Radiation Protection Dosimetry Vol. 105, No. 1 4, pp. 581 586 (2003) Published by Nuclear Technology Publishing 2003 Nuclear Technology Publishing ASSESSMENTS FOR HIGH DOSE RADIONUCLIDE THERAPY TREATMENT PLANNING* D. R. Fisher Pacific Northwest National Laboratory 902 Battelle Boulevard Richland, WA 99352, USA Abstract Advances in the biotechnology of cell specific targeting of cancer and the increased number of clinical trials involving treatment of cancer patients with radiolabelled antibodies, peptides and similar delivery vehicles have led to an increase in the number of high dose radionuclide therapy procedures. Optimised radionuclide therapy for cancer treatment is based on the concept of absorbed dose to the dose limiting normal organ or tissue. The limiting normal tissue is often the red marrow, but it may sometimes be the lungs, liver, intestinal tract or kidneys. Appropriate treatment planning requires assessment of radiation dose to several internal organs and tissues, and usually involves biodistribution studies in the patient using a tracer amount of radionuclide bound to the targeting agent and imaged at sequential timepoints using a planar gamma camera. Time activity curves are developed from the imaging data for the major organ tissues of concern, for the whole body and sometimes for selected tumours. Patient specific factors often require that dose estimates be customised for each patient. In the United States, the Food and Drug Administration regulates the experimental use of investigational new drugs and requires reasonable calculation of radiation absorbed dose to the whole body and to critical organs using the methods prescribed by the Medical Internal Radiation Dose (MIRD) Committee of the Society of Nuclear Medicine. Review of high dose studies shows that some are conducted with minimal dosimetry, that the marrow dose is difficult to establish and is subject to large uncertainties. Despite the general availability of software, internal dosimetry methods often seem to be inconsistent from one clinical centre to another. INTRODUCTION In recent years, medical internal dosimetry has undergone a shift in focus from dosimetry for safety assessment in diagnostic imaging to dosimetry for treatment planning in high dose radionuclide therapy. Developments in the biotechnology of cell specific targeting of cancer and the increased number of clinical trials involving treatment of cancer patients with radiolabelled monoclonal antibodies, peptides and similar delivery vehicles have led to an increase in the number of high dose radionuclide therapy procedures in the United States and worldwide. High dose radionuclide therapies We have entered the age of targeted radiation therapy. In February 2002, the US Food and Drug Administration (FDA) approved the first new radioimmunotherapy drug (Zevalin, or ibritumomab tiuxetan, IDEC Pharmaceuticals) for treatment of relapsed or refractory low grade, follicular or transformed B cell non-hodgkin s lymphoma. Zevalin is the monoclonal antibody 2B8 Rituxan (rituximab), which targets the CD20 antigen on B lymphocytes and is covalently linked to the therapeutic radionuclide 90 Y. New studies are under way Contact author E-mail: dr.fisher pnl.gov Prepared for the U.S. Department of Energy under Contract DE-AC06-76RL01830. Pacific Northwest National Laboratory is operated by Battelle for the Department of Energy, and cleared as document PNNL-SA-37552. to investigate the efficacy of Zevalin and other radioimmunotherapy agents in combination with standard chemotherapy agents and autologous bone marrow stem cell support (1). In related studies over the past decade, 131 I tositumomab, an anti-cd20 monoclonal antibody, has been used to treat recurrent non-hodgkin s lymphoma in combination with chemotherapy (etoposide, cyclophosphamide) and autologous stem cell transplantation (2). Numerous other studies may be cited: 131 I-BC8, an anti-cd45 hybridoma, has been used to treat advanced acute leukaemia in combination with cyclophosphamide, 12 Gy external whole-body gamma radiation and matched allogeneic or autologous bone marrow transplantation (3). 90 Y-C2B8, an anti-cd20 chimeric IgG1 antibody, has been used in pretargeted radioimmunotherapy of patients with non-hodgkin s lymphoma (4). Pre-targeted 90 Y-NRLU-10 antibody has been used to treat metastatic colon cancer and other solid tumour malignancies (5). 