Abstract
It is our responsibilities, not ourselves, that should be taken seriously.
This issue of Cancer Biotherapy and Radiopharmaceuticals contains an article entitled “Renal Dosimetry in Peptide Radionuclide Receptor Therapy,” by Siegel et al. and related Letters to the Editor. The topical material is important for PRRT and, more generally, for molecular targeted radionuclide (radioisotope) therapy (MTRT), although controversial. Controversy that heightens interest is likely to generate additional progress. However, it must not obscure the great advantage of readily generating personalized pharmacokinetic and radiation dosimetric data for a patient prior to treatment by using a safe diagnostic (tracer) radiolabeled dose of the drug to be prescribed.
The authors of the article and the associated Letter to the Editor evaluated the predictive value for delayed nephrotoxicity of radiation absorbed dose and radiobiologic models, such as the biologic effective dose (BED) and time–dose-fractionation (TDF), by reanalyzing data from two series of patients given PRRT. 1,2 These and other radiobiologic models are based on extensive data for external beam and brachy radiotherapy. Siegel et al. recommend additional evaluation before relying solely on models derived from experience for external beam radiotherapy (XBRT). 3,4
We've generated this Editorial to provide background and to point out the importance of the topic of the article and letters. Although radiation absorbed dose, the total energy absorbed per unit mass of tissue, provides information critical for tissue effect, other factors, such as tissue radiosensitivity, preexisting tissue damage, and nature and dose rate of the radiation, etc., must also be considered if prediction of tissue response is to be accurate. Of a general nature, our Editorial calls attention to the importance of radiation dosimetry for drug and drug dose selection and endorses further work on methodologies for renal dosimetry and the renal radiation dose–tissue response relationship (radiobiology) for radiation delivered by MTRT. We believe the latter to be the “bottom line” of the authors of the article and related Letters to the Editor.
The Society of Nuclear Medicine Medical Internal Radiation Dosimetry (MIRD) Committee, a prestigious group of scientists, has generated publications of considerable consequence for radiation dosimetric and for radiobiologic models intended for prediction of renal toxicity in patients given PRRT. 5,6 These publications provide many noteworthy insights. As has occurred for decades, MIRD publications reflect the “state of the art” at a point in time and are updated as additional information becomes available. One of the publications describes a multiregion model that permits estimates of regional, renal radiation doses. 5 Earlier dosimetric models assumed uniform renal radionuclide distribution and may not be adequate always for drugs that are cleared by, and retained in, regions of the kidneys. Dosimetric models that assume a uniform distribution of drug/radionuclide in the kidney average the radiation absorbed dose, can misrepresent regional doses, and fail to account for the relative biologic effectiveness of a specific kind of radiation. However, multiregion models require serial, emission tomographic renal imaging for regional dose estimates. The multiregion model may be needed to refine radiation dose estimates for use in predictive radiobiologic models of renal dose–response for PRRT, although simpler approaches may prove sufficient for MTRT in many cases. Another MIRD publication endorses the multiregion model and proposes the BED radiobiologic model for predicting renal radiation toxicity after PRRT. 6 The BED model originally was developed to compare fractionation protocols for XBRT and can be considered to be the total radiation dose required for a biologic effect even when delivered at a low dose rate and in multiple fractions.
A retrospective study of a modest-sized subset of patients given 90Y-DOTA-Tyr3-octreotide for PRRT was performed by Barone et al. 1 Patient-specific kinetics of renal uptake and clearance were determined by quantitative PET of the analog drug, 86Y DOTA-Tyr3-octreotide. A stronger correlation with nephrotoxicity was observed if the BED radiobiologic model was used rather than using radiation absorbed dose alone. The investigators observed that average renal radiation dose estimates obtained by using earlier MIRD schema that assumed uniform renal radionuclide clearance were insufficient for estimation of radiation dose adsorbed by regions of the kidney because cortical and medullary regions had markedly different uptake of the drug. Both Barone et al. 1 and Bodei et al. 2 concluded that the: (1) radionuclide radiation dose–renal response data, when expressed in terms of BED, were consistent with external beam experience for predicting renal toxicity; and, (2) nephrotoxicity predictions based on both multiregion renal dosimetry and the BED radiobiologic model should be used to guide planning for trials of radionuclide therapy.
A goal for MTRT is to ensure that each patient receives a radionuclide dose sufficient to deliver a radiation dose to the cancer that is effective and without undesired normal tissue effects. PRRT, a form of MTRT that utilizes radiolabeled somatostatin analogs as drugs, is a major tool for management of patients with inoperable or metastatic neuroendocrine cancers. The dose-limiting toxicity after PRRT is delayed nephrotoxicity. 7 Although it is important that accurate dosimetry is utilized because the radiation dose estimate is a key input into any radiobiologic model, more experience is needed to evaluate models to use for prediction of renal toxicity after PRRT, and MTRT more generally. Knowledge of drug and other patient parameters are also needed to select the preferred model for accurate predictive MTRT for a patient.
Cancer treatment requires selection of the preferred drug(s) and drug dose(s) to be administered to each patient. Designs for anticancer drugs differ from designs for drugs used in other diseases. Drugs are selected for clinical trials because of activity in animal models and human cells. A goal for cancer drug trials is to complete the evaluation in as few patients as possible because of the toxicity of the drugs. Clinical trial designs for anticancer chemotherapeutic drugs commonly begin at a “safe,” if likely ineffective, dose and incorporate dose escalation in patient cohorts until a predefined fraction of a cohort has unacceptable, or unexpected serious, toxic effects from the drug. This dose, or one near it, is referred to as the maximum tolerated dose (MTD) of the drug. Although practical, these designs have serious inadequacies as well as ethical problems. Some patients get less-effective and others less-safe drug amounts by design!
