XA0

THE ROLE OF MATHEMATICAL MODELS IN THE 102668

OPTIMIZATION OF RADIOPHARMACEUTICAL THERAPY
C. DIVGI
Cornell University Medical College, New York, N.Y., United States of America

Abstract. Mathematical models have been used in radiopharmaceutical therapy for over five decades. These
have served to determine the amount of radioactivity required to treat disease, as in the therapy of
hyperthyroidism with iodine-131, or, more frequently, to determine the largest amount of radioactivity that can be
safely administered. Mathematical models are especially useful in the determination of fractionated
radiopharmaceutical therapy. This review will briefly outline the historical development and current utility of
mathematical models in radiopharmaceutical therapy, including thyroid disorders and radioimmunotherapy; and
describe the potential of modeling in fractionated therapy. The extended application of such models to currently
used radiopharmaceutical therapy based on indices of body mass or surface area, to alleviate toxicity and increase
radiation dose to tumour, will be proposed. Finally, future applications of mathematical models in
radiopharmaceutical therapy will be outlined.

REVIEW

In contrast to external beam radiotherapy, internally administered radiopharmaceuticals have

inherent characteristics that affect effectiveness in cancer therapy, and determination of dose
estimates to target and critical normal organs is therefore crucial to optimization of
radiopharmaceutical therapy.

Existing mathematical models for calculation of maximum safe dose for differentiated thyroid
carcinoma are being refined and extended to other therapies. There is a need for development of new
models that address heterogeneity in radiation dose, permit fractionated therapy, and are directed
toward optimizing radiation dose to tumour. This review will briefly address these issues.

The first application of mathematical models was in iodine-131 therapy for thyroid disorders
(hyperthyroidism and differentiated thyroid carcinoma). Initial therapy with 1-131, first used in 1942
[1], was with a fixed dose, and this continues to be an important method of treating patients [2].
However, the safety of radioiodine therapy in hyperthyroidism, and the ability to measure uptake and
retention of 131I in the thyroid gland, has led to increasing use of dosimetric models. In these models,
an estimate of gland size is obtained, as is an estimate of 131I retention in the gland. Using these
estimates, it is possible to calculate the amount of 13II required to treat the hyperthyroidism, using
either a) the microcurie per gram desired [3, 4] or b) the desired radiation dose delivered to the gland
[5]. The efficacy of such therapy is underscored by its ubiquity; most patients with thyrotoxicosis now
receive 131I as standard therapy, and surgery is no longer the standard of care. Similarly, calculated
doses of 131I have been used to ablate residual tissue in the thyroid bed after definitive surgery for
differentiated thyroid carcinoma [6].

Almost all other models of radiopharmaceutical therapy have used dosimetry to estimate the
largest safe amount of radioactivity that can be administered. Again, radioiodine therapy of thyroid
carcinoma remains the paradigm [5, 6, 7]. As in 131I therapy for hyperthyroidism, a small amount of
131
I is first administered, and serum and whole body clearance measurements used to calculate the
largest safe (i.e. non-myelotoxic) amount of radioactivity that can be administered. Using this
approach, developed at this Center, large doses of 131I can be administered without significant
myelotoxicity [6], However, this approach does not take into account radiation dose to tumour;
Maxon et al have shown the clear relationship between tumour response and radiation dose delivered
to tumour [8].

In sharp contrast to 13iI therapy for thyroid disorders, radiopharmaceutical therapy for palliation
of bone pain has not followed any dosimetric modeling. In patients receiving 89Sr, it was not possible

51
to carry out whole body measurements (and, intuitively, serum measurements would not be
representative of marrow dosimetry since strontium is a bone-seeker), because of the lack of a suitable
radioactive emission. Consequently, the maximum tolerated dose of 89Sr was arrived through
escalation of the absolute amounts of radioactivity, rather than by radiation dose estimates [9]. It is
rather unfortunate that this pattern was continued in the evaluation of other, gamma-emitting,
radiopharmaceuticals used in the palliation of bone pain [10, 11]. The contiguity and lack of clear
demarcation between the target and the critical normal organ are among the constraints that probably
account for this feature.