166 Ho labelled DOTMP (phosphonate) has been used to treat multiple myeloma in combination with melphalan chemotherapy and allogeneic stem cell transplantation (6 8). Each of these approaches to cancer treatment requires dose assessment for treatment planning. The amounts of radioactivity that can be administered are relatively high compared to low dose therapies and the amounts that are normally administered in diagnostic nuclear medicine. For example, therapies may involve up to 52 GBq (1.4 Ci) 131 I peptide in treating bronchoalveolar carcinoma, 9.2 GBq (250 mci) 90 Y biotin in treating primary liver carcinoma, 31.5 GBq (850 mci) 131 I in treating non-hodgkin s lymphoma, 581
and up to 100 GBq (2.7 Ci) 166 Ho in treating multiple myeloma. Dose limiting normal tissues Radiation dose and toxicity to the red marrow limit the amount of a radiolabelled agent that can be safely administered without stem cell support or bone marrow transplantation. With stem cell support, the lungs (2), liver (3), intestinal tract mucosa (5) or kidneys (6 8) have become the limiting normal tissues in the above studies. The upper limits on normal tissue toxicity were determined by dose escalation studies. In lymphoma clinical trials, the normal limiting organ is usually the lungs, which can safely receive up to 25 27 Gy (2) in addition to chemotherapy. Patients with advanced lymphoma have been treated with 131 I antibody, with infusions approaching 31 GBq (850 mci) and with lung doses limited to about 27 Gy. Tumour doses are not usually determined, but prior studies have shown that lymphomas received an estimated 30 120 Gy from 131 I antibody, with a median of about 46 Gy. In leukaemia clinical trials, the normal limiting organ is usually the liver, which can safely receive up to 10.5 Gy delivered by 131 I labelled antibody in addition to 12 Gy total body gamma radiation and chemotherapy (3). Patients with advanced acute leukaemia and myelodysplastic syndrome have been treated with 131 I antibody, with infusions approaching 11 GBq (300 mci) to limit the liver dose to 8 Gy. In these studies, the red marrow, which contains leukaemic cells and constitutes the treatment target tissue, typically receives a radiation absorbed dose of 1.5 2 131 I antibody, with a median of about 1.7 Gy. Among these and other high radiation dose radionuclide therapies, several common features emerge: (1) targeted radiotherapy of cancer with systemically administered radionuclides has been effective, especially for haematopoietic cancers; (2) the radiation absorbed dose to normal organs limits the amount of radionuclide that can be safely administered to patients; and (3) radionuclide therapy is most effective when given in combination with chemotherapy agents and marrow stem cell support. In treatment planning, the upper limit on the amount of radionuclide that can be safely administered is determined by radiation dose to one or more normal tissues. This relationship requires treatment planning for optimised high dose radionuclide therapy to focus on accurate characterisation of radiation dose to the normal organs and tissues. D. R. FISHER TREATMENT PLANNING FOR SYSTEMIC RADIOTHERAPY Treatment planning requires accurate assessment of radiation dose to the major internal organs and tissues. Pre-treatment planning usually involves biodistribution studies in the patient by using a tracer amount of radionuclide bound to the targeting agent with data collected at several timepoints. The distribution of radioactivity in the body or individual source organs may be determined by sequential imaging using planar scintillation cameras or single photon emission computed tomography (SPECT). Conjugate view quantitative planar imaging with anterior and posterior measurements is the most widely used method for assessing source organ activity, because it does not require knowing the depth of the source region and does not depend on assumptions inherent in single view phantom simulations. Time activity curves are developed from the imaging data for the major organ tissues of concern, for the whole body and for selected tumours. The integral of the time activity curve is the source organ residence time. Standard methods recommended by the Medical Internal Radiation Dose Committee of the Society of Nuclear Medicine are then used to estimate internal doses from administered radionuclides (9,10). Data acquisition The following data are usually