Room: Track 3
Purpose: dosimetry has been recognized as a potentially significant predictor of treatment response and toxicity for radionuclide therapy procedures. However, the typically complicated dose estimation procedure obstructs its routine implementation in clinical practice. Single-time-point SPECT/CT based estimation of absorbed dose (STP dosimetry) is an alternative route. Our aim is to compare accuracies of two STP dosimetry methods for different radiopharmaceuticals.
Methods: accuracy was estimated for two STP dosimetry methods (M1 and M2 proposed by Hanscheid and Madsen, respectively) for scan times extending from 24h to 144h following activity administration. Method M2 invokes an estimated population-based mean effective half-life (T(p-eff)), while M1 does not make use of it. Lesions and organs-at-risk doses were estimated for ¹77Lu-DOTATATE and ?°Y-DOTATOC used in neuroendocrine tumours therapies and ¹77Lu-PSMA-617 and ¹77Lu-PSMA-I&T for prostate cancer. The accuracy of STP dosimetry for different radiopharmaceuticals was computed based on the reported T(eff) values and varying STP scan times were investigated.
Results: general, when scan time was 0.8-1.6 times of T(p-eff), dose errors (%DE=100%×(Estimate/Truth)-1) for both M1 and M2 were similar (<7%). The optimal scan time for ¹77Lu-DOTATATE would be 72h after activity injection with the lowest DE for both kidneys and lesions; while, early scans (e.g. 24h) would be preferable for ¹77Lu-PSMA-compounds and ?°Y-DOTA-compounds. However, doses for bone marrow could not be reasonably estimated using STP method. Additionally, high possibility of underestimating dose was observed for scans performed shortly after activity administration.
Conclusion: dosimetry is a viable option for routine dosimetry for all radionuclide therapy cycles. However, more accurate dose estimates would be achieved if patient-specific T(eff) is determined by multiple scans in the first treatment cycle, following which STP dosimetry could be used for the following cycles.
Not Applicable / None Entered.