Artificial Intelligence for PET Image Reconstruction.

1330 Objectives: 18F-fluciclovine and 68Ga-prostate-specific membrane antigen (PSMA) are PET radiotracers recommended for the localization of recurrent prostate cancer post-prostatectomy. We examined the positivity rates (planned secondary endpoint) of 18F-fluciclovine PET/CT and 68Ga-PSMA PET/CT in a randomized trial in patients with biochemical recurrence post-prostatectomy. Methods: The first 55 enrolled patients (accrual goal: 140) with detectable prostate-specific antigen (PSA) post-prostatectomy were randomized to undergo treatment planning based on either 18F-fluciclovine (mean dose: 9.85 ± 0.84 mCi) [Arm A] or 68Ga-PSMA (mean dose: 5.09 ± 0.15 mCi) with 20 mg Lasix given unless contraindicated [Arm B] in an ongoing prospective clinical trial. Two independent nuclear medicine physicians interpreted the PET scans by consensus. Final radiotherapy decisions were based on PET. Within the context of the trial, all lesions with abnormal radiotracer uptake, including equivocals, were treated as positive for recurrence unless proven as false positive. Positivity rates on 18F-fluciclovine and 68Ga-PSMA were compared on a whole body and region basis using chi-square or Fisher9s exact test. Results: Three patients dropped out before 18F-fluciclovine PET scanning; therefore, 52 patients were analyzed (26 patients in each arm). Median PSA was 0.28 ng/mL in the 18F-fluciclovine arm and 0.27 ng/mL in the 68Ga-PSMA arm. There was no significant difference in PSA (p = 0.96) and Gleason score at prostatectomy (p = 0.80) between groups. Nine patients had equivocal uptake in lesions: 3 prostate bed, 1 pelvic node, 1 paraaortic lymph node on 18F-fluciclovine, and 4 prostate bed on 68Ga-PSMA. No visceral lesion was detected with either modality. The overall positivity rate on 18F-fluciclovine was significantly higher than 68Ga-PSMA for whole body (96.2% vs 53.9%; p <0.01) and recurrence in prostate bed (80.8% vs 30.8%; p <0.01). Although the positivity rate on 18F-fluciclovine was higher than 68Ga-PSMA for pelvic lymph nodes, there was no significant difference (38.5% vs 19.2%; p = 0.13). Positivity rates for recurrence in extrapelvic regions (lymph nodes and bone) on 18F-fluciclovine and 68Ga-PSMA were similar (15.4% vs 19.2%; p = 1.00) (Table 1A). After stratification by PSA levels (Table 1B), the whole body positivity rate on 18F-fluciclovine compared to 68Ga-PSMA was 94.4% vs 35.3% for PSA <0.5ng/mL and 100% vs 88.9% for PSA ≥0.5 ng/mL. Conclusions: In this secondary endpoint analysis from a randomized trial for patients with biochemical recurrence post-prostatectomy, the interim positivity rate for 18F-fluciclovine was higher than 68Ga-PSMA in the prostate bed and in the pelvis. Positivity rates for distant disease were comparable. This pattern was present even at low PSA levels (<0.5 ng/mL). Further analysis in a larger cohort is recommended to confirm these findings. Change in management and failure-free survival will be reported after full accrual is reached. Funding: NCT03762759 / NIH R01CA226992; Telix Pharmaceuticals (US) Inc. provided PSMA kits to Emory University