needed to plan treatment for high dose radioimmunotherapy with 131 I labelled antibodies. (1) Organ and tumour mass. Organ volumes and tumour masses are calculated from chest and abdominal computed tomography or magnetic resonance scans, if available. Mass is an important item of information because organ dose is an approximate inverse function of organ mass. Information about the patient s weight is also important to correctly evaluate the whole-body dose. (2) Organ uptake, retention and clearance. A trace amount (185 370 MBq, or 5 10 mci) of 131 I antibody is administered intraveneously to determine the biodistribution and pharmacokinetics of the tracer. Patients are imaged using a large field of view, dual head, planar gamma camera with high energy collimator and detection window set at 364 kev ( 15% full width at half maximum). Quantitative images of the major organs, red marrow, tumours (selected cases) and the whole body are obtained immediately after infusion and at approximately 4, 8, and 24 h, and daily for 2 7 d. Following image acquisition, regions of interest are drawn by the investigator for the major source organs. The geometric means of the anterior and posterior counts are determined and corrected for background, attenuation and scatter using standard techniques (11). The counts are converted to fraction of administered activity by comparison to a counting standard. (3) Whole-body retention and clearance. Whole-body retention is determined using serial scans of the patient using the gamma camera, or using an 131 I 582
ASSESSMENTS FOR RADIONUCLIDE TREATMENT PLANNING uptake probe. Whole-body retention measurements are essential for evaluating the remainder tissue source organ component for internal dose calculations. (4) Blood clearance. Blood samples may be obtained hourly during antibody infusion and at 30, 60, 90 and 120 min thereafter, and daily for 3 5 d. Blood counts are used to evaluate the antibody pharmacokinetics and sometimes for non-specific antibody to infer the absorbed dose to red marrow. (5) Marrow biopsy. When the antibody targets red marrow (in leukaemia treatments), bone marrow biopsies may be obtained from two locations (right and left acetabulum) at 16 20 h post-infusion, and are weighed and counted against a weighed reference aliquot to calculate the fraction of the administered activity per gram. The marrow clearance curves (from gamma camera imaging) are scaled quantitatively using the bone marrow biopsy measurement values. (6) Urinary excretion. Urine samples are collected to determine the cumulative excretion of radioactivity via the kidneys and bladder for dosimetry of the bladder wall. (7) Intestinal tract clearance. Images of the intestinal tract are obtained and activity is quantified when there is evidence of significant excretion of radioactivity from the liver to bile to small intestines, or from the stomach into the small intestines. Sampling times Selecting an appropriate number of counting times requires a trade-off between the desire for sufficient data and the need to minimise overall costs and inconvenience to the patient. Thus, it is desirable to select the fewest timepoints that will provide a reasonable description of the time activity curve. The minimum number of data measurement points is typically four, and the optimum is five, including one measurement at or near the zero timepoint (or time of radionuclide infusion). Other timepoints may be selected from an estimate of the effective clearance half-time (T eff ) of the radionuclide in the source organ. For example, the second measurement may be obtained at a small fraction of the effective half-time after time zero (such as 1/12 T eff ). Additional measurements may be obtained at multiples of the effective half-time, such as 0.5, 1.0 and 3.0. For example, if T eff 48 h, then counts may be obtained at about 0 h, 1/12 T eff 4h,. T eff 24 h, 1 T eff 48 h, and at 3 T eff 144 h. Time activity curves The time activity curves for each source organ, tumour tissue, the whole body and remainder tissue are plotted. The areas under the time activity curves are integrated to infinity for accuracy and simplicity. The long-term tail of the exponential may be estimated by fitting to an exponential function. The data used for this fit should be the last two or more measurements, so long as the fit to an exponential remains strong (having a correlation coefficient greater