A model-based method for the prediction of whole-body absorbed dose and bone marrow toxicity for ¹⁸⁶Re-HEDP treatment of skeletal metastases from prostate cancer

In high-activity rhenium-186 hydroxyethylidene diphosphonate (¹⁸⁶Re-HEDP) treatment of bone metastatic disease from prostate cancer the dose-limiting factor is haematological toxicity. In this study, we examined the correlation of the injected activity and the whole-body absorbed dose with treatment toxicity and response. Since the best response is likely to be related to the maximum possible injected activity limited by the whole-body absorbed dose, the relationship between pre-therapy biochemical and physiological parameters and the whole-body absorbed dose was studied to derive an algorithm to predict the whole-body absorbed dose prior to injection of the radionuclide. The whole-body retention of radioactivity was measured at several time points after injection in a cohort of patients receiving activities ranging between 2,468 MBq and 5,497 MBq. The whole-body absorbed dose was calculated by fitting a sequential series of exponential phases to the whole-body time-activity data and by integrating this fit over time to obtain the whole-body cumulated activity. This was then converted to absorbed dose using the Medical Internal Radiation Dose (MIRD) committee methodology. Treatment toxicity was estimated by the relative decrease in white cell (WC) and platelet (Plt) counts after the injection of the radionuclide, and by their absolute nadir values. The criterion for a treatment response was a 50% or greater decrease in prostate-specific antigen (PSA) value lasting for 4 weeks. Alkaline phosphatase (AlkPh), chromium-51 ethylene diamine tetra-acetate (⁵¹Cr-EDTA) clearance rate and weight were measured before injection of the radionuclide. The whole-body absorbed dose showed a significant correlation with WC and Plt toxicity (P=0.005 and 0.003 for the relative decrease and P=0.006 and 0.003 for the nadir values of WC and Plt counts respectively) in a multivariate analysis which included injected activity, whole-body absorbed dose, pre-treatment WC and Plt baseline counts, PSA and AlkPh values, and the pre-treatment Soloway score. The injected activity did not show any correlation with WC or Plt toxicity, but it did correlate with PSA response (P=0.005). These results suggest that the administration of higher activities would be likely to generate a better response, but that the quantity of activity that can be administered is limited by the whole-body absorbed dose. We have derived and evaluated a model that estimates the whole-body absorbed dose on an individual patient basis prior to injection. This model uses the level of injected activity and pre-injection measurements of AlkPh, weight and ⁵¹Cr-EDTA clearance. It gave good estimates of the whole-body absorbed dose, with an average difference between predicted and measured values of 15%. Furthermore, the whole-body absorbed dose predicted using this algorithm correlated with treatment toxicity. It could therefore be used to administer levels of activity on a patient-specific basis, which would help in the optimisation of targeted radionuclide therapy. We believe that algorithms of this kind, which use pre-injection biochemical and physiological measurements, could assist in the design of escalation trials based on a toxicity-limiting whole-body absorbed dose, rather than using the more conventional activity escalation approach.