Prediction of [177Lu]Lu-PSMA-617 treatment response with [68Ga]Ga-PSMA-11 PET-derived variables and correlation of the overall survival with serum PSA and PSMA PET-based response criteria.

[PURPOSE] In this study we aimed to investigate prediction of response with [68Ga]Ga-PSMA PET-derived parameters and the effectiveness of RECIP criteria and PSA response in predicting survival after [177Lu]Lu-PSMA therapy.

[METHODS] Patients who received at least 2 cycles of [177Lu]Lu-PSMA therapy were retrospectively included in the study. All lesions in the PET images were segmented and PSMA PET-based parameters were extracted. The treatment response was evaluated according to RECIP and PSA criteria. ROC analyses were performed to assess their discriminatory power in predicting response. Correlation of overall survival with RECIP and PSA-based response was evaluated using Kaplan-Meier survival analysis.

[RESULTS] A total of 41 patients were included in the study. SUVmax, SUVpeak, SUVmean, qPSG-max and qPSG-mean values were significant factors in predicting PSA response. However, in predicting RECIP response (CR + PR), the AUC values for the SUV parameters were not significantly higher than 0.5. In predicting RECIP disease control (CR + PR + SD), the AUC values for SUVmax, SUVpeak, and SUVmean were 0.742(p = 0.002), 0.742(p = 0.002), and 0.722(p = 0.006). In the RECIP-responder and non-responder groups, the median survival was 33.5 and 15.2 months (p = 0.005). In patients with RECIP-disease control and progressive disease, the median survival was 25.1 and 15.2 months (p = 0.002). However, the survival difference between PSA-responder and non-responders groups did not reach significance (p = 0.127).

[CONCLUSION] PSMA PET imaging is a useful method to select patients who will most likely to benefit from [177Lu]Lu-PSMA treatment. Also, it has been shown that RECIP has significantly performed better than PSA-based response evaluation in terms of predicting overall survival.