its energy was less than or grater than 0.2 MeV. Furthermore, the AF was given a value of unity when the source is the same as the target (such as marrow to marrow dose) with no response based on energy of the source. While fitting for the time, chord based models are outdated for current use. First, chord models are only two-dimensional and do not reflec t real trabecular microstructure. Along with that, the possibility for radiation sources to deposit their energy in the outer cortical bone is not included, hence the term infinite spongiosa tran sport model. Other assumptions made in chordbased models are that beta partic les travel in straight paths (whi ch they do not) and that bone and marrow distributions are independe nt of one another. An example of bone and marrow chords achieved through random chord length sa mpling is shown in Figure 2-8. Spiers data has been updated by to include the role of adipocytes in the marrow (Watchman et al. 2005). Furthermore, dose to the endosteum has been considered for monoenergetic electrons emitte d within the bone marrow microenvironment (Eckerman, 1985; Eckerman et al. 1985; Eckerman and Stabin, 2000). Mo reover, the importance of 3D modeling to account for electron backscatter at bone-marro w interfaces have been addressed (Bouchet and Bolch, 1999; Bouchet et al. 2000). However, these models stil l implicitly assume straight line path lengths for beta particles. Newer clinical models such as OLINDA have allowed for energy variations, but still rely on the Whitwell chord length data (Stabin, 2004). Overestimations in dose due to infinite spongiosa ex ist in all chord-based models.

PAGE 37

37 Radiation Transport using Voxel-based Methods To limit the deviation from accurate trabecul ar bone modeling and overestimations found in infinite spongiosa transport, Nuclear Magnetic Resonance (NMR) images have been used to provide 3D measurements. Studi es have provided chord-length measurements of voxelized images and discussed the difficulty in calcula ting measurements and the dependency methods have in removing voxel effects (Rajon et al. 2000). Later, EGS-nrc was used to transport particles in actual marrow cavities from 3D NMR images (Patton et al. 2002a; Bolch et al. 1998; Jokisch, 1999). Significant improvements over assuming an infinite trabecular region were made, and this voxel based approach serv ed as a benchmark for future bone dosimetry models. The work by Shah et al superseded prev ious models by accounting for radionuclide deposition in cortical bone as opposed to an infin ite transport model. Furthermore, mathematical models are no longer needed as in previous sty lized models because actual patient images are used. This technique also allows for reference da ta, as these models are fu lly scalable for patient specific absorbed fractions and S values. In the fi rst part of this work, the first full image-based reference skeletal model was presented for an adult male radionuclide th erapy patient including masses of the trabecular spongiosa a nd cortical bone for all skeletal sites. Moreover, absorbed fractions and S values were obtai ned for several radionuclides used in therapeutic regimes (Shah, 2004). Paired-Image Radiation Transport (PIRT) significantly improves previous skeletal modeling approaches by incorporat ing both the ex-vivo CT image of the skeletal site with the microstructure from MicroCT images (Shah et al. 2005b). The digital macrostructure (CT image) contains the mass of both the spongios a and the cortical bone, which was digitally segmented. The microstructure contains the in terior of the spongiosa, including the fraction of

PAGE 38

38 trabeculae and marrow within. Using the novel PIRT model, these two images can be combined and radiation particles can be tracked. This is the first voxel m odel to encompass both macrostructure and microstructure to obtain absorbed fractions and is shown in Figure 2-9. The PIRT model developed by Shah et al was used for several skeletal sites containing active marrow and S values were calculated and compared to infinite transport models. Stem Cell Distributions in Normal Marrow The assumption that the stem and progenito r cells within bone marrow is uniformly distributed leads to the averaging of dose acr oss marrow cavities. To address these concerns, Charlton et al studied the distri bution of CD34+ and CD38+ cells to adipocytes. He postulated that if adipocytes are randomly distributed in bone marrow cavities, and stem cells are uniformly distributed from adipocytes, then stem cells should be randomly located in marrow. These studies found that there was a uniform distribution of adipoc ytes and an assumed uniform distribution of stem cells (Charlton et al. 1996). However, studies by Sh ah et showed a weak or no spatial gradient in adipocytes to bone trabeculae in human bone marrow, leading to inconclusive evidence of the spa tial extent of the hematopoietic stem and progenitor cells (Shah et al. 2003). Charlton did not explic itly study the spatial distribu tion of stem cells to bone trabeculae. The first studies to quantify the spatial di stribution of stem and progenitor cells to trabecular bone were performed by Lord et al. In this murine model, a definitive spatial concentration gradient was noted, where stem and progenitor cells, identified by cytokine expression, were found to decrease as a functi on of depth in the marrow cavity in the marine femoral shaft (Lord and Hendry, 1972). These we re the first results indicating that primitive cells are preferentially located clos er to bone trabeculae, while matu re cells tend to extent further into marrow cavities. Thus, using a uniform assumption to assess dose will be incorrect.

PAGE 39

39 However, the trabecular microstruc ture of the mouse is known to be different than the structure of human bone, so appl ications are limited. Studies to directly measure the spatial distri bution of HSCs to bone trabeculae were then made on human marrow (Watchman et al. 2007). CD34 and CD31 immunohistochemical staining was performed on subseque nt slides from core biopsy samples of the iliac crest from disease-free normocellular human bone marrow. After slide preparation and imaging on the Morphometric microscope, distances to the near est bone trabecular surface was measured for both hematopoietic CD34+ cells and blood vessels. Also, distances between nearest pairs of hematopoietic CD34+ cells and vessel fragment s were determined and normalized to marrow area. The data fully supported the hypothesis by Lord that CD34+ HSCs and their progenitors follow a spatial gradient with respect to bone trabeculae within human bone marrow. Also, blood vessels seem to follow the same gradient indi cating a shared sp atial niche with stem cells. Distance-dependent weighting factors are pr ovided as an update to the MIRD schema. The most current work at the University of Florida used autopsy methods rather than biopsies to assess spatial distributions in larger samples and for several skeletal sites because of potential field of view limits (Bourke et al. in press). CD34 and CD117 immunohistochemical staining was performed on autopsy s ections from the iliac crest, lumbar vertebrae, and rib of disease free, normocellular patients Moreover, cellularity was a ssessed via digital methods to provide normalization to hematopoietically active marrow only. This data shows the same trend found in previous studies, that the CD34+ cells decrease as a function of marrow depth. Although the same trend was found, differences in the concentrations of cells were found to be dependent upon bone site. Also, the method used to stain tissue was shown to alter the concentration of cells while preserving spatial gradient.

PAGE 40

40 Stem Cell Distribution in Diseased Marrow Recent studies have considered the compos ition and function of progenitor cells during treatment for malignancies specifically diffuse large cell non-Hodgkins lymphoma (MartinezJaramillo et al. 2004; Huerta-Zepeda et al. 2000; Madrigal-Velazquez et al. 2006). Patients were identified as being in remission follo wing chemotherapy treatment by hematological parameters in blood. It was shown that ly mphoma patients had a 35% reduction in progenitor cell concentration when compared to normal. In terestingly, this reductio n in cell population was even noted after the patient was in clinical remissi on. This study thus proved deficiencies in the progenitor population in lymphoma patients during active disease and at the time of complete clinical remission.

PAGE 41

41 Figure 2-1. Human comp act and spongy bone showing location of osteons and trabeculae (U. S. National Institutes of Health). Figure 2-2. Bone remodeling and resoprtion are possible by osteoblasts and osteoclasts, respectively (Marieb, 2006).

PAGE 42

42 Figure 2-3. Osteoporotic versus normal bone (US Department of Health and Human Services). Figure 2-4. Bone Minera l Density (BMD) changes with age and gender (Yao et al. 2001).

PAGE 43

43 Figure 2-5. Stem cells may become either othe r stem cells (expansion) or further specialized cells (proliferation) (DOH, 2001).

PAGE 44

44 Figure 2-6. Hematopoietic stem cell differentiation pathways (DOH, 2001).

PAGE 45

45 Figure 2-7. Lymphatic system and lym ph circulation (Yoffey and Courtice, 1970)

PAGE 46

46 Figure 2-8. Chord lengths acr oss bone trabeculae (black) an d marrow cavities (white) at scanning angle (Shah et al. 2005a).

PAGE 47

47 Figure 2-9. Paired Image Radiat ion Transport (PIRT) combines macrostructure (skeletal size) with microstructure (trabecular stru cture) for 3D radiation transport. Figure 2-10. Schema of future reference models at the University of Florida.

PAGE 48

48 CHAPTER 3 REFERENCE SKELETAL DOSIMETRY MODEL FOR THE ADULT FEMALE Introduction In order to predict the risk of bone cancer, an accurate assessment of absorbed dose to skeletal tissues is necessary. Concerning these tissues, the limiting factor in radiation protection is the toxicity of hematopoietically active bone marrow (Sgouros, 1993; Siegel et al. 2003; Siegel et al. 1990b). Bone marrow is made up of red bone marrow (RBM), which includes the hematopoietically active elements, and yellow (inac tive) fatty bone marrow (YBM). In the adult, all axial bones contain both active RBM and inactive YBM to a certain degree dependent upon the cellularity factor (CF). The CF is the fract ion of active marrow (1-fat fraction) and is 100% at birth and decreasing with age. The RBM a nd YBM, along with trabecular bone (TB) make up the spongiosa. The outer cortex of all skeletal s ites is the cortical bone (CB), and together with the TB, the entire mineral bone (MB) is formed. In this paper the RBM within spongiosa will be defined as trabecular active marrow (TAM) and the YBM will be defined as trabecu lar inactive marrow (TIM), where all bone marrow (TAM + TIM) in regions of spongiosa will be the trabecular marrow (TM). In the adult, the appendicular skeleton (long bones) contai ns only YBM without trabeculae and has a corresponding cellularity of zero. Thus, in the medullary cavities, hematopoietically active or inactive marrow will be termed cortical activ e marrow (CAM) or cortical inactive marrow (CIM), which together make up the entire cortical marrow (CM). Particular tissues of interest for cancer risk s are the hematopoietic stem cells (leukemia) and osteoprogenitor cells (bone can cer) of the trabecular endosteu m within the red active marrow (Kuznetsov et al. 2004; Charlton et al. 1996). The hematopoietic stem cells (HSC) are assumed to be uniformly distributed throughout the TA M whereas the osteoprog enitor cells (OPC) are

PAGE 49

49 thought to be localized to 50 m regions from the trabecular bone surf ace (TBS) within the spongiosa (Bolch et al. 2007). This lends itself well to su bcategorizing bone marrow by depth. Thus, TAM50 and TIM50 are the trabecular shallow active an d inactive marrow, respectively, and comprise the soft tissues 50 m from bone surfaces. All bone marrow (TAM50 + TIM50) within 50 m from the surf aces of bone trabeculae then ar e trabecular shallow marrow (TM50). In the adult, the medullary cavities of long bones contain only YBM a nd have a cellularity of zero. However, OPCs still exist, re quiring a separate 50 m region to denote the radiosensitive cell popula tion within the CB. The cort ical shallow active marrow (CAM50) and the cortical shallow inactive marrow (CIM50) then include all of the soft tissues from 50 m from CB, and make up the entire cortical shallow marrow (CM50). The cortical bone in the medullary cavities of the long bones is termed CBMC, whereas the cortical bone in the sites containing active marrow is termed CBHC, where HC denotes Haversian canals. Previous analysis has shown that absorbed dose is an accurate measure of hematological toxicity, but only if the absorbed dose estimate is specific to the particular patient (Liu et al. 1997; Shen et al. 2002; Juweid et al. 1999; O'Donoghue et al. 2002; Sgouros et al. 1997; Sgouros et al. 1996a). At present, the vast majority of skeletal reference models (SRMs) used for these purposes are based on studies in the la te 1960s and early 1970s at the University of Leeds in which a novel optical scanning met hod was used to obtain linear chord-length distributions across several skeletal sites of a si ngle 44-year male subject. These data form an essential component of the ICRP s SRM published in ICRP Publications 30, 70, and 89 as well as clinical dosimetry codes such as MIR DOSE 3.0 and OLINDA 1.0 (Stabin, 1996; Stabin et al. 2005).

PAGE 50

50 MIRD methodologies of absorbed dose, requi res separate knowledge of the cumulated activity within the source (marrow or bone), and the radionuclide S value (absorbed dose per unit cumulated activity). Much research has been do ne to asses the cumulated activity within the source, while little has been done to make patient-specific assessmen ts of absorbed dose per unit cumulated activity (Juweid et al. 1995; Macey et al. 1995; Sgouros et al. 1993; Sgouros et al. 1996b; Siegel et al. 1990a). Furthermore, these dose calculations requires known reference values such as the fraction of energy deposited in a target tissue and the mass of those target tissues. While the Leeds data did report chord-length distributions to es timate the fraction of energy deposited, the methods used are far outda ted given the technologic al improvements in the past few years. Most importantly, chord m odels are only two-dimens ional and do not reflect real trabecular microstructure. These models also assume that beta particles travel in straight paths, which is known to be untrue. Furthermor e, radiation sources cann ot deposit their energy in the outer cortical bone and instead rely on infinite spongi osa transport. This causes overestimations in dose assessment to active marrow and an unrealistic estimate of marrow toxicity (Patton et al. 2002b; Shah et al. 2005a) Skeletal tissue masses were not given for th e Leeds 44 year male subject, and thus ICRP Publications used a variety of studies to make up for this data (ICRP, 1995, 2002). Such studies include separate databases for total marrow mass (Mechanik, 1926) and bone marrow masses (Mechanik, 1926; Trotter and Hixon, 1974) as well as reference ma rrow cellularities (Custer, 1974). Furthermore, bone trabeculae surface to vol ume ratios are from other disjointed studies (Beddoe et al. 1976). While novel at the time, these met hods and values are outdated, and lead to an unrealistic ICRP Reference Man. Not onl y do the anatomical sources of absorbed fraction

PAGE 51

51 data not match the tissue masses, there are bi ased variations due to a compilation of many independent knowledge bases. Recent work has been done at the University of Florida Bone Imaging and Dosimetry Group to more accurately determine the fractiona l energy deposition in bone for alpha emitters and beta emitters as a function of bone site (Wat chman, 2005). Also, comple te sets of skeletal macrostructural and microstructural data, in a format sufficient for radiation transport simulations, have been established for an imag e-based skeletal reference model for the adult male at an age representative of cancer pa tients undergoing radionuclide therapy (Shah et al. 2005b). Relevant tissues of the microstructure, such as the bone trabeculae, bone endosteum, and marrow cavities were accounted for through imagebased techniques. Skeletal tissue masses, which are important in determin ing radionuclide S values, have b een reported to update those summarized in ICRP Publications 70 and 89. These reference skeletal dosimetry models can be used for dose assessment in radiation protection or therapeutic medicine To provide an accurate dose assessment in, the reference skeletal dosimetry model must be patient-specific. That is, it should be designed to match as closely as possible to the average individual in the worker populati on for radiation protection. If used in therapeutic radionuclid e planning, the reference model should match the patient being treated, including tissue masses and marrow cellulariti es. The skeletal tissue doses are then to be used only to establish dose limits based upon an acceptable risk of cancer induction or radiation induced effects (stochastic or non-stochastic). There are current studies underw ay to estimate marrow cellularity in patients, including non-invasive bone scanning and magnetic re sonance imaging techniques (Schick et al. 1995). Methods for estimating patient sk eletal spongiosa volume and active marrow masses, as well as

PAGE 52

52 scaling models have also been provided (Brindle et al. 2006a; Brindle et al. 2006b; Brindle et al. 2006c; Pichardo et al. 2007). Previous studies have upd ated ICRP Reference Man with tissue masses and absorbed fraction data for the male model, however, work had not been done to assess differences based on gender. To date, there is no skeletal reference female model and this paper serves to present a companion sk eletal reference model to the adult male. Materials and Methods Female Cadaver Selection A female cadaver was selected through the St ate of Florida Anatomical Board located on the University of Florida campus. The cadaver wa s selected in order to represent an average, healthy female for use in therapeutic medical dosim etry. Selection criteria included (1) an age between 50 -75 years to be representative of a cancer patient, (2) a body-mass index between 18.5 25 kg m-2 (CDC recommended healthy range), and (3) a cause of death presenting a low probability of skeletal deterioration. The subjec t selected was a 64 year old female, 63 inches tall and 135 pounds. The subject thus had a body-mass index of 23.83 kg m-3 and died of respiratory complications. In-vivo Macrostructural Image Database Several imaging modalities and image processing techniques were required in order to provide a complete skeletal data base for the reference female model. First, two whole body invivo images were acquired using a Siemens Sensation 16 multi-slice helical CT in the Department of Radiology at Shands Hospital at 1mm slice thickness, shown in Figure 3-1. In order to capture the whole body, the top of the body was imaged first, then the bottom. These invivo images were used to construct three-dimensi onal anatomical models of those skeletal sites that could not be completely harvested such as the entire rib cage or fully intact cranium.

