Darrell R. Fisher Pacific Northwest National Laboratory Richland, Washington

1. Need for Internal Dosimetry: Fundamentals to Clinical Practice –and Opportunity to Personalize Cancer Therapy Darrell R. Fisher Pacific Northwest National Laboratory Richland, Washington Joseph G. Rajendran University of Washington, Seattle, Washington Slides are not to be reproduced without permission of author

2. Educational objectives • Examine the need for and importance of radiation dosimetry in the nuclear medicine clinic • Benefits • Limitations • Approach and relevant quantities • Identify the important aspects of patient-specific dosimetry • Understand why patient-specific dosimetry is critical for achieving optimal therapy outcomes Slides are not to be reproduced without permission of author

3. “The objectives of dosimetry are to predict response and to guide clinical investigation toward safe and effective implementation of diagnostic and therapeutic radiopharmaceuticals. “These objectives can only be met if advances in computational and physics tools are coupled with advances in our understanding of the biokinetic behavior and radiobiology of radiopharmaceuticals. -- Dr. George Sgouros, Chair, MIRD Committee Slides are not to be reproduced without permission of author

4. Purpose for dosimetry Diagnostic nuclear medicine: • Provides the fundamental quantities used in radiation protection and risk assessment • Provides product safety information (package insert) Therapeutic nuclear medicine: • To propose and plan appropriate treatments • To correlate dose with biological effects in - Normal organs and tissues, and the whole body - Tumors • To predict radiation effects • To establish complete patient records Slides are not to be reproduced without permission of author

5. Rationale for therapy A. Treat to a desired radiation dose to the tumor • Thyroid cancer • Requires data on tumor mass, tumor uptake and retention OR B. Treat to a maximum tolerable radiation dose to the limiting normal organ or tissue • Radioimmunotherapy • Requires patient-specific data on patient organ mass, and the uptake, retention, and clearance of the radiopharmaceutical from the major normal organs Slides are not to be reproduced without permission of author

6. Counter-balanced goals • Establish a rigorous scientific rationale for selecting patient-individualized treatment levels • Minimize to the extent possible the complexity, costs, and time required for imaging and dosimetry rigor cost Slides are not to be reproduced without permission of author

7. Two approaches 1. Dynamic Modeling • Requires an appropriate pharmacokinetic or biokinetic model • Requires known parameter values for the model compartments and transfer rates Example: Biokinetic models established by the International Commission on Radiological Protection (ICRP) • Implemented using modeling software (SAAM II, STELLA) A1 A0 A3 l1 V1 A2 V2 l3 l2 A4 l4 Slides are not to be reproduced without permission of author

8. Two approaches (continued) 2. Direct Measurements • Planar imaging (and quantitative SPECT) - Patient positioning, anterior/posterior geometric means - Regions of interest for the major organs, tumor, whole body - Translation from counts to activity - Calibration against a radionuclide standard - Background subtraction, attenuation correction • Marrow and tumor biopsy specimens • Organ volumetrics by CT scans • Activity-time curve-fitting • Area-under-curve analysis • Dosimetry calculations using the MIRD schema (implemented using software (such as OLINDA-EXM) Slides are not to be reproduced without permission of author

9. Slides are not to be reproduced without permission of author In-111 Day 4, posterior Imaging-based time-activity curve activity • • • • Time, hr

10. MIRD schema • Simplifies dose assessment without compromising on essential details • Is evolving to meet 21st century needs • Extends from the whole-organ to the cellular and multicellular levels • May be applied to uniform or non-uniform radionuclide distributions • Patient-specific methods are preferred over use of generic model assumptions Slides are not to be reproduced without permission of author

11. Slides are not to be reproduced without permission of author

12. Relevant quantity • The relevant dosimetric quantity (and unit) is the absorbed dose (Gy) • Relevant to organ and tumor deterministic effects • Not the “effective dose” (Sv) - In the ICRP terminology, effective dose is used in radiation protection and planning, *Defined only for stochastic effects (cancer induction) * Not for individual patient dosimetry Slides are not to be reproduced without permission of author

13. Absorbed dose in biological systems • General equation where k is a unit conversion constant A is the activity in the organ (Bq) E is the total energy emitted (J) f is the fraction of energy that is absorbed m is the mass of target tissue (g) B(t) is the biological retention with time t Slides are not to be reproduced without permission of author

14. Absorbed dose Rearranged MIRD formula D(rkrh) = Ãhii φi(rkrh) / mk where Ãh is the cumulated activity, i is the mean energy emitted per unit cumulated activity, and φ(rkrh) is the absorbed fraction of energy imparted by a source organ Slides are not to be reproduced without permission of author

15. Absorbed dose Rearranged MIRD formula total number of transformations D(rkrh) = Ãhii φi(rkrh) / mk Slides are not to be reproduced without permission of author

16. Absorbed dose Rearranged MIRD formula energy imparted D(rkrh) = Ãhii φi(rkrh) / mk Slides are not to be reproduced without permission of author

17. Absorbed dose Rearranged MIRD formula organ mass D(rkrh) = Ãhii φi(rkrh) / mk Slides are not to be reproduced without permission of author