Radioimmunotherapy trials also followed a methodology that involved escalation of radioactive

amounts. Early trials [12, 13, 14] were carried out with murine antibody, precluding multiple
infusions that would permit determination of whole body and serum dose estimates. Retrospective
evaluation of a large body of data has made it clear that, while most radioimmunotherapy trials
determined that the MTD was between 60 and 75 mCi/m2 131I in most solid tumours, the most
important predictor of hematopoietic toxicity is the radiation dose to the whole body and red marrow
[15].

Patients with B-cell lymphoma constitute an exception to the rule that murine antibodies are
immunogenic. It has thus become possible to carry out dosimetric estimates in patients receiving
radioimmunotherapy with 131I-labelled antibody; this has allowed the determination of a maximum
tolerated dose (75 cGy) in patients receiving mI-anti-CD20 murine antibody with marrow rescue [16].
It has also allowed, in patients for whom marrow rescue is planned and thus for whom the critical
organ would be other than the marrow, to determine the MTD to that critical organ and treat
appropriately [17].

The development of relatively non-immunogenic antibody forms has made possible the use of
dosimetric models in the therapy of cancer. This has rekindled interest in several related issues: a)
how is tumour dose heterogeneity to be addressed; and b) is a single large dose of
radiopharmaceutical preferable to fractionated therapy? These questions are perhaps less relevant to
radioantibody therapy of leukemia and lymphoma; and extremely relevant to solid tumour
radioimmunotherapy. It is unclear at this point what the optimum therapy schedule would be for bone
metastases. The remainder of this review will focus on the role of mathematical models in
radioimmunotherapy.

We have previously shown that antibody delivery to renal carcinoma is heterogeneous, despite
homogeneous antigen distribution [18]. Various factors preclude successful targeting of radioantibody
throughout a solid tumour, among them changes in vascular flow [19], interstitial pressure, and
antigen heterogeneity. There is evidence that multiple administrations of radioantibody may result in
more homogeneous distribution of radioantibody in tumour [20, 21].

Joseph O'Donoghue has constructed elegant mathematical models to study the effect of
heterogeneity upon radiation dose. He has determined that dose heterogeneity results in inefficient
tumouricidal effect, with the efficacy being inversely proportional to both the mean dose in, and the
radiosensitivity of, the tumour.

The experimental data demonstrating the relative effectiveness of fractionated

radioimmunotherapy to single large dose radioimmunotherapy led us to initiate a clinical trial in
metastatic renal carcinoma. Mathematical models predicted that rapid fractionation would produce a
longer duration of remission than a single large dose, though the latter would result in greater
tumouricidal effect. This phenomenon was independent of tumour size, radiosensitivity, and other
characteristics. We have thus far enrolled 9 patients, and demonstrated proof-of-principle: prediction
of fractionation schema closely match actual patient clearance characteristics, both serum and whole
body. An important finding has been the tremendous variability in whole body and serum clearance of

52
radioantibody between patients, which underscores the importance of mathematical models for use in
therapy.

Decades of experience with thyroid cancer therapy have clearly demonstrated that the use of
mathematical modeling allows for patient-specific therapy, with consequent minimization of side-
effects. Currently, mathematical models are used primarily for determination of radiation dose to
critical organs. Better tumour quantitation, to determine tumour volume and mass; to determine
number of viable tumour cells, and to quantify amount and distribution of radiopharmaceutical in the
tumour, will permit even more tailored therapy. The future of radiopharmaceutical therapy, therefore,
is inextricably linked to the development of suitable mathematical models.