PAGE 53

53 Manual image segmentation of trabecular s pongiosa and cortical bone was performed on CT_Contours, a program based upon Interactive Data Language (IDL) version 5.5. Using this program, patient anatomy was overlaid with a contour containing volume measurements based on the number of voxels manually selected. These contours provide the basis for the tomographic models used in radiation transport codes, as well as the tissue masses. The total voxels of marrow in the digital image of bone site j is the spongiosa volume (SV)j, whereas all other bone voxels in that bone site are cortical bone volume (CBV) j. An image of the contoured in-vivo ribs is shown in Figure 3-2 and contoured in-vivo images of the cranium showing the separation of the left parietal, ri ght parietal, occipital, frontal, facial bones, and other bones (sphenoid, ethmoid, temporal) are show in Figures 3-3 and 3-4. Furthermore, the images will provide a basis for planning the harvest of skeletal sites. The images were reconstructed using a B80 bone filter at the best possible resolution to capture the entire body. The CT images were stored within the UF Department of Nuclear and Radiological Engineering for data storage and image processing. The right humeral head was found to be abnorma l, possibly due to some sort of accident, and was not included in mass estimates. As shown in Figure 3-5, the cortical bone in the shaft seems to extend through to the proximal humeral h ead. This deformation did not lead to a doubt in the skeletal health of the i ndividual, but merely some sort of injury at some point in the subjects life. To be safe, the macrostructural and microstructural data from the left (normal) humeral head was used in both the right and left volume and mass estimates. Ex-vivo Macrostructural Image Database Next, skeletal harvesting of those 13 skeletal sites containing the highest percentages of active marrow was performed. These included the cranial cap (parietal, occipital, and frontal lobes), clavicles (2), scapulae (2), entire vert ebral column (cervical, lumbar, thoracic, and

PAGE 54

54 sacrum), ribs (3 from each side), os coxae, prox imal humeri (2), proximal femora (2), mandible, and sternum. These sites were cleaned, labeled, bagged, and stored in a freezer until ex-vivo micro imaging could be performed. High-resolution ex-vivo CT images were acquire d from the harvested skeletal sites at 1.0 mm slice thickness, with the best possible in-plane resolution (limited by sample size). The field of view (FOV) ranged from about 5.0 cm for the ribs and 33.2 cm for the os coxae. CT data for all skeletal sites is given in Table 3-1. Thes e higher resolution CT images were used to determine volumes of both cortical bone and trab ecular spongiosa for each skeletal site through the program CT_Contours based upon Interact ive Data Language (IDL) version 6.0. Samples of the ex-vivo contoured images are shown for the femur and lumbar vertebrae in Figures 3-6 and 3-7, respectively. Spongiosa vo lume, cortical bone volume, and null space are signified by red, blue and green, respectively. The separation of the femoral head and neck is shown, and were used distinctly because of know n differences in trabecular microstructure as a result of weight loading. These images will also be used to determine the best location in that particular bone site to extract a sample for micr o imaging. Eventually, they will provide the 3D anatomic macrostructural model for paired-image radiation transport (PIR T) simulations. Using this system, particles can start within the cortical bone (blue) and the fraction of radiation entering the spongiosa (red) or escaping (green) will be known. MicroCT Microstructural Image Database After all ex-vivo CT images were acquired and volumes were determined, physical sections of marrow intact trabecular spongiosa we re cored from each bone site and imaged via microCT. There were limitations on sample size due to bone shape, cost, and the microCT imaging system. The maximum size of each bone cube was used to take advantage of microCTs large bore size compared to Nuclear Magneti c Resonance (NMR) imaging. Thus, only one

PAGE 55

55 single sample was taken from each bone, at best re presenting 80% of the total spongiosa in that bone site (vertebral body samples). This is a clear advantage over the representative 25% spongiosa acquired using NMR techniques (Shah et al. 2005b). Micro CT of the 35 cuboidal samples from the 13 major skeletal sites was pe rformed on a desktop cone beam CT80 scanner (Scanco Medical AG, Bassersdorf, Switzerland) at a 30 micron resolution. This resolution is relatively much higher than the previously 60 mi cron resolution used in the UF reference male skeletal database. Post-acquisition image processing steps were then performed using gradient assisted manual segmentation to determine marrow volume fraction (MVF), and trabecular bone volume fraction (TBVF). Steps to determine the MVF in cluded (1) extraction of a region of interest (ROI) to remove cortical bone from the image; (2) applying a median filter to improve the signal-to-noise ratio (SNR); (3) de termining a threshold to best cl assify the voxels as either bone or marrow; and (4) segmentation of the ROI into a binary image based on the threshold graylevel value that was chosen. The thresholdi ng of trabecular spongiosa was performed using gradient magnitude techniques (Rajon et al. 2006). Visual inspection of the image gradient magnitude demonstrates the ability to retrieve sample volumes with 1% accuracy at 30-m voxel resolution. Data for the 35 microC T images are given in Table 3-2. Once filtered and segmented, the MVF and TBVF were determined by the ratio of marrow voxels to total voxels and bone voxels to total vo xels (or 1-MVF), respectively. Next, MVFs were averaged dependent upon bone site and used to report tr abecular active and inactive marrow masses. To determine shallow marrow masses in trabecular bone, the shallow marrow volume fraction (SMVF) was found by processing the microCT image and summing (1) marrow voxels that are adjacent to a bone voxel and (2) 2/3 of the volume of the next bone voxel, divided

PAGE 56

56 by the total voxels in the digital image. Since microCT images are 30 m in resolution, this leads to the best estimate of the amount of tissue 50 m from bone surfaces. MVFs were weighted by SV (when available) in each respective site of acquisition to repor t total bone mass. An entire skeletal mass database was established for a female reference dosimetry model, including percentages of active marrow for thos e bone sites that contain active marrow. Reference percentages of active marrow by bone site are shown in Figure 3-8. Mass Calculations Mass estimates required both the macrostructu ral volume information (contoured ex-vivo CT images), microstructural volume inform ation (filtered and segmented microCT spongiosa images), and the volume percentage of hematopoi etically active versus inactive bone marrow for each skeletal site (marrow cellularity). Marrow cellularity can vary anywhere from 10% to 100% (no adipose tissue) with age, shown in Fi gure 3-9. Therefore, this reference individual does not have one single value of total active marrow mass, but a range of potential masses dependent upon the marrow cellularity chosen. The reported masses us e those cellularities provided by ICRP Publication 70, purely as a co mparison purpose only. Current ICRP reference cellularities based on bone site and age are shown in Figure 3-10. Bone sites containing active marrow Masses were calculated for the 13 major bone si tes containing active marrow. Densities for all skeletal regions were taken from ICRU 46 (1992) and are in Table 3-3. For those sites where multiple cuboidal samples were taken, the marrow volume fractions were volume averaged. The mass of trabecular active marro w (TAM), trabecular inactive marrow (TIM), trabecular marrow (TM), shallow active marrow (TAM50), shallow inactive marrow (TIM50), trabecular shallow marrow (TM50), trabecular bone (TB), and corti cal bone (CB) at each skeletal site are calculated as:

PAGE 57

57 RBM TAMCF MVF SV m (3-1) YBM TIMCF MVF SV m 1, (3-2) TIM TAM TMm m m (3-3) RBM TAMCF SMVF SV m 50, (3-4) YBM TIMCF SMVF SV m 150, (3-5) 50 50 50TIM TAM TMm m m (3-6) MB TBCF TBVF SV m (3-7) MB CBCBV m (3-8) where SV is the spongiosa volume, CBV is th e cortical bone volume, MVF is the marrow volume fraction, TBVF is the trabecular bone vo lume fraction (1-MVF), SMVF is the shallow marrow volume fraction, RBM is the density of RBM, YBM is the density of YBM, MB is the density of mineral bone, and CF is the cellularity f actor for skeletal site j. As noted above, in most sites the SV came from the high resolution ex-vivo image, whereas others came from the in-vivo image. Marrow volume fractions were av eraged in cases where there were more than one MicroCT sample acquired for the same bone site. Marrow volume fractions were not available for the extremities as these bone sites were not harvested and microstructura l data were used from a simi lar bone site. The marrow volume fraction of the humerus was used for the proximal and distal ends of th e lower arm bones. For the lower leg bones, the marrow volume fr action of the femoral neck was used.

PAGE 58

58 Bone sites not containing active marrow The shafts of the long bones in adult (femur, tibia, fibula, hum erus, radius, and ulna) do not contain active marrow or trabeculae and masses were calculated only fo r cortical inactive marrow (CIM), shallow inactive cortical bone (CIM50). Volumes of the CIM and entire cortical bone were reported through direct image se gmentation. ICRP 70 and 89 report marrow cellularity in the upper half and lower half and not for the entire femur and humerus, so in some cases the upper and lower halv es were reported individually. Masses of the shallow inactive cortical marrow (CIM50) in the shafts of long bones were mathematically derived. Concentric cylinder models were used to stylistically model bone site j, with a cortical marrow radius and co rtical marrow volume (CMV) defined as: CMVjRMC2LMCj, (3-10) j MC j j MCL CMV R, (3-11) where RMC is the medullary cavity radius, LMC is the medullary cavity length, and CMV is the segmented medullary volume from the contour. The shallow marrow volume fraction in the medullary cavities (SMVFshaft) is the ratio of the volume of the cortical shallow marrow (CMV50) to the volume of th e total medullary marrow (CMV). CM50 is defined as a 50 m thick cylindrical shell of ma rrow at the outermost region of the medullary cavity, which includes only and cortical inactive marrow (CIM50) in the adult. The SMVFshaft) is calculated as follows: 2 2 5050 1j MC j MC j j shaftR m R CMV CMV SMVF (3-12) The mass of the CIM, CIM50, and CB for the medullary cavity is then:

PAGE 59

59 YBM CIMCMV m (3-13) YBM shaft CIMSMVF CMV m 50, (3-14) MB CBCBV m (3-15) where the definitions are the same as above. No te that long bones may be further divided into upper and lower halves by dividi ng the mass by two. Again, volum e fractions were weighted by SV (when available) for each respective site of acquisiti on to report total bone mass. Transport Modeling Both sets of images (ex-vivo CT and ex-v ivo microCT) were combined under PairedImage Radiation Transport (PIRT) via methods de scribed previously by Shah et al (2005). Using PIRT, the 3D model of trabecular spongiosa was paired with the patient CT anatomy and beta particles were transported via EGSnrc. Possible tissue sources were the trabecular active marrow (TAM), trabecular inac tive marrow (TIM), tr abecular bone volume (TBV), trabecular bone surface (TBS), and corti cal bone volume (CBV). These sources could irradiate tissue targets including TAM and the shallow trabecular active marrow (TAM50). Cellularity was varied from 10% to 100% and energy from 0.01 MeV to 10 MeV. Particles were transported until the kinetic energy is zero, and until coefficients of variation in the absorbed fraction are less than 1%. For the shafts of the long bones, a stylisti c model was created in MCNP given the volume and lengths acquired from the segmented CT patien t data. Tissue sources can be cortical inactive marrow (CIM), cortical bone surface in the medullary cavities (CBSMC) and the CBV. Particles were transported until the source was evenly di stributed throughout the bone site volume. The tissue target in the bone s not containing active marrow was the cortical shallow inactive marrow (CIM50) and Tally 2 identified the energy (MeV) deposited in this region.

PAGE 60

60 Absorbed Fractions Both the PIRT and MCNP techniques for radi ation transport produced absorbed fractions using various source to target tissues for each bo ne site, and are fully dependent upon cellularity and energy. When divided by the mass of the ta rget (and cellularity), the specific absorbed fraction (SAF) is determined. In bone sites wher e more than one MVF existed, particles were transported several times to combine each MVF with the corresponding SV. These values were then weighted by the MVF for each individual AF for one energy-dependent absorbed fraction per bone site j as shown below: TAM TAMj MVFxTAM TAM x xMVFx x (3-16) where (TAMTAM) is the fraction of energy deposited in the TAM from a particle emitting radiation from the TAM, and MVFx is the MVF for bone site x. For bone sources, TBVF was used in place of MVF. Results The University of Florida Reference Female (UFRF) was first matched to a database of cadavers to assess whether the cad aver used was similar to the average in terms of spongiosa volume, shown in Table 3-4. Next, the volum e of spongiosa and percentage of total was compared to the University of Florida Reference Male (UFRF) in Table 3-5. Volume fractions are given in Table 3-6, which we re then used to determine bone masses. Table 3-7 shows the percentage of trabecular bone a nd cortical bone to mineral bone and is compared among UFRF, UFRM and ICRP 89 reference male. Since no ICRP female model exists, a comparison to the ICRP Reference Male must be made. To make su ch a comparison, cellularities were assigned to be the reference value, although any va lue of cellularity can be used.

PAGE 61

61 The ratio of total UFRF trabecu lar bone mass and total corti cal bone mass to ICRP 89 total masses is used as a method of fair comparison si nce there is too much variation in specific skeletal sites. Both total marrow masses and shallow marrow masses for active and inactive marrow are given for UFRF in Tables 3-8 and 3-9. Total marrow masses were compared to ICRP 89 and UFRM, however, si nce ICRP does not report shallow marrow masses, UFRF was only compared to UFRM for this case. Tabl e 3-10 lists surface-to-volume (S/V) ratios for UFRF, URFM, and ICRP 70 as well as other studies by Beddoe et al and Lloyd et al. Surface areas for trabecular bone regions and all cortical bone regions including both Haversian canals and medullary cavities is provi ded in Table 3-11. Fractional mass weights for each source and targ et region in UFRF are listed in Tables 312 and 3-13. These values were used to dete rmine skeletal averaged SAFs. UFRF used reference cellularities from ICRP Publication 70 (1995), wh ereas the ICRP model used cellularities from Eckerman & Stabin (2000). Sp ecific absorbed fraction data for UFRF is given in the appendices for any combination of source (TAM, TBS, TBV, TAM50, TIM) and target (TAM, TAM50) regions in the entire skeleton co ntaining active marrow for a range of cellularities (10% to 100%). Ta bles and graphs of SAF for each bone site, given all source-target combinations, are given in Appendix A and B, re spectively. Appendix C then contains SAF and graphs of all source-target combinations compar ing all bone sites independ ently. In bone sites not containing active marrow (shafts of the long bones), absorbed fracti on data is given for combinations of sources (CBV, CIM, CBSMC) and target (CIM50) and SAF are tabulated graphed in Appendix D. Finally, Appendix E contains tables and graphs of skeletal averaged SAFs and comparisons to models from Stabin and Siegel (2003).

PAGE 62

62 Discussion Spongiosa Volume and Percentages The spongiosa volume and percentage of to tal for UFRF was first compared to a population of database of 20 cadavers, 10 male an d 10 female, provided by Brindle et al (2006). This database included the UF Reference Male subject. The spongiosa volumes were averaged by skeletal site for female cadavers, male cadavers, and then an average of all cadavers of either sex. Table 3-4 shows the per centage of spongiosa volumes for the skeletal sites containing active marrow for UFRF, the average female, the av erage male, and the average of either sex. UF Reference Female seems to match well to the average female with respect to total spongiosa volume and percentages of to tal. The total spongiosa volume (1535 cm3) is remarkably close to the average female (1521 cm3) with a difference less than 1%. UFRF has less spongiosa in the os coxae and more in the sacrum, but overall matches well to the average female. These differences are noted in many bone sites due to normal patie nt variation. When compared to the average male, total spongios a volumes are greater in the average male population (2221 cm3) and there seems to be more inherent variation in male spongiosa volume. In particular, the percentage of spongiosa in th e ribs and scapulae vary greatly. Again, one would expect these differences due to natural skeletal variation. A comparison of spongiosa volume and percenta ge of total for UF Reference Male and Female are given in Table 3-5. Although the spo ngiosa varied, as expected due to patient size, the percentage of spongiosa volume for both refe rence individuals was comparable. A higher percentage of spongiosa volume was seen in th e os coxae of UFRF compared to UFRM, and lower percentages of spongiosa volume in the femora and humeri. ICRP 89 does not report spongiosa volume explicitly, thus no comparison can be made.

PAGE 63

63 Volume Fractions Averaged marrow volume fraction (MVF), bone volume fraction (BVF) and shallow marrow volume fraction (SMVF) by skeletal site are shown in Table 3-6. It is noted that the MVF are similar for a single bone site except in the cranium and femora. The cranium has MVF variation from 11% marrow in th e occipital lobe to about 39% 43% marrow in the frontal and occipital lobes. An image of the microstructure of the frontal and occi pital lobes are given in Figure 3-11 and 3-12, respectively. The femoral neck is 87% marrow, whereas the femoral head is 71% marrow. An image of the microstructu re showing the differences between the femoral neck and head is given in Figure 3-13 and 3-14, respectively. Since the microstructure varied greatly in these skeletal sites, volume averaging seemed best. For completeness, an image of that bone site with the highest percen t of marrow is given. The sternum is shown in Figur e 3-15, having 99% marrow. Also, to show the changes in microstructure throughout the vert ebral column, an image of the 3rd cervical vertebrae (76% marrow) and the 5th lumbar vertebrae (91% marrow) are shown in Figure 3-16 and 3-17. The marrow volume fraction increases as you move down the vertebral column. Tissue Masses Masses for the trabecular spongios a region, cortical bone regi on and total mineral bone for all skeletal sites (both active and inactive marrow) are in given in Table 3-7. ICRP reference cellularity was used for each bones site. UFRF has a trabecular bone mass of 968 g, cortical bone mass of 4723 g and total mineral bone mass of 5691 g. Thus, mineral bone is 17% trabecular bone and 83% cortical bone. A comparison to ICRP 89 Reference Male model reveals that reference values yield 20% trabecular bone (1100 g) a nd 80% cortical bone (4400 g). Furthermore, trabecular bone makes up 17% of total mineral bone and cortical bone makes up the remaining 83% and overestimates trabecular bone mass compared to UFRF (ratio of 0.88).

PAGE 64

64 UFRM was 22% trabecular bone (1 267 g) and 78% cortical bone (4388 g). Total mineral bone mass was similar in both UFRM (5656 g) and UFRF (5691 g), which is slightly higher than the ICRP value (5500 g). Compared to both UFRM and ICRP 89, the fe male model was found to be much lower in trabecular bone mass. Interestingly, the ICRP 89 value is in the middle at 20%, and perhaps represents an average for the tw o. The percentage of total co rtical bone to mineral bone was found to be slightly higher for UFRF compared to both male models with a ra tio to ICRP of 1.07. Thus, the female model seems to have less tr abecular bone and more cortical bone when compared to males. This could be a result of the differences in bone resorption by gender and the effect of osteoporosis, or simply pa tient variation in skeletal bone mass. Active marrow masses for UFRF are provided in Table 3-8. UFRF has 1984 g in inactive marrow, 686 g in active marrow, for a total marro w mass of 2670 g. Thus, marrow is 26% active and 74% inactive. It is empha sized again that the mass of UF RF is not bound by cellularity, and is given at ICRP 70/89 referen ce values only for comparison. Compared to ICRP 89, both the active marrow mass (1170 g) and inactive marro w mass (2480 g) yielding 32% active marrow and 68% inactive marrow. Thus, ICRP 89 ac tive and inactive marrow masses are severely overestimated for the UFRF with ratios of 0. 59 and 0.80, respectively. UFRM had an active marrow mass of 2406 g and an inactive marrow ma ss of 1170 g. Thus, this overestimation in marrow mass was also noted in the URFM data, although not as significan tly (ratio of 0.91 and 0.97). This is due to the mu ch lower total marrow mass of 2 670 g found in UFRF, compared to 3650 g in ICRP 89 and 3469 g in UFRM. Since the active marrow is the region for hematopoiesis, this is a red flag for marrow toxicity and cancer risk when using ICRP 89 reference values to estimate radiation dose in females.