than 0.95). The area under the time activity curve is the residence time (, h) used in MIRDOSE3 computer software (12) for dose calculations. Patient specific factors Actual patient weights and organ sizes may vary considerably from those used in the standard anthropomorphic dosimetry models. Since organ dose is approximately proportional to the inverse of target mass, corrections should be made when patient specific data are available. Actual patient weight, organ mass, tumour mass and other patient specific factors such as urinary excretion rate, cumulative excretion, blood clearance and biopsy values should be applied as appropriate in each dose assessment. Since the MIRDOSE3 software (12) does not accommodate changes in patient specific factors, they must be accounted for in other ways as part of dose assessment. One way to account for actual organ mass is to recalculate the S value for a source target organ pair using a Monte Carlo computer program. A less accurate, but perhaps more convenient, method for correcting for organ mass may be made by multiplying the calculated source organ residence time, h, by the ratio of the reference man or reference woman organ mass to the known organ mass: new ( h )(m MIRD /m actual ), where new is a patient specific residence time. This method corrects the beta radiation component but does not correct for the less important gamma component. For most radionuclides, the beta self-irradiation dose in a source organ is the greater contributor to total organ dose (usually more than 90% of the total). This correction may be appropriate when most of the organ dose is due to non-penetrating radiation. The new residence time for each source organ and for the remainder tissues may then be entered into MIRDOSE3 to estimate normal organ and whole-body doses (in rad mci 1 or mgy MBq 1 administered) for the patient. This correction is accurate for pure beta particle emitting radionuclides, such as 90 Y and 32 P. It does not correct for the gamma component. It can be used for radionuclides such as 131 I and 186 Re, which are beta/gamma emitting radionuclides, because most of the energy imparted to the source organ is from the beta particles. Dose calculations In the United States, the Food and Drug Administration regulates the experimental use of investigational new radioactive drugs. This agency requires assessment 583
of uptake and retention for all organs and tissues that image above a standard tissue background based on review of patient images. Also required is a reasonable calculation of radiation absorbed dose to the whole body and to critical organs using the methods prescribed by the Medical Internal Radiation Dose (MIRD) Committee of the Society of Nuclear Medicine, and the use of implementing software, such as MIRDOSE3 (12). However, the FDA has not officially approved any software for internal dose treatment planning. Approaches to red marrow dosimetry Red marrow is often the critical limiting tissue in radionuclide therapy. Assessing accurate dose to marrow is one of the most difficult challenges in dosimetry because the total marrow mass is not known for individual patients, and the activity concentration in marrow is difficult to determine. Marrow may receive significant irradiation from radioactivity that circulates in blood. High dose irradiation of marrow may alter the cellularity of marrow and thus marrow structure and composition. The result of high dose irradiation may be a change in marrow concentration within the marrow cavities. The marrow may have also been affected by prior chemotherapy or radiotherapy. The individual marrow status and marrow reserve from prior therapy may confound the expected correlation between marrow dose and the effects of irradiation during therapy. Several different approaches have been used to estimate absorbed doses to the red marrow. In one common but relatively inaccurate approach, the concentration of radioactivity in red marrow is merely assumed to be equal to that in the remainder tissues. Measurements are only made of radioactivity in the major source organs and in the whole body, and no specific residence time is assigned to the marrow. A second approach involves direct measurements of activity in blood serum at multiple timepoints; the assumptions are made that marrow uptake of radioactivity is proportional to the activity in circulating blood, that the clearance half-time