Multi SPECT/CT-based patient specific lesion and kidney dosimetry for verification of simpler approaches for treatment planning in Lu-177 DOTATATE PRRT

Objectives: Although there is much recent interest in dosimetry-driven treatment planning in therapies such as Lu-177 peptide receptor radionuclide therapy (PRRT) the imaging requirements (at multi-time points/multi-cycles) for an individualized approach is demanding to both the patient and clinic. Here we report initial results of fully 3D multi-SPECT/CT imaging based dosimetry of lesions and critical organ (kidneys) that can be used for comparison with estimates derived from simplified imaging protocols in patients receiving Lu-177 DOTATATE for metastatic neuroendocrine tumors (NETs).

Methods: SPECT/CT imaging post-cycle 1 was performed in 5 patients at up to 4 time points between day 0 and 7 with quantitative SPECT reconstruction including partial volume. Three of the 5 patients have also completed imaging after cycle 2, thus far. Mean absorbed dose (AD) and biological effective dose (BED) were determined for 28 NETs and 16 kidneys. Up to 5 lesions ( > 2 cc in volume) per patient were defined manually on diagnostic anatomic imaging by a radiologist and applied to co-registered SPECT/CT or defined directly on the CT portion of SPECT/CT when well visualized. Kidneys were defined by semi-automatic thresholding. Our previously developed Monte Carlo (MC) dosimetry code was used to convert SPECT/CT at each time point to 3D self-dose-rate and rest-of-body (RB) dose-rate maps. Dose maps from each scan were co-registered using deformable registration based on tumor/organ mask center of masses. The registered 3D dose-rate maps then were numerically integrated to obtain dose maps using fitted (mono- or bi-exponential) time-activity functions for the target and RB. BED was calculated using the following radiobiologic parameters taken from the literature (alpha/beta = 2.6 and 10 Gy, repair half-time = 2.8 and 1.5 h for kidney and tumor, respectively). AD and BED were compared, and dose metrics determined from imaging after cycles 1 and 2 were also compared, when available. Additionally, the ADs from the above procedure requiring multi-SPECT/CTs and multi-computations of MC dose-rate maps were compared with estimates from a simplified procedure that requires just a single SPECT/CT measurement and the corresponding computation of the MC dose-rate map. For the simplified method, time-integrated activities were approximated from the one SPECT/CT measurement using the derivation of Hanscheid et al. (JNM 2018;59:75-81).

Results: With the multi-SPECT/CT MC dosimetry procedure, the per cycle median AD for lesions was 21.6 Gy (range 5.2 to 42.6 Gy) and for each kidney (average over R and L) was 2.8 Gy (range 2.2 to 3.3 Gy). The kidney effective half-life had a median value of 50.8 h (range 40.3 to 61.9 h). The difference between AD and BED was < 8% for both kidney and lesions. The inter-cycle variability in kidney AD ranged from 13% to 21% and lesion AD from -29% to 15%, with one outlier at 54%. The best agreement between single and multi SPECT/CT MC AD estimates, considering both lesions and organs, was generally achieved when ~ day 4 post-therapy was used as the single time point: using this time point, the %difference between single and multi-point AD estimates had a median value of 4% (range -6% to 16% ) for kidney. The agreement for lesions had a wider range (median 0.3%, range -26% to 45%), but was within 20% for 70% of the lesions. In general, the differences were larger for smaller lesions and lesions that were difficult to define.

Conclusions: The generally good agreement between AD estimates based on single and multi- SPECT/CT show the potential for simplifying the protocol for lesion/kidney dosimetry guided treatment in Lu-177 PRRT, but for establishing robust dose - outcome relationships the most accurate calculation may be desired. Verification of simplified approaches by comparison with more sophisticated approaches such as presented here for this limited data set, are needed in a larger cohort.