PAGE 65

65 UFRF has a shallow marrow mass of 240.8 g, which is 20% shallow active marrow (48 g) and 80% shallow inactive marrow (193 g). These ar e shown in Table 3-9. ICRP does not report this data, therefore only a comparison to UFRM can be made. Again, these were calculated at reference cellularity, although any cellularity could be chosen. Shallow active marrow was only 100 g in UFRM, with 226 g of sha llow inactive marrow. Accordin gly, UFRM has 31% of active shallow marrow and 69% of inactive shallow marro w, with a total shallow mass of 326.3 g. Total shallow active marrow was less in UFRF due to a lower overall active marrow region. However, UFRF had a 10% higher inactive compone nt of shallow marrow, leading to dosimetry concerns. This could be simply patient variati on, or could signify hei ghtened osteoporosis in females due to lower volume of active elements in general. Surface to Volume (S/V) Ratios Table 3-10 lists S/V ratios of the trabecular micr ostructure in all skeletal sites in UFRF, along with a comparison to UFRM and data from the University of Leeds studies, which provide the basis for the ICRP Reference Man model. The mean S/V ratio was found to be 21.2 mm2/mm3 across 13 skeletal sites with a standard deviation of 7.2 mm2/mm3. This value is slightly higher than the UFRM model (18.5 mm2/mm3), but within the standard deviation of 4.2 mm2/mm3. The Leeds data indicated a smaller ratio of 16.1 4.8 mm2/mm3, while other studies in ICRP cited even lower values of 11.2 5.0 mm2/mm3 (Lloyd and Hodges et al). ICRP 70 uses a single value of 18 for all bone sites, wh ereas this data lists S/ V ratios for 13 separate skeletal sites allowing for a better esti mate of the shallow marrow masses. Surface Areas Table 3-11 lists surface areas in all skeletal sites for both trabecular and cortical bone regions. UFRF was found to have 8.23 m2 of trabecular bone surfaces, 7.38 m2 of cortical bone surfaces within regions containing haversian canals, and 0.06 m2 of cortical bone surfaces in the

PAGE 66

66 medullary cavities. This leads to 47% of th e total surface area being cortical bone within Haversian canal tissues, 0.4% of cortical bone surface in the medullary cavities and 53% of bone surface in the trabecular regions. ICRP does not list bone surface area in the cortical bone of medullary cavities, but does list 6.50 m2 of cortical bone surfaces (38%) and 10.50 m2 of trabecular bone surfaces (62%). UFRM had similar values to ICRP with 6.86 m2 of cortical bone surfaces in the Haversian canal tissues (37%) and 11.74 m2 of trabecular bone surfaces (63%). In the medullary cavities, UFRM had a surface area of 0.17 m2 (0.9%). It seems that the surface areas and S/V ratios reveal that UFRF has a higher surface to volume ratio and a lower overall trabecular bone su rface. Compared to ICRP, UFRF has a much lower (0.78) ratio of trabecular surface area and more cortical bone surface area (1.14). UFRM also had a higher cortical bone su rface area than ICRP, but had hi gher trabecular su rface area as well. This could be an inherent difference in the cadaver selected, or a real difference among female bone anatomy. It seems that there is similar cortical bone volume but less trabecular spongiosa (both bone and marrow). Due to overa ll size differences, this difference must be thicker cortical bone and smaller trabeculae and marrow cavities. Weighting Factors The fractional distribution of tissue masses at refe rence cellularities for each skeletal site is given in Tables 3-12 and 3-13. B one site weighting factors are pr ovided for the fraction of total active marrow (TAM+CAM), total inactive marro w (TIM+CIM), total shallow active marrow (TAM50+CAM50) and total inactive shallow marrow (TIM50+CIM50). For surface sources, the fractional distribution of tissue masses are gi ven for the TBS, CBS and total bone surface (TBS+CBS). For volume sources, weighting fact ors are provided for the TBV, CBV, and total bone volume (TBV+CBV).

PAGE 67

67 UFRF had 68.9% of total active marrow (fTAM+fCAM) concentrated in the os coxae (26%), thoracic vertebrae (16.2%), lumb ar vertebrae (14.5%) and ribs ( 12.2%). The same bone sites within the ICRP accounted for 79.7% of total active marrow (fTAM): os coxae (33.3%), thoracic vertebra (17.4%), lumbar vert ebrae (9.8%), and ribs (19.2%). The largest fraction of bone trabecular surfaces in the UFRF (fTBS) is from the distal femora (23.9%), proximal tibiae (19.1%), os coxae (11.6%), and proximal femo ra (8.2%). In ICRP, the TBS mass was most concentrated in the thoracic vertebrae (28.2%), lower femora (16.9%), proximal femora (16.7%), and the cervical vertebrae (11%). The fraction of trabecular bone surface in the os coxae was only 1.8% in the ICRP values. A great deviation in the os coxae was also found in UFRM, when comparing to current ICRP fractions. The largest fraction of bone tr abeculae mass in the UFRF (fTBV) is found in the distal femora (23%), proximal tibiae (18.4%), craniu m (16.8%) and proximal femora (7.9%). In the ICRP model, most of the TBV ma ss is found in the thoracic vert ebrae (27.6%), with the rest located in the lower femora (16.5%), proximal femo ra (16.3%) and cervical vertebrae (10.7%). These values seem to differ greatly in the cran ium, with only 2.6% noted in ICRP values. Finally, the fraction of corti cal bone volume in UFRF was found to be highest in the cranium (10.7%), distal femora (9.3%), os coxa e (8.7%) and the proximal tibiae (8%). ICRP values of fCBV were found to be highest in the lower femora (20.7%), proximal femora (14%), cranium (12.8%), and ribs (12%). These bone si tes comprise 36.7% of the UFRF cortical bone volume, and 59.5% of the ICRP cortical bone. The low fractional mass in UFRF is due to the cortical bone being distributed over mo re regions in the extremities.

PAGE 68

68 Specific Absorbed Fractions (SAF) SAF for individual bone sites containing active marrow SAFs for each individual bone site is provided for two targets TAM and TAM50 and all five sources in Appendix A (table s) and B (graphs). For this di scussion, a single bone site will be used since all bone sites followed the same general trend with slight differences in SAF value. The SAF for the TAM irradiating either the TAM or the TAM50 for the lumbar vertebrae is shown in Figure 3-18 and 3-19, resp ectively. Both source target combinations followed the same decrease in SAF with energy. SAF decreases with higher ener gy since higher energy particles will have enough energy to escape the TAM and de posit energy elsewhere. At low energies and at high cellularity (no fatty marrow), the absorbed fraction is at a maximum because all of the particles are absorbed in the ac tive marrow. Conversely, at lo w energies and low cellularity (fatty marrow), the particles are absorbed in the fat and there is a low absorbed fraction to the active marrow. Calculating the SAF requires a di vision of the mass of the target which depends on the cellularity, thus, cellularity trends are switched. There is a point of convergence where cellularity no longer causes dose implica tion at an intermediate energy (<1 MeV). Figure 3-20 and 3-21 shows the SAF for each s ource irradiating either the TAM or TAM50 for the lumbar vertebrae. While trends are sim ilar, there is a difference in the overall value of each source-target combination depending on the targ et. When the target is the entire active marrow, the SAF is highest in the TAM for all energies. However, when the target is the shallow active marrow, the SAF is now highest fo r the TBS, as this source would cause more dose implications to the voxels dire ctly adjacent. If this is the true location of radiosensitive cells, dose would be incorrectly assessed if averaging over the entire marrow cavity. The (TAMTBS) followed the same trend as an active marrow source, but was slightly lower in value, shown in Figure 3-20. This is because there would be less energy deposited in

PAGE 69

69 the active marrow for a particle starting in the trabecular bone su rface than there would be for actually starting in the active marrow. The (TAMTIM) slightly increases with increasing energy and levels off to the same value as activ e marrow and trabecular bone surface values. A particle would need to have enough energy to leave the inactive marro w in order to deposit energy into the active marrow regions. The (TAMTBV) is the second lowest value and follows the same trend as an inactive marrow s ource. The lowest SAF is found for a cortical bone volume source where the (TAMCBV) increases with increasing energy and then decreases at very high energies. This is because a particle needs to have a significant amount of energy to leave the cortical bone volume and deposit energy into the active marrow. At very high energies, the particle leaves the active marrow to deposit dose elsewhere. The SAF for all sources to a TAM50 target follows the same trends with differences only in values (Figure 3-21). For a TAM50 target, the highest SAF is fo r a TBS source at all energies or a TBV source at intermediate energies. This is because this is closest in proximity to the TAM50 since, by definition, this region is 50 micr ons from the surface of all trabecular bone regions. The SAF is also large for the TAM source, and this is expected as well. SAF for either a TIM or CBV is very low for the shallow active marrow, as most particles cannot escape the adipose tissue or cortical bone to deposit energy into the bone trab eculae which are well within marrow cavities. SAF for all bone sites co ntaining active marrow Appendix C lists the SAF data comparing all so urce-target combinati ons for all bone sites containing active marrow independently Figures 3-22 and 3-23 graph the (TAMTAM) and (TAM TAM50) for all bone sites, re spectively. The TAM50 target followed the same trends as the TAM target. At low energies, the (TAMTAM) was highest for those bone sites with

PAGE 70

70 small marrow cavities (e.g. mandible, cervical vertebrae, sternum, cr anium). At higher energies, the same trend was followed except for the mandible. The (TAMTAM) is decreasing over energy because at higher energies more particle s leave the marrow regions. The bone site with the lowest (TAMTAM) was the os coxae, where there is a large amount of active marrow and large cavities. The (TAMTAM50) was also high for bone sites with low marrow volume fractions and more bone volume fractions (e.g. mandible, cranium, cervical vertebrae, scapulae) at low energies. The same trend was followed except for the cranium, which does not decrease as much as other bone sites at high energies. The (TAMTAM50) is decreasing over energy because at higher energies more particles l eave the shallow marrow regions. Again, the bone site with the lowest (TAMTAM) was the os coxae, where there is a large amount of active marrow and large cavities. SAF for all bone sites not containing active marrow SAF tables and graphs for all source-target combinations in the sh afts of long bones are given in Appendix D. The SAF for the CIM irradiating the CIM50, CBSMC, and CBV for all skeletal sites, are given in Figures 3-24 through 3-26, respectively. As expected, (CIM50CIM) decreases with energy, as particles can escape the shallo w inactive marrow, as shown in Figure 3-24. This seems to become mo re drastic after 1MeV. The long bones of the arms (radius, ulna) have the largest SAF, with th e smallest being the humerus and femur. This seems plausible since these bone sites are larg er and contain more marrow space to absorb particles, with less being deposited in the shallow regions. The (CBSMCCIM) also decreases with energy, and is shown in Figure 3-25. As pa rticles gain energy they can escape the cortical

PAGE 71

71 bone surface and deposit energy in to the shallow cortical marrow For this case the ulna and radius also had the largest SAF, with the sm allest again being the humerus and femur. Figure 3-26 plots (CBVCIM), which decreases with energy drastically from 0.1 MeV to 1 MeV and then tapers off until 10 MeV. Due to the low count of particles escaping the cortical bone volume, good statisti cs are not as realizable for th is source. Again the long bones of the arms had the largest SA F (radius, ulna), with the humer us and femur having the lowest SAF. SAF from sources emitting from the bone surfaces provide less radiation dose to the shallow cortical marrow than would a source starti ng within the cortical inactive marrow itself. This is due to the high density of the cortical bon e which is difficult for particles to traverse, and those that do will be significantly reduced in energy. Skeletal Averaged Absorbed Fractions (AF) Current models for marrow dose estimates do not provide specific absorbed fractions by bone site, and instead use one value that is averaged over all skeletal sites. These include models used in MIRDOSE 3.0 (Eckerman & Stabin 1994 ) and OLINDA 1.0 (Stabin and Siegel 2003). Furthermore, these models do not use Paired Im age Radiation Transport to take advantage of true trabecular microstructure, nor do the tissue masses come from the same patient anatomy. To provide a fair comparison, absorb ed fractions for each bone site have been averaged over the entire skeleton. To facilitate this, UFRF bone site weighting f actors (Table 3-12 and 3-13) were used, along with ICRP reference cellularities (Figure 3-10). A ppendix E gives all tables and graphs for skeletal averaged UFRF AFs, as well as comparisons to the Stabin and Siegel models (2003). The skeletal averaged SAF for UFRF compar ed to Stabin and Siegel (SS) for a marrow source irradiating the total active marrow and for bone sources irradiating the total active marrow

PAGE 72

72 are shown in Figures 3-27 and 3-28, respectivel y. Under current ICRP assumptions, the total active marrow is the surrogate tissue for the hema topoietic stem cell (HSC ) population, so this will be the estimated dose to that radiosensitive population. When the marrow itself is a source, the UFRF model matches well with the SS model at energies less than 1 MeV. However, beyond this point, SS converges to 0.431, whereas AF continues to decrease until 0.103 for UFRF. This discrepancy has been noted to be due to energy independe nt modeling approaches in the SS model (Shah et al., 2005a). Also shown in Figure 327 is the AF for a TIM source, which is not provided in the SS models. Here it is noted that the decrease in AF for the TAM source follows the TIM source past 1 MeV only in the UFRF model, which is to be expected. Discrepancies in bone source models irradiati ng the total marrow are shown in Figure 3-28. Both the TBS and TBV sources seem to be a re asonably accurate when comparing UFRF and SS models at low energies. Differences in values are due to tissue definiti ons, which are described below. The SS model continues to overestimate dose at energies beyond 1 MeV, as explained above. Also shown in Figure 3-28 is the AF to the total marrow from a CBV source for UFRF. This was not provided for comparison in the SS models. The skeletal averaged SAF fo r UFRF compared to Stabin and Siegel (SS) for a marrow source irradiating the shallow trabecular marrow and for bone s ources irradiating the shallow trabecular marrow are shown in Figures 3-29 and 3-30, respec tively. Under current ICRP assumptions, the shallow trabecular marrow is the surrogate tissue for the osteoprogenitor cell (OPC) population, so this will be th e estimated dose to that radios ensitive population. It is noted that this region was once termed the endos teum, which was updated from a 10m region adjacent to bone surfaces in the SS model to a 50m region adjacent to bone surfaces in the UFRF modeling approach. Figur e 3-29 shows marrow sources ir radiating the shallow active

PAGE 73

73 marrow for the UFRF and SS model. In this case, differences are noted at low energies, whereby the UFRF has a higher fraction of energy being depos ited in the shallow active marrow. This is due to the updated definition of the endosteum, with a higher vol ume encompassing a higher AF. The same divergence at high energi es is again prevalent, as discussed previously. Again, a TIM source for the UFRF is provided. Bone trabeculae sources irradi ating the shallow active marrow are shown in Figure 3-30. Here, similar re sults are found between UFRF and SS modeling approaches for both the TBS and TBV sources as described above. Differe nces at low energies are explained by the updated tissue definitions th at confirm the OPC population. The SS model has a much higher value of AF for a CBV s ource when compared to the UFRF model. Discrepancies between the SAFs include differences in modeling approach (voxel based versus chord based), imaging technique (3D Mi croCT versus 2D optical scanning), and gender differences in macrostructure a nd microstructure (64y female vers us 44y male). Perhaps one of the main differences is the tissue region definition used for both trabecular active marrow and shallow trabecular active marrow as described in Shah et al (2003). This also lends to differences in fractional mass wei ghting factors used in the determ ination of skeletal averaged AFs. Effect of Update in Shallow Active Marrow Size This reference female model is the first to use the newly proposed definition of the endosteum (50 m from 10 m), which leads to a larger volume of shallow trabecular active marrow. This should cause a greater number of particles depositing dose into the TAM50, when compared to previous models. However, the mass of the target will be larger, as there are more voxels, thus the dose implications may be spread out for a similar SAF. It was shown that for bone sources irradiating the sh allow active marrow, the updated definition of the endosteum

PAGE 74

74 allows for a much higher dose to the OPC popula tion. Due to these discrepancies, the new definition of shallow active marrow should be used to predict marrow toxicity, as current models do not provide an accurate measure of dos e to the radiosensitive cell population. Conclusion A complete skeletal mass database has been gi ven for a reference female model to be used to provide a patient specific scalable model fo r accurate dose assessment in radiation protection and medical applications. No model for a female subject has previously been given, and this paper serves to provide a comp anion skeletal reference model to the adult male. The mass database comes from actual macrostructural pa tient data (ex-vivo CT) as well as accurate trabecular structure (microCT). Specific absorb ed fractions are provided using Paired Image Radiation Transport, where the macrostructure and microstructure of the actual, same patient is combined. It is hoped that this model will be used to better assess dose estimates of the female population.

PAGE 75

75 Table 3-1. Ex-vivo CT data for all skeletal sites.

PAGE 76

76 Table 3-2. MicroCT data for the skeletal sites with active marrow.

PAGE 77

77 Table 3-3. Tissue compositions (% by mass) and mass densities used in skeletal mass calculations. Table 3-4. Percentage of spongi osa of reference female compared to a database of 20 cadavers.

PAGE 78

78 Table 3-5. Spongiosa volume and percentage of total for UF reference male and female.

PAGE 79

79 Table 3-6. Averaged marro w volume fractions, bone volume fractions and shallow marrow volume fractions by skel etal site for UFRF.

PAGE 80

80Table 3-7. University of Florida Refere nce Female masses for the trabecular spongio sa, cortical bone, and mineral bone regions for all skeletal sites.

PAGE 81

81Table 3-8. University of Florida Reference Female active marrow masses.

PAGE 82

82Table 3-9. Shallow marrow masses for UFRF.

PAGE 83

83Table 3-10. S/V Ratios.

PAGE 84

84Table 3-11. Surface Areas for UFRF in both trabecular and cortical bone regions.

PAGE 85

85 Table 3-12. Weighting factors for marro w sources for each bone site for UFRF.

PAGE 86

86 Table 3-13. Weighting factors for bo ne surface and volume sources for UFRF.

PAGE 87

87 Figure 3-1. Top in-vivo s can of UF Reference Woman

PAGE 88

88 Figure 3-2. In-vivo contoured imag e of the ribs in transverse (top) coronal (middle) and sagittal (bottom) planes.

PAGE 89

89 Figure 3-3. In-vivo contoured image of the craniu m, showing the separation of the lobes: left parietal (orange), right pariet al (yellow), occipita l (green), facial bones (pink), cortical bone (blue) and other bones (red).

PAGE 90

90 Figure 3-4. In-vivo contoured image of the craniu m showing the separated l obes: left parietal (orange), right parietal (yellow), and frontal (green). Figure 3-5. Transverse (top) and coronal (bo ttom) CT images showing the abnormality of the right humerus. The cortical shell of th e shaft extends throug h to the proximal humeral head.

PAGE 91

91 Figure 3-6. Transverse (top) and Coronal (bottom) Contoured ex-vivo image of the femur, showing the separation of th e femoral head and neck Figure 3-7. Transverse (top) and sagital (bottom) contoured ex-vivo image of the lumbar vertebrae with cortical bone in blue and spongiosa in red.