REFERENCES

[I] HERTZ, S., ROBERTS, A., "Radioactive iodine as indicator in thyroid physiology: use of
radioactive iodine in differential diagnosis of 2 types of Graves' Disease", J. Clin. Invest.
(1942)21:31.
[2] WERNER, S.C., "Radioiodine, in Werner SC, Ingbar SH (eds): The Thyroid", Harper & Row,
Hagerstown, MD (1978), pp. 827-35.
[3] IRVINE, W.J., TOFT, A.D., "The diagnosis and treatment of thyrotoxicosis", Clin. Endocrinol.
(1976)5:687.
[4] ROUDEBUSH, C.P., HOYE K.E., DEGROOT, L.J., "Compensated low-dose 131I therapy of
Graves' Disease", Ann. Intern. Med. (1977) 87: 441.
[5] JACKSON, G.L., "Calculated low dose radioiodine therapy of thyrotoxicosis", Int. J. Nucl.
Med. Biol. (1975)2:80.
[6] HURLEY, J.R., BECKER, D.V., "The use of radioiodine in the management of thyroid
cancer", In: Freeman, L.M., Weissman, H.S. (eds.): Nuclear Medicine Annual. Raven Press,
New York, (1983), p. 329.
[7] LEEPER, R., "Controversies in the treatment of thyroid cancer", The New York/Memorial
Hospital approach. In: Oppenheimer, J.H. (ed.): Thyroid Today, (1982) 5: 1-4.
[8] MAXON, H.R., THOMAS, S.R., HERTZBERG, V.S., et al., "Relation between effective
radiation dose and outcome of radioiodine therapy for thyroid carcinoma", N. Engl. J. Med.
(1983)309:937.
[9] ROBINSON, R.G., PRESTON, D.F., SCHIEFELBEIN, M., BAXTER, K.G., "Strontium 89
therapy for the palliation of pain due to osseous metastases", JAMA (1995) 274: 420^4.
[10] SERAFINI, A.N., HOUSTON, S.J., RESCHE, I., et al., "Palliation of pain associated with
metastatic bone cancer using samarium-153 lexidronam: a double-blind placebo-controlled
clinical trial", J. Clin. Oncol. (1998) 16: 1574-81.
[II] MCEWAN, A.J., "Pain palliation and nuclear medicine [editorial; comment]", Eur. J. Nucl.
Med. (1993)20: 1-3.
[12] SCHEINBERG, D.A., LOVETT, D., DIVGI, C.R., et al., "A phase I trial of monoclonal
antibody Ml95 in acute myelogenous leukemia: specific bone marrow targeting and
internalization of radionuclide", J. Clin. Oncol. (1991) 9: 478-90.
[13] WELT, S., DIVGI, C.R., KEMENY, N., et al., "Phase I/II study of iodine 131-labeled
monoclonal antibody A33 in patients with advanced colon cancer", J. Clin. Oncol. (1994) 12:
1561-71.
[14] WELT, S., SCOTT, A.M., DIVGI, C.R., KEMENY, N.E., et al., "Phase I/II study of iodine
125-labeled monoclonal antibody A33 in patients with advanced colon cancer", J. Clin. Oncol.
(1996)14:1787-97.
[15] SGOUROS, G., DIVGI, C.R., SCOTT, A.M., et al., "Hematologic toxicity in
radioimmunotherapy: An evaluation of different predictive measures", J. Nucl. Med. (1996)
43P-44P.
[16] KAMINSKI, M.S., ZASADNY, K.R., FRANCIS, I.R., et al., "Radioimmunotherapy of B-cell
lymphoma with [131I]anti-Bl (anti-CD20) antibody", N. Engl. J. Med. (1993) 329: 459-65.

53
[17] PRESS, O.W., EARY, J.F., APPELBAUM, F.R., et al., "Radiolabeled-antibody therapy of B-
cell lymphoma with autologous bone marrow support", N. Engl. J. Med. (1993) 329: 1219-24.
[18] OOSTERWIJK E., BANDER, N.H., DIVGI, C.R., et al. "Antibody localization in human renal
cell carcinoma: a phase I study of monoclonal antibody G250". J. Clin. Oncol. (1993) 11: 738—
50.
[19] JAIN, R.K. "Delivery of novel therapeutic agents in tumours: physiologic barriers and
strategies". J Natl Cancer Inst 1989; 81: 1-7.
[20] SCHLOM, J., MOLINOLO, A., SIMPSON, J.F., et al. Advantage of dose fractionation in
monoclonal antibody-targeted radioimmunotherapy. J. Natl. Cancer Inst. (1990) 82: 763-771.
[21] BUCHSBAUM, D., KHAZAELI, M.B., LIU, T., et al. "Fractionated radioimmunotherapy of
human colon carcinoma xenografts with 1311-labeled monoclonal antibody CC49". Cancer Res
(1995) 55 (Suppl): 5881-5887.

54