for marrow is equal to that in blood serum, and that the spatial distribution of the activity in marrow is uniform. The patient s haematocrit is factored into the calculation, and a residence time for marrow is calculated from the plasma serum concentration. A third approach to bone marrow dosimetry involves repetitive direct measurement of activity in a constant marrow region (such as the acetabulum, sacrum, pelvis or femur) over time. These measurements provide data for evaluating the time activity curve. The time activity curve may be normalised to a known mass of tissue. For example, the sacral marrow may be assumed to contain about 10% of total body marrow. The disadvantage of direct marrow imaging is that the activity in the sacrum or other regions of interest may not always be distinguishable from the remainder tissue background count rates. Alternatively, a marrow biopsy may be obtained, D. R. FISHER weighed and counted, and the data point for this measurement may be normalised to a marrow time activity curve. It is also difficult to measure and interpret the biological response of red marrow to high dose radionuclide therapies. The treatment plan The therapeutic index is the ratio of the dose to tumour divided by the dose to the normal limiting tissue. The amount of a radiolabelled therapeutic agent that may be administered to a patient is limited by the maximum tolerable radiation dose to the limiting normal tissue. The treatment plan is not determined by the tumour dose, but the plan usually requires a therapeutic index greater than 1.0, meaning that the tumour will receive more radiation dose than the limiting normal organ. The treatment dosage is determined by dividing the upper limit on normal dose by the calculated radiation dose per unit administered activity. UNCERTAINTIES IN ABSORBED DOSE CALCULATIONS The validity of an internal dose assessment depends on the quality and accuracy of the measurement data and on the way in which these data are used to calculate internal dose. The author acknowledges uncertainties inherent in methods and models, patient variability and measurement difficulties. The two major sources of uncertainty in treatment planning arise from (1) measurements to determine the amount of activity in each of the source organs; and (2) extension of simplified MIRD anthropomorphic phantoms to living subjects. Direct measurements are always subject to analytical errors that may arise from efforts to measure overlapping and irregularly shaped organs, attenuation correction, scatter correction, measurement of patient thickness and background correction factors. The likely error associated with direct measurements of activity in patients organs can be as great as about a factor of 2, or about 100% (10). Anthropomorphic models used in the MIRD schema only crudely represent the human physical and anatomical form. The likely error associated with use of a mathematical construct such as the MIRD phantom to represent the actual size and mass of the patient and internal organ is about 20 to 60% (10). Given the above, the total uncertainty in an organ absorbed dose estimate is likely to be about a factor of 3 ( 300%). When measurement techniques are optimised and dosimetry calculations are customised to actual patient size and organ weight, the overall uncertainty of an organ dose estimate can be reduced to about 30% (13). 584
ASSESSMENTS FOR RADIONUCLIDE TREATMENT PLANNING CONSISTENCY FROM ONE MEDICAL CENTRE TO ANOTHER The MIRD schema and implementing software provide standard tools for dose assessment and treatment planning. The purpose of these schema is to provide a common, consistent approach to dose assessment. However, inherent differences in dose estimates that may arise from measurements, the differences in methods used by different investigators and the differences in assumptions used may lead to inconsistent application of the MIRD schema and implementing software from one clinical centre to another. This author has casually observed varying levels of commitment, priority and expertise, and differences in (1) data acquisition methods, equipment and analysis software; (2) anatomical models; (3) assumptions and methods of data analysis; (4) software tools; and (5) approaches to organ level dosimetry, such as the several different methods and S values (the mean absorbed dose per unit cumulative activity) used to estimate red marrow