PAGE 92

92 Figure 3-8. Active marrow distributi on as a function of age used as reference values in ICRP 70 (Cristy, 1981).

PAGE 93

93 Figure 3-9. Cellularity varies by bone si te and decreases with age (Custer, 1974) Figure 3-10. Current ICRP 70 reference cellulari ties based on bone site and age (Cristy, 1981).

PAGE 94

94 Figure 3-11. Microstructure of the frontal lobe (40% marrow). Figure 3-12. Microstructure of th e occipital lobe (11% marrow). Figure 3-13. Microstructure of the femoral neck (87% marrow).

PAGE 95

95 Figure 3-14. Microstructures of the femoral head (71% marrow). Figure 3-15. Microstructure of the sternum (99% marrow). Figure 3-16. Microstructure of the 3rd cervical vertebrae (76% marrow).

PAGE 96

96 Figure 3-17. Microstructure of the 5th lumbar vertebrae (91% marrow).

PAGE 97

97 Figure 3-18. Specific absorbed fraction for th e TAM irradiating the TAM for the lumbar vertebrae.

PAGE 98

98 Figure 3-19. Specific absorbed frac tion for the TAM irradiating the TAM50 for the lumbar vertebrae.

PAGE 99

99 Figure 3-20. Specific absorbed fraction for a ll sources irradiating the TAM for the lumbar vertebrae.

PAGE 100

100 Figure 3-21. Specific absorbed fracti on for all sources irradiating the TAM50 for the lumbar vertebrae.

PAGE 101

101 Figure 3-22. Specific absorbed fraction for the TAM irradiating the TAM for all bone sites. Figure 3-23. Specific absorbed frac tion for the TAM irradiating the TAM50 for all bone sites.

PAGE 102

102 Figure 3-24. Specific absorbed fraction for the CIM irradiating the CIM50 in the shafts of all long bones.

PAGE 103

103 Figure 3-25. Specific absorbed fraction for the CBSMC irradiating the CIM50 in the shafts of all long bones.

PAGE 104

104 Figure 3-26. Specific absorbed frac tion for the CBV irradiating the CIM50 in the shafts of all long bones.

PAGE 105

105 Figure 3-27. Skeletal averaged AF for marrow sources irradiating the TAM for both UFRF and Stabin and Siegel (2003).

PAGE 106

106 Figure 3-28. Skeletal averaged AF for bone trabeculae sources irradi ating the TAM for both UFRF and Stabin and Siegel (2003).

PAGE 107

107 Figure 3-29. Skeletal averaged AF fo r marrow sources irradiating the TAM50 for both UFRF and Stabin and Siegel (2003).

PAGE 108

108 Figure 3-30. Skeletal averaged AF for bone trabeculae sources irradiating the TAM50 for both UFRF and Stabin and Siegel (2003).

PAGE 109

109 CHAPTER 4 COMPARISON OF HUMAN STEM CELL MEASUREMENTS IN BIOPSIES VERSUS AUTOPSY SPECIMENS Introduction Bone marrow biopsies are widely used to a ssess the health of bone marrow and prediction of disease. Advances in the understanding of blood disorders and th eir ties to cellular elements in bone marrow have vastly increased the use of biopsies in the past decad e. Consequently, bone marrow biopsies are taken frequently in both norm al and abnormal patients as a part of routine diagnosis for conditions in hematology, intern al medicine, and oncology. Bone biopsies are valuable not only for diagnosis of cytopenia, my eloand lymphoproliferativ e disorders, but also for vast problems relating to immunology and hematopoietic regulator y mechanisms (Hoffman et al., 1991). Typical bone marrow biopsies are taken from eith er the anterior or posterior iliac crest, with core sizes ranging from 4-8 mm x 10-22 mm depending on the acquisition site and procedure followed. Hematopoietic tissue a nd trabecular bone account for 40% and 26% by volume of the biopsy, respectively, assuming 28% fatty tissue. The remaining bone marrow biopsy is osteoid, sinuses, and edema. Disease specific processing techniques are applied to the biopsy to quantify the degree of hematological di sorder through a cellula r count such as red blood cell quantity in the case of potential anemia or Reed-Ste rnberg cells in the case of Hodgkins lymphoma (McPhe rson and Pincus, 2007). In radioimmunotherapy, the dose limiting orga n is the hematopoietically active (red) bone marrow and estimates of the absorbed dos e require knowledge of the marrow health. Moreover, the red marrow is in reality a surr ogate for the underlying radiosensitive cell populations, namely that of the hematopoietic stem cells. Thus, a measure of marrow health is related to the quantity of hematopoietic stem cel ls. In particular, the concentration of CD34+

PAGE 110

110 stem and progenitor cells has been shown to be di rectly related to the pr ognosis of therapeutic outcome following marrow ablation (Sgouros, 1993). Recent studies have shown that a spatial gradient exists for this stem cell population in human bone marrow, with a higher concentration of cells localized near the trabecular bone and decreasing further into the marrow cavities (Watchman et al., 2007). This relationship is seemi ngly found to exist in both large field specimens (autopsies) as well as smaller fields (b iopsies). The spatial di stribution itself does not prove to have any particular bone site dependenc e, and thus any bone site should yield the same linearly decreasing trend (Bourke et al., in press). The first studies to directly measure the sp atial distribution of HS Cs to bone trabeculae were performed at the University of Florida by Watchman et al (2007). In that study, CD34 and CD31 immunohistochemical staining was performed on subseque nt slides from core biopsy samples of the iliac crest from disease-free nor mocellular human bone marrow. The data fully supported the hypothesis that CD34+ HSCs and their progenitors fo llow a spatial gradient with respect to bone trabecu lae within human bone marrow. Also blood vessels seem to follow the same gradient indicating a shared spatial niche with stem cells. Distance-dependent weighting factors are provided as an upda te to the MIRD schema. The most current work at the University of Florida by Bourke et al (2009) used autopsy methods rather than biopsies to assess spatial distributions in larger samples and for several skeletal sites because of poten tial field of view limits. CD34 and CD117 immunohistochemical staining was performed on au topsy sections from the iliac crest, lumbar vertebrae, and rib of disease free, normocellular patients. This data shows the same trend foun d in previous studies, that the CD34+ cells decrease as a function of marrow depth. Also, the method used to stain tissue was shown to alter the concentration of cells while preserving spatial gradient.

PAGE 111

111 Cellular measurements are often made in small biopsy samples, with hopes that the sample is representative of the entire marro w. Accuracy may be limited by the field of view in a biopsy, when compared to a full field autopsy specimen. Shown in Fi gure 4-1, a biopsy may only show the white portion of the marrow, leading to cell distances being meas ured to the nearest trabeculae in view for total dist ances of 331, 287, 404, and 365 pixe ls. If, however, the full field was seen and the pink region of marrow was included, cellular measurements would now correctly be 331, 287, 241, and 271 pixels. This brings to question the validity of using limited field biopsy specimens as a surrogate for the entire marrow cavity. Artificial cell-to-bone distances by trabeculae lying just outsi de the processed specimens field-of-view has been termed a potential edge effect, and is postulated as a potential limitation of the use of biopsies to a ssess marrow health (Watchman et al., 2007). This has implications not only for cellular distributions, but could lend to inaccuracies in any cell count expressed as a percentage of total marrow. Biopsies are wi dely used to assess bl ood disease, but their usefulness is solely dependent upon the accuracy of such a measurement technique. The objective of the present study is to directly quantify the differen ce in hematopoietic stem cells concentration and distribution due to areal effects. The objective is thus to determine whether hematopoietic cellular distributions found in biop sy specimens will yield the same result that would have been found had a larger tissue field-of-view been used. Furthermore, any directional dependence (anisotropy) in HSC concentration will be discussed through use of simulated biopsies taken in vertical or horizontal planes. Materials and Methods Artificial biopsy sections were taken from previously collected and stained autopsy specimens from the L1 vertebrae. The skeletal site of the autopsy specimen was chosen due to its planar symmetry. Unlike the Iliac cres t, autopsies from the L1 vertebrae are 5000m on end

PAGE 112

112 and allows for identical sized masks in both tw o dimensional planes (Figure 4-2). This is advantageous for assessing anisotro py within bone sites. Six artifi cial sections were taken, three in the horizontal and three in the vertical plane (Figure 4-3). Hori zontal and vertical planes were chosen to study any anisotropy or directional de pendence. Each simulated biopsy specimen was treated independently, i.e. each was looked at as if there was no other information other than what was in that view. Accord ingly, cellularity was measured for each biopsy section separately (Figure 4-4). Autopsy Collection and Preparation Previously acquired post-mortem bone samples from Bourke et al (2009), sectioned and fixed in paraffin from the L1 vertebrae of nine recently deceased patients were assembled. All samples were collected within 24 hours of death and were determ ined to be absent of marrow disease under a HIPAA-compliant and IRB-approved protocol. Au topsy blocks were decalcified and embedded in paraffin for preservation. Sequential sections were collected from each specimen at a thickness of 5 mm, and the most intact section was kept. Table 4-1 summarizes the age, gender, and cause of death for each subj ect in the study. These sections were then manually stained using mouse anti-CD34 (d ilution 1:25, QBEND10, DAKOCytomation; Carpinteria,CA). CD34 staining of endothelial cells served as an internal control. Digital Imaging and Processing After staining, the autopsy specimens were imaged at the Optical Microscopy (OM) facility of the University of Florida McKnight Brain Institute (MBI). Image resolution was verified via a caliper slide giving a calibration factor of 1.2 pixels/m m. Using these digital images, bone trabeculae were digitally segm ented using Adobe Photoshop. Following the acquisition of digital autopsy sp ecimens, masks the average width of a clinical biopsy were overlaid to achieve a simulated biopsy w ithin the full field autopsy specimen.

PAGE 113

113 Using 16 previously collected biopsy samples fr om Watchman et al (2007), the width of a typical clinical biopsy was found to be 1.24 mm with a standard devia tion of 0.46 mm, or 1489 550 pixels. This large variation in biopsy width is to be expe cted and based on the individual procedure used to collect biopsies. Three samp le width measurements were averaged per biopsy slide, for a total sample size of 48. Six s imulated biopsy masks were overlaid in nonoverlapping regions of each autops y slide, three in the horizontal plane and three in the vertical plane. In the horizonta l plane, sections were in the upper, middle, and lower regions, whereas sections were in the left, middle, and right regions in the vertical plane. An example of both the horizontal (yellow) and vertical (purple) middle simulated biopsy sections are shown in Figure 42. Measurement of Hematopoietic CD34+ Cells Cells were optically identified using an OM36LED contour microscope (Microscope Store, Rock Mount, VA) with a 20X objective as a compromise between field of view and image resolution. SPOT Advanced imaging software (Diagnostic Instruments In c, Sterling Heights, MI) was used to make distance measurements in pixel units. A CD34+ cell was found in the slide under the microscope, and the distance fr om that corresponding cel l to the nearest bone trabeculae was measured on the digital image. Pixel values were then transformed to micrometers using a calibration factor of 1.2 pi xels/mm. Histological distance measurements were determined for each of the six simulate d biopsy, and compared to the cell count and concentration in the total autopsy specimen. Al so, the average concentration of the horizontal masks and the average of the verti cal masks were compared to the total to explore anisotropy. Cells were not considered CD34+ unless they exhibited visible memb rane expression of the antigen, in addition to small, primitive morphol ogy consistent with a HSC. This is shown in Figure 4-5. When bone trabecul ae were obviously missing or were partially separated from the

PAGE 114

114 tissue during the sectioning process, distance measurements and trabeculae segmenting were made to the boundary of its original anatomical space. No boundary masks were used, and only those cells within each simulated biopsy were counted and compar ed to the total ce llular count in the autopsy as a whole. Sample measurements from cells to nearest bone trabeculae are found in Figure 4-1. Cellularity Measurement Regions of active bone marrow were determin ed via manual segmentation of adipocytes using Adobe Photoshop. For each simulated biopsy, a 500 mm x 500 mm location within each bone site, deemed visually to be representative of the entire marrow space was confirmed. Three sample locations were averaged, when possible. At the very least, two cellularity measurements were sampled for more accurate determination of fat percentage (Figure 4-4). The resulting color image was converted to grayscale, filter ed, and converted to a binary black and white image, and then pixel counted. The ratio of the final non-zero pixel count to the original image pixel count was used as the precise digital estimate of th e amount of total marrow area contributed by the adipocytes popu lation, and is termed the ma rrow cellularity factor (CF). Marrow Area Measurement In order to determine cellular concentration, the area of marro w must be known. First, bone trabeculae was segmented using Adobe Photos hop. Shown in Figure 4-6, each specimens digital image was converted to gr ayscale, filtered, and then conve rted to binary black and white image. In this image, white and black signify marrow regions and trabecular regions, respectively. Trabecular surfaces were then expanded in 50 m increments using an in-house edge dilation algorithm developed for use with Im ageJ (NIH, Bathesda, MD). Using this data, binned areal measurements were determined fo r each simulated biopsy through pixel counting of the merged dilated trabecular bone image with the total marrow image. These binned cell counts

PAGE 115

115 were normalized to the amount of marrow c ontributed by each region, yielding cells/mm2 in 50 m increments from the bone surface. For each simulated biopsy s, the average cellular concentration is simply the ratio of total num ber of CD34+ hematopoietic stem and progenitor cells (HSPC) to marrow area for that site: s HSPC s HSPC sA N (4-1) where N is the number of cells and A is the area. This was performed for each 50 m bin, substituting the cell count and subsequent area for that bin. For example, for the first 50 m bin would the number of cells in the first 50 m from all bone trabeculae, divided by the marrow area 50 m from all bone trabeculae in the simulated biopsy. Cellular concentration of he matopoietically active (red) bone marrow regions were obtained by multiplying total marrow area contribut ed by each region by its cellularity factor (CF) within the specimen: s HSPC s HSPC active sA CF N (4-2) Each specimens cellular concentration gradient was weighted by its fractional contribution (fs) to the total marrow area measured in that bone s ite across all specimens. This leads to two fractional contributions dependent upon the measurement made. Fo r the case of total simulated biopsies, fs was weighted to the total marrow of all speci mens. If horizontal and vertical biopsies are assessed separately, fs was weighted to the amount of marro w in either all horizontal or all vertical biopsies, respectively. The fractiona l weights for each case are shown in Table 4-2. Specimen-weighted mean cell concentrations and we ighted standard deviations of the mean cell concentration were also calcula ted and plotted at each distan ce interval, in each bone site. Specifically,

PAGE 116

116 x HSPC s s HSPCf (4-3) where the fractional weight and areal density are decribed above. The weighted standard devi ation is then given as: 12 x x HSPC x HSPC s s HSPCn fx (4-4) where n is the number of samples. Results The simulated biopsy cellular concentrations were compared to the entire autopsy from which they came. Cellularity, specimen fractiona l weight for total biopsies, total marrow area, active marrow area, and cell counts for each patient are given in Table 4-1. Specimen fractional weights are shown in Table 4-2. The mean ce llular concentration and standard error for both simulated biopsies and autopsies in the total and active marrow among all patients are provided in Table 4-3. Mean cellular concentration and standard error for verti cal and horizontal biopsies in both total and active marrow if given in Tabl e 4-4. Figure 4-7 show s the concentration of cells for the simulated biopsies versus autopsies. In this case, simulated biopsies are all six simulated biopsy cases separately weighted and added together to provide a comparison to the autopsy cell count. Finally, Fi gure 4-8 is a comparison between the cellular concentration in vertical biopsies, horizontal biopsies, and the autopsies. Discussion Full field Autopsy Measurements Mean average CD34+ cell concentration in the total marrow was found to be 12.55 + 2.84 cells per mm2 among all patients in the to tal vertebrae autopsy. In the active marrow, cellular concentration was 7.62 + 1.38 cells per mm2. The cellular concentra tion was best modeled by a

PAGE 117

117 linear curve fit, and wa s found to have an R2 of 0.848. Most HSCs were found between found between 0 and 250 m from the bone surfaces (80.4% ), and this region contained most of the active marrow area (71.3%). The cellular concentr ation was essentially equal to the mean up to 250 m, and diminished rapidly after this point with no cells found beyond 700 m. Total Simulated Biopsy Measurements Each of the six simulated biopsy measuremen ts were compared independently, and the mean cellular concentration in both total and active ma rrow was found to be 7.04 + 0.98 cells per mm2 and 10.97 + 1.58 cells per mm2 respectively. A linear fit provided an R2 of 0.939. Again, most cells were found in the fi rst 250 m from the bone surfaces (79.6%), comprising 70.4% of the active marrow area. HSCs dropped off quite rapidly after 300 m, with no cells past 700 m. Vertical Small Field Measurements CD34+ mean cell concentration in the ve rtical fields was found to be 7.30 + 1.22 cells per mm2 in the total marrow and 11.37 + 1.99 cells per mm2 in the active marrow. A linear fit was performed, with an R2 of 0.911. As in the previous cas e, most cells and active marrow were within the first 250 m, with 80.8% and 70.7% resp ectively. A drastic li near decrease was noted after 300 m, with no cells past 700 m. Horizontal Small Field Measurements Mean cell concentration of HSCs in the horizontal fields were 6.81 + 1.17 cells per mm2 and 10.45 + 1.88 cells per mm2 in the total and active marrow, respectively. A linear fit provided an R2 of 0.920. In this study, 78.3% of the cel ls were within the fi rst 250 m from bone trabecular surfaces, and this region accounted for 70% of the active marrow. The cellular concentration was found to be near the mean until 250 m, where the concentration of cells decreased until no cells were found as fa r as 700 m into the marrow cavities.