18. Absorbed dose D(rkrh) = Ãhii φi(rkrh) / mk patient-specific quantitative imaging energy transport and deposition patient organ CT imaging specific absorbed fractions Slides are not to be reproduced without permission of author

19. Absorbed dose D(rkrh) = Ãhii φi(rkrh) / mk patient-specific quantitative imaging energy transportand deposition patient organ mass by CT imaging specific absorbed fractions Three items to focus on for “personalized dosimetry” Slides are not to be reproduced without permission of author

20. Dosimetry is not a “precise science” • It is not possible to determine exact dose values for any tissue in the intact patient • Uncertainties: - Applying a mathematical model to represent the actual size and mass of the patient and internal organs: ± 20 to 60% • Measurement of activity in a source organ: ± 100% • Total error: ± 0 to 300% (ICRP-53) - Probable goodness of an organ dose estimate: ± 30% • The quality and accuracy of the direct measurements are the most important aspect of internal dose assessment Slides are not to be reproduced without permission of author

21. Model and reality Gastrointestinal tract Pelvic CT slice Slides are not to be reproduced without permission of author

22. Limitations • Requires well-trained people who understand the appropriate application and factors that can lead to errors - Common problem: confusing “biological” and “effective” • Mismatch between model and patient • Input and transposition errors • Chronic problem with software errors and validation Slides are not to be reproduced without permission of author

23. Relative improvement in dosimetry bias and precision Slides are not to be reproduced without permission of author

24. Phantom development and radiation transport • Tomographic, voxel, and hybrid models provide measured improvements in ability to calculate the self-organ and cross-organ gamma component • Not yet scalable or patient-specific • Revised S values not yet available MIRD “stylized” Realistic Slides are not to be reproduced without permission of author

25. Slides are not to be reproduced without permission of author Newborn model courtesy Wes Bolch, University of Florida

26. Correcting for organ mass • A correction may be made for patient-specific organ masses that differ from the standard values (OLINDA-EXM) • This correction is appropriate when most of the organ dose is due to non-penetrating radiations • Alternative approach: multiply the source organ residence time by the ratio of the MIRD organ mass to the actual organ mass patient (h) (mMIRD/mactual) Rajendran et al., 2004. J. Nucl. Med. 45(6):1059-1064. Slides are not to be reproduced without permission of author

27. Red marrow - Sensitive to ionizing radiation - Target tissue for certain radionuclide treatments (ablation therapy) - Often the dose-limiting normal tissue in radioimmunotherapy • In some protocols, marrow dosimetry is essential for treatment planning and evaluating biological response after treatment Slides are not to be reproduced without permission of author

28. Marrow dosimetry challenges • The geometry of red marrow structure with respect to bony spicules is difficult to describe and model • The amount of radionuclide in marrow is difficult to measure and quantify • Marrow is also irradiated by activity in circulating blood • High-dose irradiation of marrow may alter the cellularity of marrow and therefore marrow composition and volume Slides are not to be reproduced without permission of author

29. Approaches to marrow dosimetry • Make no direct measurements of marrow and assume that red marrow is part of the “remainder” tissues • Measure blood serum radioactivity, obtain the hematocrit, and assume that marrow uptake is proportional to activity in circulating blood (AAPM Task Group) • Measure the blood serum radioactivity (Sgouros method) • Direct imaging of a marrow space - repetitive counts over the same region (acetabulum, sacrum, pelvis, femur) - biopsy specimen (weighed and counted) Slides are not to be reproduced without permission of author

30. Sacral marrow scintigraphy • Obtain sacral marrow region-of-interest • Determine the time-activity curve • Assume that the sacral marrow = 9.9% of total-body marrow • Advantage: may be more accurate than serum activity counting • Disadvantage: activity in the sacrum is not always distinguishable from background (Siegel et al., 1989) Slides are not to be reproduced without permission of author

31. Direct measurement with calibration • Serial gamma camera (a/p) imaging to determine the marrow time-activity curves • Bone marrow biopsies obtained at 24-h post-infusion to determine %ID/g, to normalize the marrow time-activity curve Anterior view Posterior view Iliac Crest Sacrum Slides are not to be reproduced without permission of author

32. Normalizing a time-activity curve Plot the iliac or sacral direct measurements to a time-activity curve Correct the needle biopsy measurement for bone, blood, and fatty marrow Normalize the curve to the biopsy measurement normalize Slides are not to be reproduced without permission of author

33. Dose and biological response • Radiation dose correlates with biological response • The correlation always improves with ability to assess with accuracy both dose and end point Kidney dose vs. renal damage Wessels et al, JNM 49:1884; 2008 Slides are not to be reproduced without permission of author

34. Challenges • Uncertainties in quantitative (direct) imaging • Variability in red marrow volumes for different regions of the skeleton in a patient • Accuracy of biopsy specimen measurements • Variable fluid, cellular, fat, and bone contents • Small sample size • Marrow activity in the rib cage may interfere with lung activity assessment Slides are not to be reproduced without permission of author

35. Summary • Personalized dosimetry is important for selecting appropriate therapy amounts • based on absorbed dose to the normal limiting organ • Most important efforts in dose improvement will be: • Quantitative imaging for major normal organs • CT organ volumetrics • Red marrow imaging calibrated by biopsy (activity/gram) Slides are not to be reproduced without permission of author