Renal Dosimetry: Ready for BED? Response to the Italian Researchers

The authors state that models are not used robotically and that researchers are aware that refinements are compulsory steps for reliability. Unfortunately, this has not been true in practice. For example, the same values for the α/β ratio and repair half-time have been used by different investigators. We have shown that these assigned values make them mathematically meaningless in the biological effective dose (BED) model. Furthermore, because of this particular selection of parameter values, the time–dose fractionation (TDF) and BED models are mathematically closely related. Modification of these important parameters would, of course, alter BED values.

According to the authors, “The BED concept enables researchers to distinguish the tolerability related to different fractionation schemes and to highlight such a sparing effect. This task is assigned to the dose rate and also to a repair parameter unlike the time–dose fractionation." ¹ Apparently, the authors misunderstood our original article and our subsequent Letters to the Editor. We illustrated that the BED model currently used is insufficient for routine clinical use in peptide radionuclide receptor therapy (PRRT) and requires more careful crafting to fit a diverse patient population with and without risk factors better. The choice of using the same assigned values for the two fixed parameters of the BED model not only places a restriction on the overall predictive accuracy of the model, requiring a deep understanding of the related caveats, but its mathematical basis is also inadequate to provide clinically meaningful guidance. In fact, in its current incarnation, TDF is as good as BED in its ability to distinguish tolerability for any fractionation scheme likely to be encountered in PRRT. We also estimated a mean kidney dose per treatment cycle of 8.6 Gy (range: 3.0–28.6 Gy), using the combined data of Barone et al. and Bodei et al. as presented in our publication. Based on an effective half-time of 30 hours for the radioactivity in the kidney, the resulting dose rates are low, typically much less than 1 Gy/hour, and may cause less renal toxicity than that expected with conventional external beam fractionation. ³

We are disappointed that the authors seemed to misinterpret our intent. We never stated, and certainly do not believe, that “dosimetric estimates utilized in clinical trials to predict renal toxicity in patients have no substantiation” ¹ and we apologize for any misunderstanding we may have caused. We provided an objective critique of the analysis given in Medical Internal Radiation Dose (MIRD) Pamphlet 20⁴ and reiterate that, it is by no means clear which absorbed dose and radiobiological models to use, how to assign BED parameters values appropriately, and, perhaps even more importantly, how to modify a patient's treatment plan effectively in the presence of certain reported preexisting conditions (i.e., to lower dose enough to limit toxicity but not enough to compromise efficacy), as radiation dose by itself or even dose coupled to a radiobiological model may be unable to predict renal toxicity accurately in such patients. An important quantitative analysis of normal tissue effects in the clinic (QUANTEC) article published in 2010⁵ identified these same conditions as well as additional patient- and treatment-related factors as being responsible for enhancing radiotherapy-associated kidney injury (i.e., reducing the kidney's tolerance to radiation): chemotherapy: underlying renal insufficiency; diabetes; hypertension; liver disease; heart disease; and smoking. However, the magnitude of these effects was reported as unclear, indicating, as we have suggested, the need for further study to appropriately include pretherapy kidney radiosensitivity in the PRRT treatment paradigm.