PAGE 118

118 Statistical Analysis A students t-test reveals that there is no st atistically significant difference between mean cell concentrations in the simulated biopsies an d the autopsies in the total marrow (t=1.21, p=.24) or in the active marrow (t=1.63, p=0.14) with a 95% confidence leve l. There is also no difference between the vertical biopsy and the au topsy in the total marrow (t=0.62, p=0.55) or active marrow (t=1.16, p=.29). Furthermore, there was no difference between the mean cell concentration in the horizontal biopsy and autopsy in the total ma rrow (t=1.58, p=0.06) or active marrow (t=2.07, p=0.07) with a 95% confidence level. Lastly, there was no statistically significant difference found between the vertical and horizontal bi opsies in the to tal (t=1.51, p=0.14) and active marrow (t=0.96, p=0.09). Field of View The possibility of artifacts when assessing th e spatial distribution of stem cells in bone marrow is greater at smaller fields of view. The less marrow viewed in the field of view, the greater the chance of inducing an edge effect, a nd artificially skewing measurements to longer distances. In this study, small field of views are the simulated biopsies, with large views of views comprising the full autopsy specimen. The mean spatial concentration of simulated biopsies was compared to the mean cellular conc entration of the autopsy from which it came to assess the degree of artifact induction. Comparing the average spatial concentration of each simulated biopsy to the autopsy from which it came yields the same linear decrease as de pth into the marrow cavity increases. In the active marrow, mean cellular concentration is so mewhat lower in small fields (10.97 cells per mm2) than in large fields (12.55 cells per mm2), but are within one sta ndard deviation from the mean (1.58 cells per mm2). The same trend in cellular co ncentration is noted in the active marrow, with small and large fi elds yielding 7. 62 cells per mm2 and 7.04 cells per mm2

PAGE 119

119 respectively. Again, both are with in the error bars (0.98 cells per mm2), deeming no difference between the two. Identical mean concentrations are not noted si nce the simulated biopsies did not cover the entire auto psy error and were instead measured to have the same average width of an actual patient biopsy. Although the simulated biopsy yiel ds a slightly smaller mean cellular concen tration when compared to the autopsy, there is no difference gi ven the error in counts. It would be expected that a smaller area would reveal lower cell counts, however, this data is re lative to the amount of area in the viewing area. Because of this normalization to area cellular concen tration is the same regardless of field of view. Furthermore, the same decreasing spatial gradient is noted in both cases. Thus, the small field simulated bi opsy can be used to predict the spatial concentration within the larger field of view. This is noted when accounting for both total and active marrow. This lends itself well to de termining the accuracy of using biopsies as a surrogate of the entire marrow cavity to asse ss marrow health in both normal and diseased patients. Anisotropy One concern in using biopsies to determin e marrow health is the proper location of insertion. To determine whether there is a ny directional dependence (anisotropy) in cellular concentration measurements, vertical and hor izontal simulated biopsies were compared independently. Vertical biopsies in the total marrow provided a mean cellular concentration of 7.30 cells per mm2, whereas vertical biopsies in the tota l marrow showed a mean of 6.45 cells per mm2. Both are within one standard devi ation of the mean: 1.22 cells per mm2 and 1.10 cells per mm2 in the vertical and horizontal biopsies, respectively. In th e active marrow, mean spatial concentration was 11.37 cells per mm2 in the vertical biopsies and 9.97 cells per mm2 in the

PAGE 120

120 horizontal biopsies. Again, these are within both vertical biopsy e rror (1.99 cells per mm2) and horizontal biopsy error (1.80 cells per mm2). Both the vertical and horizontal mean cellular concentrations matched the autopsy from which they came, and did not vary amongst each ot her. This was noted in both active and total marrow. Moreover, the same decrease in spa tial gradient as depth into the marrow cavity increases was seen. This proves that there is no directional dependence among specimens. It can be assumed that a bone marrow biopsy in one dire ction will yield a similar cellular concentration as any other direction and an accu rate assessment of marrow health. Dose Implications The inclusion of artifacts such as thos e caused by edge effect s would have great implications in proper assessmen t of dose to the red bone marrow. As proposed by Watchman et al (2007), the MIRD dose equa tion should be modified to include a dist ance-dependent weighting factor as follows: x s x x s sr RBM S A r RBM D ~ (4-5) where A is the cumulated activity with the source, S is the radionuclide S value for marrow depth x, and is a distance-dependent wei ghting factor which is the mean fractional mass of red bone marrow at marrow depth x. Had a biopsy not correctly measured the concentration and spatial gradient in the red bone marrow, the weighting factor would ha ve been incorrectly measured, and the dose to the red bone marro w would be incorrectly assessed. This can be significant for radioimmunotherapy, as low ener gy particles with short ranges now irradiate a large portion of the radiosensitive cell population, leading to increase marrow toxicity.

PAGE 121

121 Conclusion In this study, the mean cellular concentration and spatial gradient of HSCs were compared in simulated biopsies and the autopsies from which they came. Six small field simulated biopsies were superimposed on each full field autopsy, and CD34+ measurements were made. These were normalized to the area of either total or active marrow. The mean concentration of HSCs was found to be the same in both simulated biopsies and autopsies, signifying the accuracy in using small fields to measure cellular concentration as a surrogate for larger field of views. Furthe rmore, horizontal and vertical si mulated biopsies were compared independently to assess anisotropy in cellular m easurements. Again, no difference was found in both mean and spatial concentration and thus we conclude that th ere is no directional dependence. This study has shown that a biop sy should provide an accurate assessment of marrow health.

PAGE 122

122 Table 4-1. Pertinent values for each patient in the simulated biopsy study. Table 4-2. Specimen fractional weight in total simulated biopsies or ho rizontal and vertical biopsies.

PAGE 123

123 Table 4-3. Cellular concentration for simulate d biopsies versus autopsies in active and total marrow. Table 4-4. Cellular concentra tion for horizontal vers us vertical biopsies in active and total marrow.

PAGE 124

124 Figure 4-1. Limitations of cell count in biopsies: A biopsy ma y only show the white portion of the marrow, leading to cell distan ces being measured incorrectly. Figure 4-2. The lumbar vertebrae was chosen due to planar symmetry as compared to the illiac crest (Bourke et al., in press).

PAGE 125

125 Figure 4-3. Artificial biopsy sections, 3 in the horizonal plane (yellow) and 3 in the vertical plane (purple). Figure 4-4. Independent cellularity measurem ents were made for each biopsy (red squares). Figure 4-5. Stained CD34+ cel ls in human bone marrow (Bourke et al., in press).

PAGE 126

126 Figure 4-6. Image processing to determine 50m areal contours from surfaces of bone trabeculae (Watchman et al., 2007).

PAGE 127

127 Figure 4-7. Spatial gradient of HSCs in the ac tive marrow in simulated biopsies versus autopsies

PAGE 128

128 Figure 4-8. Spatial gradient of HSCs in the active marrow for vertical and horizontal biopsies versus autopsies.

PAGE 129

129 CHAPTER 5 EFFECT OF CHEMOTHERAPY ON THE SPAT IAL DISTRIBTUION OF STEM CELLS IN HUMAN BONE MARROW Introduction To make predictions of radiation dose in pa tients undergoing targeted radionuclide therapy for cancer, an accurate model of skeletal tissues is necessary. Concerni ng these tissues, the doselimiting factor in these therap ies is toxicity of the hema topoietically active bone marrow (Sgouros, 1993; Siegel et al., 1990). In addition to acute effect s, one must also be concerned with long-term stochastic effects such as radiation-induced leukemia. Cells of particular interest for both toxicity and cancer risk are the hematopoi etic stem cells (HSC), found within the active marrow regions of the skeleton (Lim et al., 1997). At present, cellu lar-level dosimetry models are complex, and thus we cannot model individu al stem cells in an anatomic model of the patient. As a result, one revert s to looking at larger tissue regi ons where these cell populations may reside. Regions of bone marrow exclusive of marrow adipocytes thus serve as surrogate target tissue for underlying radiosensitive cell populations. The current assumption made in internal radi ation dosimetry is that the HSC population is uniformly distributed throughout the marrow cav ities of all skeletal sites (ICRP, 1977). Absorbed dose is then averaged through all re gions of bone marrow accordingly. Also, the number of target cells is considered to be proportional to the volume of active bone marrow due to a lack of measured data. However, recent st udies have shown that a sp atial gradient of HSCs may exist in human bone marrow (Watchman et al., 2007). It is proposed that the hematopoietic stem and progenitor cells are preferentially locat ed closer to the bone trabecular surfaces, and decrease in concentration further into the bone marrow cavities. This may cause significant errors in calculating the marrow dose estimate s using the current dose-response models for

PAGE 130

130 internal radionuclide therapy sinc e the tissue surrogate (the en tire active bone marrow) no longer matches the spatial location of the radiosensitive cells of interest. In radioimmunotherapy (RIT), radiolabeled antibodies guide the radioactive element to a specific site of interest. Fo r treatment of Non-Hodgkins lym phoma (NHL), beta sources are used to deliver radiation to abnormal cells in a ffected body sites. In th e process of targeting tumor cells, radiolabeled antibodies come in contact with marrow elem ents causing radiation damage to diseased and normal cells. All radi olabeled antibodies cause some amount of bone marrow suppression, however the degree of marrow t oxicity is difficult to ascertain using current methodology. The inability to predict and modify the degree of marrow toxicity adds significant clinical complexity to utiliz ing these agents. As a genera l rule however higher marrow involvement with lymphoma, greatly affects ra diation dosimetry levels (Martinez-Jaramillo et al., 2004). As the antibody travels to the tumor, a higher concentration will stay in the marrow and the crossfire from cancerous cells to normal cells will be high. Severe myelotoxicity cases as a result of radioimmunotherapy ca n affect much of the hematopoie tic stem cell population. A bone marrow transplant may be required to rest ore the depleted population in extreme cases. Systemic therapy, consisting of traditional cytotoxic agents and more recently immunotherapeutic agents such as Rituximab, is the corner stone of front line therapy for nonHodgkins lymphoma. External beam radiation therapy is sometime s used in select histological variants and clinical presentations with ra dioimmunotherapy commonly used in the setting of relapsed or refractory disease. The most comm only used chemotherapeutic regimen for NHL is R-CHOP consisting of Cyclophospha mide, Doxorubicin, Vincristin and Prednisone, along with the anti CD 20 monoclonal antibody, Rituximab. Rituximab targets B lymphocytes, including abnormal lymphocytes, through the CD20 antigen that is present on their ce ll surface. While the

PAGE 131

131 traditional cytotoxic agents target all rapi dly growing cells including the cancerous cells, Rituximab more specifically targ ets cells expressing CD20 such as the NHL cells. Some degree of toxicity to normal marrow cells including stem cells is anticipated in all systemic therapies including cytotoxic agen ts, therapeutic antibodies and RIT. To provide a more accurate marrow radiati on dose assessment, the skeletal dosimetry model must also be patient-speci fic. That is, it should be de signed to match as closely as possible to the patient under going treatment. Patients with relapsed NHL, depending on the degree of bone marrow disease involvement and pr evious therapy, may ha ve variable amounts of marrow reserve including numbers of HSC. Studi es have shown hematopoietic alterations in patients with diffuse large-cell lymphoma, with as much as a 35% reduction in progenitor cell numbers (Huerta-Zepeda et al., 2000; Chavez-Gonzalez et al., 2004). Current radioimmunotherapy procedures used for the trea tment of lymphoma may result in a large dose to the bone marrow, and will result in increase d marrow toxicity if the stem cell population is lower than normal. Additionally, the ability to accurately esti mate marrow toxicity from RIT will improve our ability to include these agents in clinical protocols utilizing cytotoxic agents also. There has been much debate over whether RIT procedures should be used in conjunction with chemotherapy, as well as their use as frontline treatment (Wiseman et al., 2001; Kaminski et al., 2005). Compared to chemotherapy, RIT has the ability to i nduce apoptosis through radiation damage, a strategy that may prove more effective than current a pproaches to targeting the CD20 antigen present on lymphocytes. Unfo rtunately, current intern al dosimetry models do not account for any difference in stem cell population due to hema topoietic alterations known to exist in diseased patients during treatment, increasing the difficulty of predicting marrow

PAGE 132

132 toxicity. While it is known that HSC will be disr upted to some degree in all therapeutic regimes, finding the optimal timing based on marrow health is key to successful in corporation of these agents treatment regimens. Thus, not only is it important to accurately m odel the target tissues of interest in a normal patient, it is important to do so for differing levels of marrow health and among treatment regimes. This study serves to better define the HSC population in NHL patients at diagnosis and after standard chemot herapy and will help establish a more accurate assessment of dose estimates for patients who are candidates for RIT. Materials and Methods The overall concentration and spatial distribution of CD34+ /CD31hematopoietic stem and progenitor cells from preand post-chemoth erapy lymphoma patients was compared to that seen in non-NHL patients. Patients were scre ened and confirmed to have had CHOP-R or similar treatments, have low or no marrow involve ment at diagnosis and in follow up. Twelve patient biopsies pairs both pre-chemotherapy and post-chemothe rapy, were collected, stained for CD34/CD31, mapped and compared to prev iously acquired measur ements provided by Watchman et al (2007). Pathol ogy reports of patients in that study included positive screens for Hodgkins lymphoma, HIV+, semi noma, CNS lymphoma, ITP, and abdominal mass. Those slides were cleared to have no marrow i nvolvement and were all negative for NHL. Patient Selection and Slide Preparation Twelve previously treated Bcell non-Hodgkins lymphoma patie nts were identified from the University of Florida Hematological pat hology archives under an approved institutional review board protocol. Eight females and four males were enrolled in the study, with ages ranging from 14 to 75 years and confirmed to m eet the requirements stated above. Two bone marrow samples, one pre-chemotherapy and one pos t-chemotherapy, from samples collected as part of their routine were masked of identity a nd evaluated. All patients were screened for low

PAGE 133

133 marrow involvement, similar chemotherapy regimes, disease status and de gree of hematological toxicity after therapy. Most pa tients received CHOP or CHOP-R or other similar regimens as specified in Table 5-1. Post -chemotherapy biopsies were co llected when the patient was determined to be in complete remission, with the timing since pre-chemotherapy biopsy noted. Thus, a total of 24 biopsies we re collected: 12 pre-chemotherapy biopsy samples and 12 postchemotherapy samples (two per patient). Pa tient data is shown in Table 5-1. Paraffin blocks were stained for CD34/CD31 using methodology defined in Watchman, et al (2007). As noted, 4mm sections were cut and placed on slides to be dried and freed of all paraffin. Co-staining was preformed, using CD34 antibody (QBend/10 clone prediluted from Ventana) and CD31 antibody (Dako Corp) at a diluti on of 1:20. Antigen presence was visually determined using detection kits. Slides were then dehydrated and pe rmanently mounted for further use. Digital Imaging and Processing After CD34/CD31 staining, slides were imaged at the University of Florida Molecular Pathology Core using an Aperio ScanScope CS system to provide a digital image of each specimen. Slides were submitted with all iden tifying marks masked, and viewed via Spectrum software. Using Aperio ImageScope, a region of interest was defi ned and resolution was verified. Digital images were then stored lo cally and adjusted based on memory and field of view limitations. A sample co-stained biopsy is shown in Figure 5-1. Measurement of Hematopoietic CD34+ Cells Cells were optically identified using an OM36 LED contour microscope (Microscope Store, Rock Mount, VA) with a 20x objective as a comp romise between field of view and image resolution. SPOT Advanced imaging software (Diagnostic Instruments In c, Sterling Heights, MI) was used to make distance measurements in pixel units. A CD34+ /CD31(red) cell was

PAGE 134

134 found in the slide under the microscope, and th e distance from that corresponding cell to the nearest bone trabeculae was measured on the digital image. Pixel values were then transformed to micrometers using a calibration factor of 1.2 pixels/mm. Histological distance measurements were determined for each patient both before an d after chemotherapy and compared to the mean cell count and concentration gradient in previously ac quired non-NHL marrow. Cellularity Measurement Regions of active bone marrow were determin ed via manual segmentation of adipocytes using Adobe Photoshop. For each biopsy, a 500 mm x 500 mm location with in each bone site deemed visually to be represen tative of the entire marrow spa ce was confirmed. Three sample locations were averaged for each biopsy. The re sulting color image was converted to grayscale, filtered, converted to a binary black and white im age, and then pixel counted. The ratio of the final non-zero pixel count to th e original image pixel count wa s used as the precise digital estimate of the amount of total marrow area cont ributed by the adipocyt es population, and is termed the marrow cellularity factor (CF). Marrow Area Measurement In order to determine cellular concentration, the area of marro w must be known. First, bone trabeculae was segmented using Adobe Photos hop. Shown in Figure 5-2, each specimens digital image was converted to gray scale, filtered, and then convert ed to a binary black and white image. In this image, white and black signify marrow regions and trabecular regions, respectively. Trabecular surfaces were then expanded in 50 mm increments using an in-house edge dilation algorithm developed for use with Im ageJ (NIH, Bethesda, MD). Using this data, binned areal measurements were determined fo r each simulated biopsy through pixel counting of the merged dilated trabecular bone image with the total marrow image. These binned cell counts were normalized to the amount of marrow c ontributed by each region, yielding cells/mm2 in 50

PAGE 135

135 m increments from the bone surface. For each biopsy s, the average cellular concentration is simply the ratio of total number of CD34+ he matopoietic stem and pr ogenitor cells (HSPC) to marrow area for that site: s HSPC s HSPC sA N (5-1) where N is the number of cells and A is the area. This was performed for each 50m bin, substituting the cell count and subsequent area for that bin. For example, for the first 50m bin would the number of cells in the first 50m from all bone trabeculae, di vided by the marrow area 50m from all bone trabeculae in the simulated biopsy. Cellular concentration of he matopoietically active (red) bone marrow regions were obtained by multiplying total marrow area contribut ed by each region by its cellularity factor (CF) within the specimen: s HSPC s HSPC active sA CF N (5-2) Each specimens cellular concentration gradient was weighted by its fractional contribution (fs) to the total marrow area measured in that bone site across all specimens. Specimen-weighted mean cell concentrations and weighted standard deviations of the mean cell concentration were also calculated and plotted at each distance interval, in each bone site. Specifically, x HSPC s s HSPCf (5-3) The weighted standard devi ation is then given as: 12 x x HSPC x HSPC s s HSPCn fx (5-4) where n is the number of samples.

PAGE 136

136 Results Table 5-1 lists the cellularity, specimen frac tional weight, total and active marrow areas, and total cell count for each patient both pre and post-chemotherapy. Mean cellular concentrations are given in Table 5-2, along data used in the statis tical analysis. Figure 5-3 plots the spatial gradient of HSCs to compare chem otherapy regime and dise ase states to non-NHL marrow provided in Watchman et al. Figur e 5-4 shows the percent change in HSPC concentration depending on the time between biopsies, and Fi gure 5-5 shows the cellularity changes for each patient. Mean Cellular Concentration In the pre-chemotherapy total marrow, HSPC concentration ranged from 0.46 cells per mm2 in patient 2 to 16.05 cells per mm2 in patient 11. Post-chemotherapy showed similar variation, with 1.25 cells per mm2 in patient 4 and 13.76 cells per mm2 in patient 12. In the active marrow, pre-chemotherapy HSPC con centration ranged from 0.99 cells per mm2 in patient 2 to 41.47 cells per mm2 in patient 11. Cellular concentrat ions in the active marrow for postchemotherapy biopsies range d from 1.48 cells per mm2 in patient 4 and 26.61 cells per mm2 in patient 12. The mean concentration of HSCs in the total marrow among all 12 patients was found to be 5.08 1.13 cells per mm2 pre-chemotherapy and 5.72 0.93 cells per mm2 postchemotherapy. Non-NHL marrow using similar s lide preparation and staining methods reports 9.03 1.47 cells per mm2. In the active marrow, pre-chemotherapy specimens yield 10.18 2.81 cells per mm2 and post-chemotherapy specimens yield 11.37 2.28 cells per mm2. There is no reported cell concentration in ac tive marrow for the non-NHL marrow.