Comparisons from one mammalian species to another are based on the seminal analysis of Freirich et al. 8 on the toxicities of anticancer chemotherapeutic drugs. Since the 1950s, considerable data on the relative toxicity of anticancer drugs in humans compared to animals have accumulated. Human and animal toxicology that was assembled in the late 1970s led the United States Food and Drug Administration to accept that preclinical studies in mice could be used to select a safe starting dose for trials in patients. The consistency of toxic doses across species when normalized to body surface area reflects the general relationship between surface area and physiological processes that determine drug clearance. 9
Extrapolation of toxicity from animal to a specific patient poses serious difficulties. Each patient is unique; there are substantial inter-patient differences in drug distribution and pharmacokinetics that are unrelated to patient size. Better approaches are needed to improve our ability to identify the preferred drug dose for a patient.
The article by Siegel et al. in this issue is a reminder that selection of a safe and effective drug dose can be as important as selection of the appropriate drug for the patient. Preferred doses vary widely among patients. Consideration of clinical and pharmacological data, including that obtained by using the radiolabeled drug, makes sense for prescribing anticancer drugs. Reliance on patient size as the sole determinant of dose should be considered to be absurd. 10 If a radionuclide is attached to a drug, its movement in a patient can be followed over time. Drug-based imaging has important roles in the development of a treatment plan for a patient. In MTRT, radiation dose distributions, and radiobiologic and risk factors available pretherapy for a patient, provide a distinct advantage for selection of the preferred drug and dose for a patient. Pharmacokinetics and radiation dosimetry can be used to determine drug amount. Personalized drug dosing is especially attractive; few people suggest that a drug has the same behavior in all patients.
Delayed nephrotoxicity is dose-limiting for PRRT because the kidneys are the primary route of excretion for small radionuclide conjugates and concentrate some radiometals. Drugs excreted through the kidneys may have a nonuniform renal distribution that can be associated with a nonuniform distribution of radiation dose. Renal radiation dosimetry can be the basis for simple algorithms for managing PRRT for a patient. 11,12 Radiobiologic models represent efforts to extrapolate radiation absorbed dose to a tissue biologic effect that reflects the nature of the radiation. Pretherapy assessment of renal radiation dose coupled with an appropriate radiobiologic model and existing risk factors should reduce the risk of nephrotoxicity after PRRT. At present, it is not certain which radiation absorbed dose (uniform versus multiregion) and which radiobiologic model should be used and how existing renal conditions should be used in PRRT and other MTRT planning. Although accurate dosimetry must be utilized because the radiation dose estimate is a key input into any radiobiologic model, more experience is needed to evaluate models to use for prediction of renal toxicity after PRRT, and MTRT generally. Knowledge of drug and of other patient parameters is also needed to select the preferred model for accurate predictive MTRT for a patient.
When the effective radiation renal dose and delayed nephrotoxicity can be predicted better, then radiation dose to the cancer can be intensified. 13
The current dialogue is evidence for the power of these methods and their progression in recent years. Dialogue among the experts is now related to preferred models rather than the usefulness and practicality of the methods. None of these experts deny that robust biologic data is needed for pharmacokinetic and dosimetric estimates. Although dosimetry has been of enormous value for drug development, dosimetry's value for selection of the preferred drug and drug amount for a patient requires further documentation. A purpose for dosimetry is to improve patient management; radiation absorbed dose is most useful if it can be used to predict tissue response. Radiation dose estimated for a uniform distribution of drug/radionuclide in the kidney using reference phantoms and without regard to the dose rate and spatial nonuniformities may not provide meaningful dose–response correlations. Although the largest amount of human dose–response data is from experience for XBRT, there is a substantial amount of dose–response data now available for radionuclide therapy. Radiobiologic models shown preferable for XBRT provide a “starting point” for MTRT although the latter is quite different from XBRT. Radionuclide radiation is of “long-duration,” continuous, low-dose rate, and is somewhat analogous to “seed” brachytherapy, whereas external beam radiation is of short duration and high-dose rate.
In summary, MTRT based on prescribed and patient-specific radiation dose rather than based on radionuclide dose or patient size should lead to greater realization of its potential to be effective and safe! Future data should be used to select preferred models for MTRT. Their relevance to patient management, their cost-effectiveness, etc., must be documented better. Paraphrasing others, large prospective trials are needed to document: (1) the relationship between “radiation dose” and tissue response for MTRT; (2) the relative predictive value of uniform and regional methods for estimating radiation absorbed dose and for modeling its biologic effect on the dose-limiting tissue; and, (3) the cost–benefit of radiation dosimetry. Delayed nephrotoxicity after PRRT seems to be an especially fertile area for some of these investigations for reasons evident in this issue of Cancer Biotherapy and Radiopharmaceuticals. Analogous methods used routinely for conventional radiotherapy became well-established over the past one-half century because of trials of this kind.
Members of the scientific community are encouraged to generate Letters to the Editor for publication on this topic of timely interest; they should be supported by data and references, as appropriate.