dose. These differences in dose assessment often lead to inconsistencies in the interpretation of therapeutic outcomes for both normal tissue toxicity and for tumour response. SUMMARY AND CONCLUSIONS We have entered the exciting new era of molecular nuclear medicine, with a shift in dosimetry focus from REFERENCES diagnostic radipharmaceutical safety to high dose radionuclide therapy treatment planning. Treatment planning for high dose radionuclide therapies may be based on an assessment of radiation absorbed dose to a critical, dose limiting normal tissue, which may be the red marrow, lungs, liver, kidneys, intestinal tract or other specifically targeted normal tissues. Treatment planning requires a dedicated effort to obtain essential data on the pharmacological behaviour of the radionuclide and carrier, prior to infusion, to ensure a favourable biodistribution and therapeutic index, and to minimise dosimetric uncertainties. Radiation doses to red marrow are difficult to establish and are subject to both large uncertainties in dose and radiobiological assessment of dose response. In some cases, high dose patient studies are conducted with minimal or insufficient dosimetry. Dosimetry methods and results obtained vary widely across centres. A potential solution may be to establish internal dose intercomparison studies using standard datasets. ACKNOWLEDGEMENTS This work was supported by the National Institutes of Health grant PO1-CA44991 and by NeoRx Corporation (Seattle, WA). Participation at this Workshop was made possible by grants from the Department of Energy and from Antisoma (London, UK). 1. Raubitschek, A. Phase I/II trial of escalating Zevalin in combination with high-dose etoposide and cyclophospamide followed by autologous stem cell transplant (ASCT) for patients with poor risk/relapsed B-cell NHL. Zevalin Investigators Newsletter 1(1), 1 (IDEC Pharmaceuticals, 1 August 2002). 2. Press, O. W., and 18 others. A phase I/II trial of iodine-131-tositumomab (anti-cd20), etoposide, cyclophosphamide, and autologous stem cell transplantion for relapsed B-cell lymphomas. Blood 96, 2934 2941 (2000). 3. Matthews, D. C. and 10 others. Phase I study of I-131-anti-CD45 antibody plus cyclophosphamide and total body irradiation for advanced acute leukemia and myelodysplastic syndrome. Blood 94(4), 1237 1247 (1999). 4. Weiden, P. L., Breitz, H. B., Press, O., Appelbaum, J. W., Bryan, J. K., Gaffigan, S., Stone, D., Axworthy, D., Fisher, D. R. and Reno, J. Pretargeted radioimmunotherapy (PRIT ) for treatment of non-hodgkin s lymphoma (NHL): initial phase I/II study results. Cancer Biother. Radiopharm. 15(1), 15 29 (2000). 5. Knox, S. J., and 13 others. Phase II trial of yttrium-90-dota-biotin pretargeted by NR-LU-10 antibody/streptavidin in patients with metastatic colon cancer. Clin Cancer Res. 6, 406 414 (2000). 6. Eary, J., Rajendran, J., Bensinger, W. I., Girault, S., Champlin, R., Thoelke, K. and Bryan, J. K. Holmium-166 with melphalan and total body irradiation (TBI) as a preparative regimen for autologous stem cell transplantation. Proc. Chemotherapy Foundation Symposium XIX, 8 11 November 2000, New York, NY. 7. Bayouth, J., Macey, D. and Kasi, L. Pharmacokinetics, dosimetry and toxicity of holmium-166 DOTMP for bone marrow ablation in multiple myeloma. J. Nucl. Med. 36, 730 (1995). 8. Bensinger, W., Giralt, S. and Eary, J. Phase I/II study of holmium-166-dotmp in combination with melphalan and total body irradiation with autologous peripheral blood stem cell transplant for patients with multiple Myeloma. Proc. Am. Soc. Clin. Oncol. 19(26), 9a (2000). 9. Loevinger, R., Budinger, T. F. and Watson, E. E. MIRD primer for absorbed dose calculations. ISBN 0-932004-38-5, rev. ed. (New York: The Society of Nuclear Medicine) (1991). 10. ICRP. Radiation dose to patients from radiopharmaceuticals. ICRP Publication 53. Ann. ICRP 18(1 4) (Oxford: Pergamon) (1988). 585
D. R. FISHER 11. Siegel, J. A., Thomas, S. R. and Stubbs, J. B. MIRD pamphlet no 16: techniques for quantitative radiopharmaceutical biodistribution data acquisition and analysis for use in human radiation dose estimates. J. Nucl. Med. 40, 37S 61S (1999). 12. Stabin, M. G. MIRDOSE: personal computer software for internal dose assessment in nuclear medicine. J. Nucl. Med. 37, 538 546 (1996). 13. Fisher, D. R. Internal dosimetry for systemic radiation therapy. Semin. Radiat. Oncol. 10(2), 123 132 (2000). 586