PAGE 137

137 Statistical Analysis A students t-test reveals that there is a statistically significant difference between both non-NHL marrow and pre-chemotherapy marrow (t =7.38, p=0) and post-chemotherapy marrow (t=6.592, p=0) with a 95% conf idence level. However, ther e is no statistically significant difference between pre and post-chemothera py marrow for both total marrow (t=1.515, p=0.1372) and active marrow (t=0.1342, p=0.2686). Spatial Gradient As shown in Figure 5-3, the CD34+/CD31cells decreased in concentration as depth into the marrow cavity increased. Specifically, the fi rst 50 m of diseased marrow comprised 17% of the total bone marrow, and this region c ontained 42.6% of the HSC population in prechemotherapy patients and 46.6% of the HSC populat ion in post-chemotherapy patients. Nearly all cells (88.4% pre-chemotherapy and 87.08% pos t-chemotherapy) were found within the first 200 m of total marrow. No CD34+/CD31cel ls were found beyond 600 m of bone trabeculae in diseased marrow. Previously documented non-NHL marrow was best fit to a linearly decreasing trend (R2=0.935). Both pre and post-chemothe rapy data were best fit to an exponentially decreasing model (R2=0.988 and R2=0.990, respectively). Discussion The data presented in Figure 5-3 confirms prev ious findings of a spa tial gradient in human bone marrow for NHL patients. In both pre and post-chemotherapy lymphoma patients, a higher concentration of cells is seen close to the bone trabeculae, whic h decreases with depth into the marrow cavities. Thus, a larger proportion of ce lls are prudentially located in the first 50 m from bone surfaces, and a more specialized surrog ate for marrow toxicity would be to assume a spatial gradient instead of a single value. Fu rthermore, red bone marrow dose should not be simply averaged over marrow cavities as is curren tly assumed, and should instead be weighted

PAGE 138

138 by the fractional weight of HSPC present as noted by Watchman et al (2007). This would provide a more accurate prediction of marrow toxicity than is currently performed. Comparison to non-NHL Marrow The mean cellular concentration of the 12 pr e-chemotherapy lymphoma patients in this study was found to be less than the nonNHL marrow (5.08 versus 9.03 cells per mm2, respectively). This result was statistically si gnificant with 95% confidence, signifying a real decrease in the average HSC population in NHL patients when compared to non-NHL. The same spatial gradient was seen in nonNHL, however more pronounced and following an exponential instead of linear fit. It seems there may be a highe r population of stem cells in postchemotherapy diseased marrow within the firs t 50m from bone surface s when compared to non-NHL, although the reason for this is not fully unders tood. It is specula ted that this may be HSC translocation and repopulati on of the bone marrow niche dur ing hematopoietic recovery. Possible decreases in progenitor cell count s due to malignant lymphoma has been confirmed in studies as far back as 2000 (Huerta-Zepeda et al., 2000). In that study, the composition and function of the hematopoietic mi croenvironment was fo und to be compromised in the presence of diffuse large-cell lymphoma. Specifically, a 35% redu ction in progenitor cell numbers was found. In the present study, a 44% reduction in hematopoietic stem and progenitor cells is noted. While the presence of deficiencies in the hematopoietic system is noted in lymphoma patients, the reason for th is is still not fully understood. Further studies were made to determine wh ether the alterations in the hematopoietic microenvironment were due to an intrinsic defect in the progenitor cell compartment of lymphoma patients, or to an altered microe nvironment altogether. It was found that the expansion ability of the hematopoi etic progenitor cells to produ ce new cells capable of forming hematopoietic colonies was limited in lymphoma patients (Martinez-Jaramillo et al., 2004).

PAGE 139

139 Interestingly, most patients showed normal he matological parameters even though they had reduced progenitor cell numbers. It is speculated that while the expansion process is deficient, proliferation is not and normal numbers of ma ture blood cells can be maintained. This may suggest an intrinsic deficiency in the hematopoietic stem cells themselves in lymphoma or possibly the presence of a pa raneoplastic mechanism causing depression of HSC during the development of lymphoma. Moreover, recent st udies have also indicated that molecular alterations are present in hematopoietic precursers, mature cells, as well as primitive HSCs, some of which may be morphologically recognizable (Madrigal-Velazquez et al., 2006; ChavezGonzalez et al., 2004). Still other studies have shown a possible microenvironment defect as well as the well established inherent flaw in expansi on potential (Soligo et al., 1998). Effect of Chemotherapy Chemotherapy was found to have no affect on the mean concentration of HSCs, and there was no statistical difference when comparing pre and post-chemothera py biopsies among the 12 patients. Thus, it is believed that chemothera py regime did not affect the stem cell population when averaged over all patients. Previous st udies by Huerta et al (2000) also found that hematopoietic alterations were st ill present in patients at co mplete clinical remission after chemotherapy. It has been shown that a la rge heterogeneity exists in the response of progenitor cell growth after chemotherapy. In studies by Mar tinez-Jaramillo et al (2004), of 24 patients, 8 patients had reduced progenitor cell numbers, 11 showed an increase in progenitor cells, and 5 showed no difference. There was no correla tion found, however, between the change in progenitor cell number and clinical outcome. In the present study, a decrease in the HSC population post-chemotherapy occurred only in patient s 4, 6, 9, and 11. All 8 other patients had higher HSC counts after chemotherapy, thus a simila r heterogeneity was found. It has also been

PAGE 140

140 noted that after 18 months of chemotherapy and clinical remission, normal progenitor cell proliferation and expansion is expected lead ing to an indication of normal hematopoietic function (Yao et al., 2000). Timing Effect The percent change in average areal concentr ation versus the timing between pre and postchemotherapy patients is shown in Figure 5-4. Po sitive values indicate a higher number of cells post-chemotherapy. It appears that there is a gr eater change in the number of stem cells per area when the time between both pre and post chemothera py is small. There was a decrease in HSCs post-chemotherapy in patients 4, 6, 9, and 11. An increase in HSCs post-chemotherapy was seen for Patients 2, 3, 5, 7, 8, 10 and 12. Howeve r, no consistent trend has been found to suggest a reason for such changes in HSC population. Si milar heterogeneity post -chemotherapy has been previously found, and marrow only seems to retu rn to normal after 18 months of clinical remission (Martinez-Jaramillo et al., 2004; Yao et al., 2000). Effect on Cellularity Cellularity is used to determine the amount of fat in the marrow, where 100% cellularity indicates purely active marrow and 0% cellularity indicates entirely fatty marrow. Bone marrow cellularity has been noted to increase as a result of damage to the hematopoietic microenvironment (Jensen et al., 1990; Orazi et al., 1992). Thus, patients with lymphoma seem to have a higher fat content (low er cellularity) than that found in normal marrow at diagnosis, which increases with treatment. However, studies have shown that this decrease in fat content and subsequent increase in hemat opoietic tissue subsides once the patient is in remission, and a higher cellularity indicates a normal hema topoietic microenvironment (Moulopoulos and Dimopoulos, 1997).

PAGE 141

141 In the present study, cellularity was found to be lower than the normal value of 48% (Cristy, 1981) for 6 of 12 patients pre-chemothera py. Fat fraction ranged from 30.6% for patient 7 to 80.4% for patient 8. As expected, ce llularity increased or stayed the same postchemotherapy for 10 of 12 patients, as shown in Figure 5-5. Intere stingly, the two patients who did not have an increase in hematopoietic tissu e post-chemotherapy were patients 3 and 8 whom both had experimental (non CHOP) treatments. Dose Implications Radioimmunotherapy guarantees some amount of marrow toxicity since there is no specific marker to target only di seased cells. If the radiopharm aceutical is uniformly distributed throughout the marrow, absorbed do se would accurately be meas ured by averaging the dose over marrow cavities, as all radiosensitive cells would receive the same radiation dose. This is what is implicitly assumed in current dosimetry met hods to predict marrow toxicity. However, radiopharmaceuticals that localize on bone su rfaces (non-uniform distri butions) would result in higher marrow toxicity, as this is the preferenti al location of the radiosensitive cell population. In this case, absorbed dose s hould not simply be averaged ov er the marrow cavities, but should instead to weighted by the amount of cells at individual locations throughout bone marrow, as proposed by Watchman et al (2007). Lymphoma patients have shown to have less radiosensitive cells in the red bone marrow overall when compared to non-NHL marrow, with no changes in HSC population due to chemotherapy. This signifies that (1) marrow t oxicity may be inherently higher in NHL patients simply due to hematopoietic alterations, and (2) chemotherapy does not seem to affect the hematopoietic environment negatively and does not seem to result in marrow toxicity as a function of HSC death. For radioimmunotherapy, th e timing or nature of previous chemotherapy regimens appears to be less important to numbers of HSC although the numbers of more

PAGE 142

142 differentiated progenitors were not evaluated in our study. However, care should be taken in general when predicting absorbed dose in lymphoma patients, as current methods do not assume preferential location or a lower concentra tion of the radiosensitive cell population. Conclusion This paper serves to determine the mean concentration and spatial gradient in NHL patients to provide a better pred iction of marrow toxicity. Th e effect of chemotherapy on the bone marrow was also explored due to its co mmon use with radioimmunotherapy. Twelve patients with diffuse, non-Hodgki ns lymphoma undergoing chem otherapy were studied both before and after treatment. It was found that th e mean concentration of HSCs was lower in NHL patients when compared to marrow from patients without NHL, and did follow a spatial gradient whereby cell concentration decreased with de pth into marrow cavities. Chemotherapy was found to have no negative effect on the HSC popul ation. Thus, current procedures used to predict marrow toxicity should be updated to include a spatial grad ient and inherently lower HSC concentration. It is hoped that this will help better estimate marrow toxi city in diseased marrow

PAGE 143

143Table 5-1. Pertinent data for all 12 patients, in both pre and post-chemotherapy biopsies.

PAGE 144

144 Table 5-2. Mean cellular con centration and error fo r both normocellular a nd diseased marrow.

PAGE 145

145 Figure 5-1. Co-stained biopsy showing CD34+/CD31(red) hema topoietic stem and progenitor cells and CD34+/CD31+ (brown) endothelial tissue.

PAGE 146

146 Figure 5-2. Binary digital images of total ma rrow (white) and trabeculae (black) are combined and trabeculae are expanded to determine area of marrow in 50 m bins (Watchman et al., 2007).

PAGE 147

147 Figure 5-3. CD34+ cell concentration as a function of depth into marrow cavities.

PAGE 148

148 Figure 5-4. Change in HSC concentration as a function of time between biopsies.

PAGE 149

149 Figure 5-5. Cellularity changes before and after chemotherapy.

PAGE 150

150 CHAPTER 6 CONCLUSIONS AND FUTURE WORK Conclusions To predict the risk of bone can cer, an accurate assessment of sk eletal tissues is necessary. This requires both a patient specific model of the active marrow, as well as proper modeling of the location and concentration of radiosensitive ce lls. The individual patie nts skeletal structure must be known on a macrostructura l scale (correct skeletal si ze) and a microstructural size (trabecular structure). Furthermore, the os teoprogenitor cell and he matopoietic stem cell populations must be accuratel y known for realistic absorb ed dose estimates. In radioimmunotherapy, little work has been done to determine these factors. Also, there are no models to date that assess the differences in gender for patient specific assessments of radiation absorbed dose. Current ICRP models used to predict radia tion dose and toxicity rely on a Reference Man model that is outdated and unsalab le. Skeletal factors such as marrow masses and cellularities come from several independent sources dating back to 1926 (Custer, 1974; Trotter and Hixon, 1974; Mechanik, 1926). Moreover, absorbed fr actions are based on 2D optical scanning methods of only seven bone sites (Whitwell, 197 3). Perhaps the most striking problem is the inability of scaling to tailor abso rbed fractions and S values to specific patients. Recent work has also shown that ICRP tissue definitions for th e endosteum may be incorrect, and may require increasing the size of the radiosensiti ve region from 10 m to 50 m (Bolch et al., 2007). Furthermore, studies have identifi ed a spatial gradient in the he matopoietic stem cell leading to inaccuracies in current dose estimates which re ly on methodologies that average absorbed dose over marrow cavities (Watchman et al., 2007). These methods also assume normal bone marrow

PAGE 151

151 in terms of the concentration and location of stem cells, whic h is unrealistic of patients undergoing radioimmunotherapy. Chapter 3 provides a complete skeletal mass da tabase for a reference female model to be used to provide a patient specific scalable model for accurate dose assessment in radiation protection and medical app lications. No model for a female s ubject has previous ly been given, and this paper serves to provide a companion skeletal reference model to the adult male. The mass database comes from actual macrostructural patient data (ex-vivo CT) as well as accurate trabecular structure (microCT). Specific absorb ed fractions are provided using Paired Image Radiation Transport, where the macrostructure and microstructure of the actual, same patient is combined (Shah et al., 2005b). Hopefully, this model will be used to better assess dose estimates of the female population. Biopsies are widely used to assess blood di sease, and their usef ulness for predicting marrow health was explored in Chapter 4. In that study, it was determined whether hematopoietic cellular distributions found in biop sy specimens would yield the same result that would have been found had a larger field of view been used. Also possible directional dependence (anisotropy) in the hematopoietic stem cell concentration was discussed. The mean concentration of HSCs was found to be the same in both simulated biopsies and autopsies, signifying the accuracy in using small fields to measure cellular concentration as a surrogate for larger field of views. Furthe rmore, no anisotropy was found, signifying that a biopsy should provide an accurate asse ssment of marrow health. Chapter 5 served to determine the mean con centration and spatial gradient in lymphoma patients to provide a better pred iction of marrow toxicity. Th e effect of chemotherapy on the bone marrow was also explored due to its comm on use with radioimmunoth erapy. It was found

PAGE 152

152 that the mean concentration of HSCs was lo wer in lymphoma patients when compared to normocellular marrow, and did follow a spatial gr adient whereby cell concentration decreased with depth into marrow cavitie s. Chemotherapy was found to have no negative effect on the HSC population. Thus, current pr ocedures used to predict marro w toxicity should be updated to include a spatial gradient and inherently lower HS C concentration. It is hoped that this would help better estimate marrow t oxicity in diseased marrow. Future Work This dissertation enhances work already performed in the ALRADS group at the University of Florida, and allows room for fu rther applications. Specifically, improvements in the skeletal database can be made by adding more reference in dividuals of varying skeletal health through use of PIRT fo r 3D radiation transport. Al so, reference S values can be determined for the reference female model for several radionuclides of interest in current therapeutic regimes. Then, clin ical applications through scalab ility are achievable. Lastly, cellular measurements can be improved for faster and more realistic determinations of concentrations and gradients. Improvements in Skeletal Database The current skeletal database of UF refere nce individuals now consists of a male and female cancer patient, but needs to be expanded to include more individuals of varying skeletal health. The trabecular microstr ucture is known to change with age, thus to accurately predict absorbed dose for an individual, age related changes should availa ble in our database (Atkinson, 1965). Perhaps more prevalent is the known changes in trabecul ae with osteoporosis, signifying the need for individuals with varying skeletal health (Berne 1993). Studies have also shown changes in bone mineral density with menstruation status, with a markedly larger population of women with osteoporosis post-menopause (Yao et al., 2001). Thus, it is proposed that future

PAGE 153

153 reference individuals for women are assembled ba sed on menopausal status instead of age. Building a library of trab ecular microstructures and skeletal macr ostructures, one will be able to mix and match images for more patient specific scaling. Work is cu rrently underway in the ALRADS group for younger reference subjects incl uding pediatric patients Furthermore, new methods of modeling such as hybrid phantoms that combine both the realistic an atomy of voxel phantoms and flexibility of sty lized phantoms are realized. Reference S Values for UFRF Current methods of assessing absorbed dose re quire knowledge of the cumulated activity and radionuclide S value through the MIRD schema. The S va lue is the dose received by they target organ per disintegration in the source, and depends on the absorbed fraction, mass of the target tissue, and the mean energy emitted for th e radionuclide used in th erapy. These values should be assembled for the reference female mo del for several radionuclides used in current therapeutic regimes. Possible radionuclides to be used should include P-33, Sm-153, Sr-89, P32, Y-90, and I-131. After averaging over all skel etal sites, S values should be compared to those currently in use. Improvements in Cellular Measurements To alleviate the time to count CD34+ cells, fl uorescent staining should be used to eliminate use of visually defining positive cells in the micr oscope. Using this method, several stains can be applied at once to decrease the chance of falsel y selecting a co-stained (brown) cell as a stem cell (red). With sufficient code writing, this will also allow for automation in measurements from cells to the nearest bone tr abeculae. Most notably, work should be performed to develop 3D models of bone marrow to re duce errors induced by edge eff ects. The use of fiducially markers, novel imaging techniques, and software development should be e xplored to co-register the z plane in biopsies. Moreover, more patient s should be enrolled for both normal and diseased

PAGE 154

154 patients to increase relia bility and allow for age and gender assessments of the hematopoietic stem cell lines. Variation in spatial gradient with bone si te is recognized, but further investigations would more conclusively evaluate any bone site dependence (Bourke et al., in press). Lastly, the osteoprogenito r cell population should be studied to better define the region of the endosteum for better pr edictions of cancer risk. Scalability and Clinical Applications The absorbed fractions, masses, and S values from this work lends itself well to better estimates of absorbed dose in clinical applications, but only if the models are conveniently scalable. Studies have been ma de to predict and scale spongio sa volumes and masses and using statistical methods (Brindle et al., 2006b; Brindle et al., 2006c; Pichardo et al., 2007). However, these methods are not feasible on a large scale and software development is necessary. Using image based software, a quick pelvic CT scan of a radionuclide therapy patient could estimate patient specific S values almost immediately. Furthermore, image based measurements of marrow cellularity will allow for active marrow to be known in that radioimmunotherapy patient for a better prediction of marrow toxicity.

