Improving the Dose-myelotoxicity Correlation in Radiometabolic Therapy of Bone Metastases with 153Sm-EDTMP

Purpose: (153)Sm-ethylene diamine tetramethylene phosphonic acid ((153)Sm-EDTMP) is widely used to palliate pain from bone metastases, and is being studied for combination therapy beyond palliation. Conceptually, red marrow (RM) dosimetry allows myelotoxicity to be predicted, but the correlation is poor due to dosimetric uncertainty, individual sensitivity and biological effects from previous treatments. According to EANM guidelines, basic dosimetric procedures have been studied to improve the correlation between dosimetry and myelotoxicity in (153)Sm-EDTMP therapy.

Methods: RM dosimetry for 33 treatments of bone metastases from breast, prostate and lung tumours was performed prospectively (with (99m)Tc-MDP) and retrospectively, acquiring whole-body scans early and late after injection. The (153)Sm-EDTMP activity was calculated by prospective dosimetry based on measured skeletal uptake and full physical retention, with the RM absorbed dose not exceeding 3.8 Gy. Patient-specific RM mass was evaluated by scaling in terms of body weight (BW), lean body mass (LBM) and trabecular volume (TV) estimated from CT scans of the L2–L4 vertebrae. Correlations with toxicity were determined in a selected subgroup of 27 patients, in which a better correlation between dosimetry and myelotoxicity was expected.

Results: Skeletal uptakes of (99m)Tc and (153)Sm (Tc% and Sm%) were well correlated. The median Sm% was higher in prostate cancer (75.3 %) than in lung (60.5%, p = 0.005) or breast (60.8%, p = 0.008). PLT and WBC nadirs were not correlated with administered activity, but were weakly correlated with uncorrected RM absorbed doses, and the correlation improved after rescaling in terms of BW, LBM and TV. Most patients showed transient toxicity (grade 1–3), which completely and spontaneously recovered over a few days. Using TV, RM absorbed dose was in the range 2–5 Gy, with a median of 312 cGy for PLT in patients with toxicity and 247 cGy in those with no toxicity (p = 0.019), and 312 cGy for WBC in those with toxicity and 232 cGy in those with no toxicity (p = 0.019). ROC curves confirmed the correlations, yielding toxicity absorbed dose thresholds of 265 cGy for PLT and 232 cGy for WBC.

Conclusion: The best predictor of myelotoxicity and blood cells nadir was obtained scaling the RM absorbed dose in terms of the estimated TV. It seems clear that the increase in skeletal uptake due to the presence of bone metastases and the assumption of full physical retention cause an overestimation of the RM absorbed dose. Nevertheless, an improvement of the dose–toxicity correlation is easily achievable by simple methods, also leading to possible improvement in multifactorial analyses of myelotoxicity.