Accuracy and precision of few-time-points renal dosimetry for [177Lu]Lu-PSMA-617 therapy: Analysis with nonlinear mixed-effects modeling.

BACKGROUND

Determination of time-integrated activity (TIA) with a reduced number of imaging sessions is essential for minimizing patient burden and clinical workload in the dosimetry of [177Lu]Lu-PSMA-617. One approach to achieve this is by performing single-time-point (STP) dosimetry. However, STP dosimetry may be associated with significant accuracy errors in certain patients, potentially impacting the reliability of dose calculations. Moreover, the assessment of precision is crucial to evaluating the stability and reliability of estimated doses.

PURPOSE: This study aims to investigate the accuracy and precision of few-time-points (FTP) TIA calculation in kidneys for [177Lu]Lu-PSMA-617 using nonlinear mixed-effects modeling (NLMEM).

METHODS

Biokinetic data of kidneys from 63 patients with metastatic castration-resistant prostate cancer (mCRPC) treated with [177Lu]Lu-PSMA-617 in the first treatment cycle were used. The SPECT/CT measurement was done at time points (TP) 1) (1.8 ± 0.8) h, 2) (18.7 ± 0.9) h, 3) (42.6 ± 1.0) h, 4) (66.3 ± 0.9) h, and 5) (160.3 ± 24.2) h after injection. This study used the sum-of-exponentials function (SOEF) with six parameters, previously selected as the best fit function for the biokinetic data (PMID: 38423787). Reference TIAs (rTIAs) were derived from fitting the SOEF parameters to all-time-points (ATP) data within the NLMEM framework. Estimated TIAs (eTIAs) were calculated by fitting the FTP data, which consist of one-, two-, three-, and four-time points combinations of the biokinetic data. The accuracy of FTP-NLMEM TIA calculations was quantified using the root-mean-square error (RMSE) and mean absolute percentage error (MAPE) of the relative deviation between eTIAs and rTIAs. Precision was assessed from the coefficient of variation (CV) of individual TIA estimates at each optimal time-point combination.

RESULTS

For each optimal TP combination, the RMSEs and MAPEs were (11.0 ± 2.5)% and (7.0 ± 2.3)% for TP3, (6.3 ± 1.6)% and (4.8 ± 1.3)% for TP25, (3.9 ± 1.1)% and (2.3 ± 1.0)% for TP135, and (1.4 ± 0.8)% and (0.9 ± 0.8)% for TP1235. The %CV values of individual TIAs for each best TP combination were (17.1 ± 4.9)% for TP3, (8.5 ± 2.4)% for TP25, (6.3 ± 0.9)% for TP135, (5.1 ± 0.6)% for TP1235, and (4.4 ± 0.3)% for ATP.

CONCLUSION

The accuracy and precision of various FTP schemes have been determined and can be used to decide on the number of measurements required. Our study showed that incorporating TP3 in FTP dosimetry could lead to a high accuracy and precision of calculated individual TIAs in [177Lu]Lu-PSMA-617therapy.