PAGE 155

155 APPENDIX A TABLES OF SPECIFIC ABSORBED FRACTION This appendix tabulates values of Specific Absorbed Fraction (SAF) for UFRF for any combination of source (TAM, TBS, TBV, TAM50, TIM) and target (TAM, TAM50) regions in the entire skeleton containing active marrow for a range of cellularities (10% to 100%). Absorbed Fraction (AF) values were normalized to the mass of the target (and cellularity) to report SAF. For the TAM irradiating the TAM, marrow cellularity is varied. For all sources irradiating the TAM or TAM50, ICRP reference cellularity is used. Table A-1. UFRF absorbed fraction for all sour ces irradiating the active marrow in the cranium. Table A-2. UFRF absorbed fraction for all sour ces irradiating the shallow active marrow in the cranium.

PAGE 156

156 Table A-3. UFRF absorbed frac tion for all sources irradiating th e active marrow in the clavicles. Table A-4. UFRF absorbed fraction for all sour ces irradiating the shallow active marrow in the clavicles.

PAGE 157

157 Table A-5. UFRF absorbed fraction for all sources irradiating the active marrow in the mandible. Table A-6. UFRF absorbed fraction for all sour ces irradiating the shallow active marrow in the mandible.

PAGE 158

158 Table A-7. UFRF absorbed fraction for all sour ces irradiating the active marrow in the scapulae. Table A-8. UFRF absorbed fraction for all sour ces irradiating the shallow active marrow in the scapulae.

PAGE 159

159 Table A-9. UFRF absorbed fraction for all sour ces irradiating the active marrow in the sternum. Table A-10. UFRF absorbed fraction for all sour ces irradiating the shallow active marrow in the sternum.

PAGE 160

160 Table A-11. UFRF absorbed fraction for all so urces irradiating the active marrow in the ribs. Table A-12. UFRF absorbed fraction for all sour ces irradiating the shallow active marrow in the ribs.

PAGE 161

161 Table A-13. UFRF absorbed fraction for all sour ces irradiating the active marrow in the cervical vertebrae. Table A-14. UFRF absorbed fraction for all sour ces irradiating the shallow active marrow in the cervical vertebrae.

PAGE 162

162 Table A-15. UFRF absorbed fraction for all sour ces irradiating the active marrow in the thoracic vertebrae. Table A-16. UFRF absorbed fraction for all sour ces irradiating the shallow active marrow in the thoracic vertebrae.

PAGE 163

163 Table A-17. UFRF absorbed fraction for all sour ces irradiating the active marrow in the lumbar vertebrae. Table A-18. UFRF absorbed fraction for all sour ces irradiating the shallow active marrow in the lumbar vertebrae.

PAGE 164

164 Table A-19. UFRF absorbed fraction for all sour ces irradiating the active marrow in the sacrum. Table A-20. UFRF absorbed fraction for all sour ces irradiating the shallow active marrow in the sacrum.

PAGE 165

165 Table A-21. UFRF absorbed fraction for all so urces irradiating the active marrow in the ossa coxae. Table A-22. UFRF absorbed fraction for all sour ces irradiating the shallow active marrow in the ossa coxae.

PAGE 166

166 Table A-23. UFRF absorbed fraction for all sources irradiating the active marrow in the proximal humeri. Table A-24. UFRF absorbed fraction for all sour ces irradiating the shallow active marrow in the proximal humeri.

PAGE 167

167 Table A-25. UFRF absorbed fraction for all sources irradiating the active marrow in the proximal femora. Table A-26. UFRF absorbed fraction for all sour ces irradiating the shallow active marrow in the proximal femora.

PAGE 168

168 APPENDIX B GRAPHS OF SPECIFIC ABORBED FRACTION This appendix graphs Specific Absorbed Fraction (SAF) data for UFRF for any combination of source (TAM, TBS, TBV, TAM50, TIM) and target (TAM, TAM50) regions in the entire skeleton containing active marrow for a range of cellularities (10% to 100%). Absorbed Fraction (AF) values were normalized to the mass of the target (and cellularity) to report SAF. For the TAM irradiating the TAM, marrow cellularity is varied. For all sources irradiating the TAM or TAM50, ICRP reference cellularity is used. Figure B-1. Specific absorbed fraction for th e TAM irradiating the TAM for the cranium.

PAGE 169

169 Figure B-2. Specific absorbed fraction for all sources irradiating the TAM for the cranium.

PAGE 170

170 Figure B-3. Specific absorbed frac tion for the TAM irradiating the TAM50 for the cranium.

PAGE 171

171 Figure B-4. Specific absorbed fracti on for all sources irradiating the TAM50 for the cranium.

PAGE 172

172 Figure B-5. Specific absorbed fraction for th e TAM irradiating the TAM for the mandible.

PAGE 173

173 Figure B-6. Specific absorbed fraction for all sources irradiating the TAM for the mandible.

PAGE 174

174 Figure B-7. Specific absorbed frac tion for the TAM irradiating the TAM50 for the mandible.

PAGE 175

175 Figure B-8. Specific absorbed fracti on for all sources irradiating the TAM50 for the mandible.

PAGE 176

176 Figure B-9. Specific absorbed fraction for th e TAM irradiating the TAM for the scapulae.

PAGE 177

177 Figure B-10. Specific absorbed fraction for all sources irradiating the TAM for the scapulae.

PAGE 178

178 Figure B-11. Specific absorbed frac tion for the TAM irradiating the TAM50 for the scapulae.

PAGE 179

179 Figure B-12. Specific absorbed fracti on for all sources irradiating the TAM50 for the scapulae.

PAGE 180

180 Figure B-13. Specific absorbed fraction for th e TAM irradiating the TAM for the clavicles.

PAGE 181

181 Figure B-14. Specific absorbed fraction for all sources irradiating the TAM for the clavicles.

PAGE 182

182 Figure B-15. Specific absorbed frac tion for the TAM irradiating the TAM50 for the clavicles.

PAGE 183

183 Figure B-16. Specific absorbed fracti on for all sources irradiating the TAM50 for the clavicles.

PAGE 184

184 Figure B-17. Specific absorbed fraction for th e TAM irradiating the TAM for the sternum.

PAGE 185

185 Figure B-18. Specific absorbed fraction for al l sources irradiating the TAM for the sternum.

PAGE 186

186 Figure B-19. Specific absorbed frac tion for the TAM irradiating the TAM50 for the sternum.

PAGE 187

187 Figure B-20. Specific absorbed fracti on for all sources irradiating the TAM50 for the sternum.

PAGE 188

188 Figure B-21. Specific absorbed fraction for the TAM irradiating the TAM for the ribs.

PAGE 189

189 Figure B-22. Specific absorbed fraction for a ll sources irradiating the TAM for the ribs.

PAGE 190

190 Figure B-23. Specific absorbed frac tion for the TAM irradiating the TAM50 for the ribs.

PAGE 191

191 Figure B-24. Specific absorbed fracti on for all sources irradiating the TAM50 for the ribs.

PAGE 192

192 Figure B-25. Specific absorbed fraction for the TAM irradiating the TAM for the cervical vertebrae.

PAGE 193

193 Figure B-26. Specific absorbed fraction for all sources irradiating the TAM for the cervical vertebrae.

PAGE 194

194 Figure B-27. Specific absorbed frac tion for the TAM irradiating the TAM50 for the cervical vertebrae.

PAGE 195

195 Figure B-28. Specific absorbed fracti on for all sources irradiating the TAM50 for the cervical vertebrae.

PAGE 196

196 Figure B-29. Specific absorbed fraction for the TAM irradiating the TAM for the thoracic vertebrae.

PAGE 197

197 Figure B-30. Specific absorbed fraction for a ll sources irradiating the TAM for the thoracic vertebrae.

PAGE 198

198 Figure B-31. Specific absorbed frac tion for the TAM irradiating the TAM50 for the thoracic vertebrae.

PAGE 199

199 Figure B-32. Specific absorbed fracti on for all sources irradiating the TAM50 for the thoracic vertebrae.

PAGE 200

200 Figure B-33. Specific absorbed fraction for the TAM irradiating the TAM for the lumbar vertebrae.

PAGE 201

201 Figure B-34. Specific absorbed fraction for a ll sources irradiating the TAM for the lumbar vertebrae.

PAGE 202

202 Figure B-35. Specific absorbed frac tion for the TAM irradiating the TAM50 for the lumbar vertebrae.

PAGE 203

203 Figure B-36. Specific absorbed fracti on for all sources irradiating the TAM50 for the lumbar vertebrae.

PAGE 204

204 Figure B-37. Specific absorbed fraction for th e TAM irradiating the TAM for the sacrum.

PAGE 205

205 Figure B-38. Specific absorbed fraction for a ll sources irradiating the TAM for the sacrum.

PAGE 206

206 Figure B-39. Specific absorbed frac tion for the TAM irradiating the TAM50 for the sacrum.

PAGE 207

207 Figure B-40. Specific absorbed fracti on for all sources irradiating the TAM50 for the sacrum.

PAGE 208

208 Figure B-41. Specific absorbed fraction for the TAM irradiating the TAM for the os coxae.

PAGE 209

209 Figure B-42. Specific absorbed fraction for all sources irradiating the TAM for the os coxae.

PAGE 210

210 Figure B-43. Specific absorbed frac tion for the TAM irradiating the TAM50 for the os coxae.

PAGE 211

211 Figure B-44. Specific absorbed fracti on for all sources irradiating the TAM50 for the os coxae.

PAGE 212

212 Figure B-45. Specific absorbed fraction for th e TAM irradiating the TAM for the proximal humeri.

PAGE 213

213 Figure B-46. Specific absorbed fraction for a ll sources irradiating th e TAM for the proximal humeri.

PAGE 214

214 Figure B-47. Specific absorbed frac tion for the TAM irradiating the TAM50 for the proximal humeri.

PAGE 215

215 Figure B-48. Specific absorbed fracti on for all sources irradiating the TAM50 for the proximal humeri.

PAGE 216

216 Figure B-49. Specific absorbed fraction for th e TAM irradiating the TAM for the proximal femora.

PAGE 217

217 Figure B-50. Specific absorbed fraction for a ll sources irradiating th e TAM for the proximal femora.

PAGE 218

218 Figure B-51. Specific absorbed frac tion for the TAM irradiating the TAM50 for the proximal femora.

PAGE 219

219 Figure B-52. Specific absorbed fracti on for all sources irradiating the TAM50 for the proximal femora.

PAGE 220

220 APPENDIX C SPECIFIC ABSORBED FRACTIONS FOR ALL BONE SITES Specific Absorbed Fraction (SAF) data for UF RF is given for any combination of source (TAM, TBS, TBV, TAM50, TIM) and target (TAM, TAM50) regions in the entire skeleton containing active marrow. ICRP reference cel lularity for each bone site was used. Figure C-1. Specific absorbed fraction for the TAM irradiating the TAM for all bone sites. Figure C-2. Specific absorbed frac tion for the TAM irradiating the TAM50 for all bone sites.

PAGE 221

221 Figure C-3. Specific absorbed fraction for the TBS irradiati ng the TAM for all bone sites. Figure C-4. Specific absorbed frac tion for the TBS irradiating the TAM50 for all bone sites.

PAGE 222

222 Figure C-5. Specific absorbed fraction for the TBV irradiati ng the TAM for all bone sites. Figure C-6. Specific absorbed frac tion for the TBV irradiating the TAM50 for all bone sites.

PAGE 223

223 Figure C-7. Specific absorbed fraction for the CBV irradiating the TAM for all bone sites. Figure C-8. Specific absorbed frac tion for the CBV irradiating the TAM50 for all bone sites.

PAGE 224

224 Figure C-9. Specific absorbed fraction for the TIM irradiatin g the TAM for all bone sites. Figure C-10. Specific absorbed frac tion for the TIM irradiating the TAM50 for all bone sites.

PAGE 225

225 APPENDIX D SPECIFIC ABSORBED FRACTIONS FOR LONG BONES This appendix provides Specific Absorbed Frac tion (SAF) data in bon e sites not containing active marrow for combinations of sources (CBV, CIM, CBSMC) and target (CIM50) regions. For the shafts of the long bones of th e adult, cellularity is 0%. Absorbed fractions are divided by the mass of the shaft of the long bones to report SAF. Table D-1. Specific absorbed fracti ons in the shafts of the leg bones.

PAGE 226

226 Table D-2. Specific absorbed fracti ons in the shafts of the arm bones.

PAGE 227

227 Figure D-1. Specific absorbed frac tion for the CIM irradiating the CIM50 in the shafts of all long bones.

PAGE 228

228 Figure D-2. Specific abso rbed fraction for the CBSMC irradiating the CIM50 in the shafts of all long bones.

PAGE 229

229 Figure D-3. Specific absorbed frac tion for the CBV irradiating the CIM50 in the shafts of all long bones.

PAGE 230

230 APPENDIX E SKELETAL AVERAGED AB SORBED FRACTIONS Skeletal averaged Absorbed Fraction (AF) data for UFRF is given for any combination of source (TAM, TBS, TBV, TAM50, TIM) and target (TAM, TAM50) regions in the entire skeleton containing active marrow. Bone site weighting fact ors were used to average AF for each skeletal site. Values of AF at ICRP reference cellularity were then compared to values from Stabin and Siegel (2003). Table E-1. UFRF skeletal aver aged absorbed fraction (AF) for various sources irradiating the trabecular active marrow.

PAGE 231

231 Table E-2. UFRF skeletal aver aged absorbed fraction (AF) for various sources irradiating the shallow trabecular active marrow. Table E-3. UFRF skeletal av eraged specific absorbed frac tion (SAF) for various sources irradiating the trabecular active marrow.

PAGE 232

232 Table E-4. UFRF skeletal specific averaged absorbed fraction (SAF) for various sources irradiating the shallow trabecular active marrow.

PAGE 233

233 Figure E-1. Skeletal averaged AF for marrow sources irradiating the TAM for both UFRF and Stabin and Siegel (2003).

PAGE 234

234 Figure E-2. Skeletal averaged AF for trabecular bone source s irradiating the TAM for both UFRF and Stabin and Siegel (2003).

PAGE 235

235 Figure E-3. Skeletal aver aged AF for marrow sources irradiating the TAM50 for both UFRF and Stabin and Siegel (2003).

PAGE 236

236 Figure E-4. Skeletal averag ed AF for trabecular bone s ources irradiating the TAM50 for both UFRF and Stabin and Siegel (2003).

PAGE 237

237 LIST OF REFERENCES Armitage J O 2005 Stagi ng Non-Hodgkin Lymphoma CA Cancer J Clin 55 368-76 Atkinson P J 1965 Changes in resorption spaces in femoral cortical bone with age. J. of Pathol. and Bacteriol. 89 173-8 Beddoe A H, Darley P J and Spiers F W 1976 Meas urements of trabecular bone structure in man Phys. Med. Biol. 21 589-607 Berne R M 1993 Physiology (St. Louis: Mosby) Bolch W, Jokisch D, Patton P, Bouchet L, Rajon D, Inglis B and Myers S 1998 NMR microimaging of trabecular bone: a new tool for the development of bone dosimetry models Health Phys. 74 S13 Bolch W E, Shah A P, Watchman C J, Jokisch D W, Patton P W, Rajon D A, Zankl M, PetoussiHenss N and Eckerman K F 2007 Skeletal absorb ed fractions for elect rons in the adult male: Considerations of a revised 50-m icron definition of the bone endosteum Radiat Prot Dosimetry Bouchet L G and Bolch W E 1999 A three-dimensional transport model for determining absorbed fractions of energy for electrons in cortical bone J. Nucl. Med. 40 2115-24 Bouchet L G, Bolch W E, Howe ll R W and Rao D V 2000 S values for radionuclides localized within the skeleton J. Nucl. Med. 41 189-212 Bourke V A, Watchman C J, Reith J D, Jorg ensen M L, Dieudonne A and Bolch W in press Spatial gradient of blood vesse ls and hematopoietic stem a nd progenitor cells within the marrow cavities of the human skeleton Blood Brindle J, Trindade A, Pichardo J, Myers S, Shah A and Bolch W 2006a CT volumetry of the skeletal tissues Med Phys 33 3796-803 Brindle J M, Myers S L and Bolch W E 2006b Correlations of total pelvic spongiosa volume with both anthropometric parameters and CT-based skeletal size measurements Cancer Biother Radiopharm 21 352-63 Brindle J M, Trindade A A, Shah A P, Jokisch D W, Pa tton P W and Bolch W E 2006c Linear regression model for predicting patient-specific total skeletal spongiosa volume for use in molecular radiotherapy dosimetry J Nucl Med 47 1875-83 Calvi L M, Adams G B, Weibrech t K W, Weber J M, Olson D P, Knight M C, Martin R P, Schipani E, Divieti P, Bri nghurst F R, Milner L A, Kronenberg H M and Scadden D T 2003 Osteoblastic cells regulate the haematopoietic stem cell niche Nature 425 841-6

PAGE 238

238 Charlton D E, Utteridge T D a nd Beddoe A H 1996 Microdosimetry of haemopoietic stem cells irradiated by alpha particles from the shortlived products of 222Rn decays in fat cells and haemopoietic tissue Int J Radiat Biol 69 585-92 Chavez-Gonzalez A, Rosas-Cabral A, Vela-Oje da J, Gonzalez J and Mayani H 2004 Severe functional alterations in vitr o in CD34+ cell subpopulations from patients with chronic myeloid leukemia Leukemia research 28 639-47 Cristy M 1981 Active bone marrow distri bution as a function of age in humans Phys Med Biol 26 389-400. Cui Y F, Lord B I, Woolford L B and Testa N G 1996 The relative spatial di stribution of in vitroCFCs in the bone marrow, responding to specific growth factors Cell Prolif 29 243-57 Custer R 1974 An atlas of the blood and bone marrow (Philadelphia, PA: W.B. Saunders) DOH 2001 Stem Cells: Scientific Progress and Future Research Directions. Department of Health and Human Services) DOH 2006 Regenerative Medicine. Department of H ealth and Human Services) Eckerman K F 1985 Aspects of the dosimetry of ra dionuclides within the sk eleton with particular emphasis on the active marrow. In: Proceedings of the Fourth International Radiopharmaceutical Dosimetry Symposium, ed A T Schlafke-Stelson and E E Watson (Oak Ridge, Tennessee: Oak Ridge A ssociated Universities) pp 514-34 Eckerman K F, Ryman J C, Taner A C and Ke rr G D 1985 Traversal of cells by radiation and absorbed fraction estimates for electrons and alpha particles. In: Proceedings of the Fourth International Radiopharmaceutic al Dosimetry Symposium, ed A T Schlafke-Stelson and E E Watson (Oak Ridge, Tennessee: ORAU) pp 67-81 Eckerman K F and Stabin M G 2000 Electron absorb ed fractions and dose conversion factors for marrow and bone by skeletal regions Health Phys. 78 199-214 Forero A and Lobuglio A F 2003 History of antibody therapy for non-Hodgkin's lymphoma Semin Oncol 30 1-5 Foucar K 2001 Bone marrow pathology (Chicago, IL: American Society of Clinicial Pathologists) Frassoni F, Testa N G and Lord B I 1982 The rela tive spatial distribution of erythroid progenitor cells (BFUe and CFUe) in the normal mouse femur Cell Tissue Kinet 15 447-55 Gossner W 2003 Target cells in internal dosimetry Radiat Prot Dosimetry 105 39-42 Gossner W, Masse R and Stather J W 2000 Cells at risk for dosimetric mode ling relevant to bone tumour induction Radiat Prot Dosimetry 92 209-13

PAGE 239

239 Harris N, Jaffe E, Stein H, Banks P, Chan J, Cleary M, Delsol G, De Wolf-Peeters C, Falini B and Gatter K 1994 A revised European-American classification of lymphoid neoplasms: a proposal from the International Lymphoma Study Group [see comments] Blood 84 1361 Heddy Z and Swart B 2005 The human leucocyte differentiation antigens (HLDA) workshops: the evolving role of antibodies in research, diagnosis and therapy Cell Research 15 691-4 Hoffman R, Benz Jr E, Shattil S, Furie B a nd Cohen H 1991 Hematology: basic principles and practice Blood 235 Huerta-Zepeda A, Talavera A, Avils A, Neri N and Mayani H 2000 In v itro hematopoiesis in patients with malignant lympho ma during active disease and at complete clinical remission after chemotherapy Leukemia & lymphoma 39 613 ICRP 1977 Recommendations of the Internati onal Commission on Radi ological Protection. Pergamon, Oxford) ICRP 1995 Basic anatomical and physiological data for use in radiological protection: the skeleton. (Oxford, UK: Internationa l Commission on Radiol ogical Protection) ICRP 2002 Basic anatomical and physiological data for use in radiological protection: reference values. (Oxford; Pergamon Press: Interna tional Commission on Radi ological Protection) Jensen K, Jensen M, Grundtvig P, Thomsen C, Karle H and Henriksen O 1990 Localized in vivo proton spectroscopy of the bone marrow in patients with leukemia Magnetic resonance imaging 8 779 Jokisch D 1999 Beta particle dosimetry of the tr abecular region of a thorac ic vertebra utilizing NMR microscopy. In: Nuclear and Radiologi cal Engineering, (Gainesville, Florida: University of Florida) p 235 Juweid M, Sharkey R M, Siegel J A, Behr T and Goldenberg D M 1995 Estimates of red marrow dose by sacral scintigraphy in radioi mmunotherapy patients having non-Hodgkin's lymphoma and diffuse bone marrow uptake Cancer Res 55 5827s-31s. Juweid M E, Zhang C H, Blumenthal R D, Hajjar G, Sharkey R M and Goldenberg D M 1999 Prediction of hematologic toxicity afte r radioimmunotherapy with (131)Ilabeled anticarcinoembryonic antigen monoclonal antibodies J Nucl Med 40 1609-16. Kaminski M, Tuck M, Estes J, Kolstad A, Ross C, Zasadny K, Regan D, Kison P, Fisher S and Kroll S 2005 131I-tositumomab therapy as initial treatment for follicular lymphoma The New England journal of medicine 352 441 Kuppers R, Klein U, Hansmann M and Raje wsky K 1999 Cellular origin of human B-cell lymphomas New England Journal of Medicine 341 1520-9 Kuznetsov S A, Mankani M H, Gronthos S, Satomura K, Bianco P and Robey P G 2001 Circulating skeletal stem cells J Cell Biol 153 1133-40

PAGE 240

240 Kuznetsov S A, Riminucci M, Ziran N, Tsutsu i T W, Corsi A, Calvi L, Kronenberg H M, Schipani E, Robey P G and Bianco P 2004 The interplay of osteogenesis and hematopoiesis: expression of a constitutivel y active PTH/PTHrP receptor in osteogenic cells perturbs the establishment of hematopoiesis in bone and of skeletal stem cells in the bone marrow J Cell Biol 167 1113-22 Kvinnsland Y, Skretting A and Bruland O S 2001 Radionuclide therapy with bone-seeking compounds: Monte Carlo calculations of dos e-volume histograms for bone marrow in trabecular bone Phys Med Biol 46 1149-61 Lim S M, DeNardo G L, DeNar do D A, Shen S, Yuan A, O' Donnell R T and DeNardo S J 1997a Prediction of myelotoxicity us ing radiation doses to marro w from body, blood and marrow sources J Nucl Med 38 1374-8. Lim S M L, DeNardo G L, DeNardo D A, Shen S, Yuan A, O'Donnell R T and DeNardo S J 1997b Prediction of myelotoxicity using radiat ion doses to marrow from body, blood, and marrow sources J. Nucl. Med. 38 1374-8 Liu T, Meredith R F, Saleh M N, Wheeler R H, Khazaeli M B, Plott W E, Schlom J and LoBuglio A F 1997 Correlation of toxicity with treatment parameters for 131I-CC49 radioimmunotherapy in three phase II clinical trials Cancer Biother Radiopharm 12 79-87. Loevinger R, Budinger T F and Watson E E 1991 MIRD primer for absorbed dose calculations (New York: The Society of Nuclear Medicine) Lombardi M H 2006 Radiation safety in nuclear medicine: CRC Press) Lord B I and Hendry J H 1972 The distribution of hematopoietic colony forming units in the mouse femur and its modification by x-rays Br. J. Radiol. 45 110-5 Macey D J, DeNardo S J, DeNa rdo G L, DeNardo D A and Shen S 1995 Estimation of radiation absorbed doses to the red marrow in radioimmunotherapy Clin Nucl Med 20 117-25 Madrigal-Velazquez M, Avilas A, Neri N, Huer ta J, Martinez-Jaramillo G and Mayani H 2006 Expression of ice, bcl-2, c-myc and p53 in different bone marrow cell populations from patients with diffuse large B-cell lymphoma Leukemia and Lymphoma 47 665-73 Marieb E N 2006 Human anatomy and physiology (San Francisco, CA: P earson Education, Inc.) Martinez-Jaramillo G, Aviles A, Neri N, Huerta J, Madrigal-Velazquez M, Flores-Guzman P and Mayani H 2004 In vitro proliferation and e xpansion of hematopoietic progenitors from patients with diffuse large B-cell lym phoma before and after chemotherapy Leukemia and Lymphoma 45 1247-54 McPherson R and Pincus M 2007 Henry's clinical diagnosis and management by laboratory methods: Saunders Elsevier)

PAGE 241

241 Mechanik N 1926 Studies of the weight of bone marrow in man Zeitschrift fur die Gest. Anatomy 79 58-99 Mosekilde L 1989 Sex differences in age-relate d loss of vertebral trabecular bone mass and structure biomechanical consequences Bone 10 425-32 Moulopoulos L A and Dimopoulos M A 1997 Magnetic Resonance Im aging of the Bone Marrow in Hematologic Malignancies Blood 90 2127-47 Muller A, Ihorst G, Mertelsmann R and E ngelhardt M 2005 Epidemiology of non-Hodgkin's lymphoma (NHL): trends, geogra phic distribution, and etiology Annals of Hematology 84 1-12 O'Donoghue J A, Baidoo N, Deland D, Welt S, Divgi C R and Sgouros G 2002 Hematologic toxicity in radioimmunothera py: dose-response relationships for I-131 labeled antibody therapy Cancer Biother Radiopharm 17 435-43 Orazi A, Cattoretti G, Schiro R, Siena S, Bregni M, Di Nico la M and Gianni A 1992 Recombinant human interleukin-3 and reco mbinant human granulocytemacrophage colony-stimulating factor administered in vivo after highdose cyclophosphamide cancer chemotherapy: effect on hematopoiesis a nd microenvironment in human bone marrow Blood 79 2610-9 Patton P W 2000 NMR microscopy for skeletal dosimetry: an invest igation of marrow cellularity on dose estimates. In: Nuclear and Radiologi cal Engineering, (Gainesville, FL: University of Florida) p 426 Patton P W, Jokisch D W, Raj on D A, Shah A P, Myers S L and Bolch W E 2002a Skeletal dosimetry via NMR microscopy: investigations of sample reproducibility and signal source Health Phys 82 316-26 Patton P W, Rajon D A, Shah A P, Jokisch D W, Inglis B and Bolch W E 2002b Site-specific variability in trabecular bone dosimetry: cons iderations of energy loss to cortical bone Med Phys 29 6-14 Pichardo J C, Trindade A A, Brindle J M a nd Bolch W E 2007 Method fo r estimating skeletal spongiosa volume and active marrow mass in the adult male and adult female J Nucl Med 48 1880-8 Rafii S, Shapiro F, Rimarachin J, Nachman R L, Ferris B, Weksler B, Moore M A and Asch A S 1994 Isolation and characteri zation of human bone marrow microvascular endothelial cells: hematopoietic progenitor cell adhesion Blood 84 10-9 Rajon D, Pichardo J, Brindle J, Kielar K, Jokisch D, Patton P and Bolch W 2006 Image segmentation of trabecular spongiosa by visual inspection of the gradient magnitude Phys Med Biol 51 4447-67

PAGE 242

242 Rajon D A, Jokisch D W, Patton P W, Shah A P and Bolch W E 2000 Voxel size eff ects in threedimensional nuclear magnetic resonance microscopy performed for trabecular bone dosimetry Med Phys 27 2624-35 Schick F, Weiss B and Einsele H 1995 Magneti c resonance imaging reveals a markedly inhomogeneous distribution of marrow cellula rity in a patient with myelodysplasia Annals of Hematology 71 143-6 Sgouros G 1993 Bone marrow dosimetry for radi oimmunotherapy: Theoretic al considerations J. Nucl. Med. 34 689-94 Sgouros G, Chiu S, Pentlow K S, Brewster L J, Kalaigian H, Baldwin B, Daghighian F, Graham M C, Larson S M and Mohan R 1993 Th ree-dimensional dosimetry for radioimmunotherapy treatment planning J. Nucl. Med. 34 1595-601 Sgouros G, Deland D, Loh A C, Divgi C R and Larson S M 1997 Marrow and whole-body absorbed dose versus marrow toxicity folli ng 131-I-G250 antibody therapy in patients with renal-cell carcinoma J. Nucl. Med. 38 252P Sgouros G, Divgi C R, Scott A M, Willams J and Larson S M 1996a Hematologic toxicity in radioimmunotherapy: An evaluation of different predictive measures J. Nucl. Med. 37 43P Sgouros G, Howell R, Bolch W and Fisher D 2009 MIRD Commentary: Proposed Name for a Dosimetry Unit Applicable to Deterministic Biological Effects--The Barendsen (Bd) Journal of Nuclear Medicine 50 485 Sgouros G, Jureidini I M, Scott A M, Grah am M C, Larson S M and Scheinberg D A 1996b Bone marrow dosimetry: regional variab ility of marrow-localizing antibody J. Nucl. Med. 37 695-8 Shah A 2004 Reference skeletal dosimetry model for an adult male radionuclide therapy patient based on 3D imaging and paired-i mage radiation transport. In: Department of Biomedical Engineering, (Gainesville, FL: University of Florida) p 488 Shah A, Jokisch D, Watchman C, Rajon D, Patton P and Bolch W 2005a Chord-based versus voxel-based methods of electron tran sport in the skeletal tissues Med Phys 32 3151-9 Shah A P, Bolch W E, Rajon D A, Patton P W and Jokisch D W 2005b A pa ired-image radiation transport model for skeletal dosimetry J Nucl Med 46 344-53 Shah A P, Patton P W, Rajon D A and Bolch W E 2003 Adipocyte spatial distributions in bone marrow: Implications for skeletal dosimetry models J Nucl Med 44 774-83 Shen S, Meredith R F, Duan J, Macey D J, Khazaeli M B, Robert F and LoBuglio A F 2002 Improved prediction of myelotoxi city using a patient-specific imaging dose estimate for non-marrow-targeting (9 0)y-antibody therapy J Nucl Med 43 1245-53

PAGE 243

243 Siegel J A, Goldenberg D M and Badger C C 1993 Radioimmunot herapy dose estimates in patients with B-cell lymphoma Med. Phys. 20 579-82 Siegel J A, Pawlyk D A, Lee R E, Sasso N L, Horowitz J A, Sharkey R M and Goldenberg D M 1990a Tumor, red marrow, and organ dosimetry for 131I-labeled anti-carcinoembryonic antigen monoclonal antibody Cancer Res. (Suppl.) 50 1039s-42s Siegel J A, Stabin M G a nd Sparks R B 2003 Total-body and red marrow dose estimates J Nucl Med 44 320-1; author reply 1-3 Siegel J A, Wessels B W and Watson E E 1990b Bone marrow dosimetry and toxicity for radioimmunotherapy Antibody Immunoconj ugates Radiopharm. 3 213-33 Snyder B D, Piazza S, Edwards W T and Hayes W C 1993 Role of trabecular morphology in the etiology of age-related vertebral fractures Calcified Tissue International 53 (Suppl 1) S14S22 Soligo D, Lambertenghi Delilieris G, Servid a F, Quirici N, Campiglio S and E T 1998 Hematopoietic abnormalities after autologous stem cell transplantation in lymphoma patients Bone Marrow Transplantation 21 15-22 Spiers F W, Whitwell J R and Beddoe A H 1978 Ca lculated dose factors for the radiosensitive tissues in bone irradiated by surface-deposited radionuclides Phys Med Biol 23 481-94. Stabin M 2004 The OLINDA/EXM Personal Co mputer Code. Vanderbilt University) Stabin M G 1994 MIRDOSE3. (Oak Ridge: Oak Ri dge Institute for Science and Education) Stabin M G 1996 MIRDOSE: personal computer software for internal dose assessment in nuclear medicine J. Nucl. Med. 37 538-46 Stabin M G, Sparks R B and Crowe E 2005 OL INDA/EXM: the second-generation personal computer software for internal dose assessment in nuclear medicine J Nucl Med 46 1023-7 Sugiyama T, Kohara H, Noda M and Nagasawa T 2006 Maintenance of the hematopoietic stem cell pool by CXCL12-CXCR4 chem okine signaling in bone marrow stromal cell niches Immunity 25 977-88 Swerdlow S, Campo E, Harris N, Jaffe E, Pile ri S, Stein H, Thiele J and Vardiman J 2008 WHO Classification of tumours of Haematopoietic and Lymphoid Tissues: International Agency for Research on Cancer (IARC)) Taichman R S 2005 Blood and bone: two tissues w hose fates are intertwined to create the hematopoietic stem-cell niche Blood 105 2631-9 Trotter M and Hixon B 1974 Sequential changes in weight, density, and pe rcentage ash weight of human skeletons from an ear ly fetal period through old age Anat. Rec. 179 1-18

PAGE 244

244 U. S. National Institutes of Hea lth N C I SEER Training Modules. US Department of Health a nd Human Services O o t S G Vogler J 1988 Bone marrow imaging Radiology 168 679 Watchman C, Jokisch D W, Patton P, Raj on D, Sgouros G and Bolch W 2005 Absorbed fractions for alpha-particles in tissues of trabecular bone : Considerations of marrow cellularity within the ICRP reference male J Nucl Med 46 1171-85 Watchman C J 2005 Skeletal dosimetry models for alpha-particles for use in molecular radiotherapy. In: Nuclear and Radiological Engineering, (Gainesville, FL: University of Florida) Watchman C J, Bourke V A, Lyon J R, Knowlt on A E, Butler S L, Grier D D, Wingard J R, Braylan R C and Bolch W E 2007 Spatial di stribution of blood vessels and CD34+ hematopoietic stem and progenitor cells with in the marrow cavities of human cancellous bone J Nucl Med 48 645-54 Whitwell J R 1973 Theoretical inve stigations of energy loss by ionizing particles in bone. In: Department of Medical Physics, (Leeds, UK: University of Leeds) p 268 Whitwell J R and Spiers F W 1976 Calculated beta-ray dose factors for trabecular bone Phys. Med. Biol. 21 16-38 Wiseman G, White C, Sparks R, Erwin W, Podo loff D, Lamonica D, Bartlett N, Anthony Parker J, Dunn W and Spies S 2001 Biodistribution a nd dosimetry results from a phase III prospectively randomized controlled trial of Zevalin radioimmunotherapy for low-grade, follicular, or transformed B-cell non-Hodgkin's lymphoma Critical Reviews in Oncology and Hematology 39 181-94 Yao M, Fouillard L, Lemoine F, Bouchet S, Fira t H, Andreu G, Gorin N, Douay L and Lopez M 2000 Ex vivo expansion of CD34positive peripheral blood progenitor cells from patients with non-Hodgkin's lymphoma: no evidence of concomitant expansion of contaminating bcl2/JH-positive lymphoma cells Bone Marrow Transplantation 26 497 Yao W, Wu C, Wang S, Chang C, Chiu N and Yu C 2001 Differential changes in regional bone mineral density in healthy Chines e: age-related and sex-dependent Calcified Tissue International 68 330-6 Yoffey J and Courtice F 1970 Lymphatics, lymph and th e lymphomyeloid complex: Academic Press) Zhang J, Niu C, Ye L, Huang H, He X, Tong W, Ross J, Haug J, Johnson T and Feng J 2003 Identification of the haematopoietic stem cell niche and control of the niche size Nature 425 836-41

PAGE 245

245 BIOGRAPHICAL SKETCH Kayla N. Kielar was born in Pittsburgh, Pennsylvania, but grew up in Palm Beach, FL. Kayla is the daughter of David and Susan Kielar a nd the sister of Karla. Kayla graduated from the Medical Magnet program at La ke Worth High School in 2001 and her career as a Gator at the University of Florida began. She went on to receive a Bachelor of Science in Nuclear Engineering, graduating Summa Cum Laude in 2004. In 2006, Ka yla earned a Master of Science in Nuclear Engineering Sciences. While engaged in doctoral research, Kayl a pursued an M.S. in Business Management from the Warrington College of Business, which she earned in May 2009. She also finished a full Ironman triathlon the sa me year. Upon completion of her Doctorate in Philosophy, Kayla will be joining the Department of Radiation Oncology at the Stanford Cancer Center as a radiation oncology physics resident.