Outcome of patients with PSMA PET/CT screen failure by VISION criteria and treated with 177Lu-PSMA therapy: A multicenter retrospective analysis.

Abstract

5042 Background: The VISION trial showed that 177Lu-PSMA prolongs survival and improves the quality of life of patients with advanced mCRPC. It used PSMA-PET/CT to select patients for inclusion. The screen failure rate was “only” 12.6% (126/1003) and some have argued that the trial could have been positive even in an unselected population. It remains unknown whether the VISION-PET criteria were appropriate to stratify patients into likely responders versus non-responders to177Lu-PSMA. Here, we evaluated the outcome of patients who would have been excluded by the VISION-PET criteria and were nevertheless treated with 177Lu-PSMA. Methods: Patients treated with ≥1 cycle of 177Lu-PSMA between 11/2017 and 07/2021 (n = 74) who were registered in our institutional database and those enrolled in a multicenter dataset published previously (n = 230) (Gafita A, Lancet Oncol 2021) were retrospectively studied. One dual certified nuclear medicine and radiology reader reviewed baseline PSMA PET/CT studies and classified the patients as eligible (VISION-PET-E) or screen failure (VISION-PET-SF) based on the VISION-PET criteria: 1) absence of metastatic lesion(s) with uptake > liver background (i.e., low PSMA expression) or 2) presence of ≥ 1 metastatic lesion measurable by CT (≥ 1 cm for bone lesions with soft-tissue component (M1b) and solid/visceral organs lesions (M1c); ≥ 2.5 cm for lymph nodes lesions (N1-M1a)) with uptake ≤ liver background (i.e., PSMA-negative lesions). Outcome measures were PSA response rates (decline of ≥50% (PSA50RR), any decline (anyPSARR)), PSA-progression free-survival (PSA-PFS), and overall survival (OS). Results: 3/304 patients (1%) were excluded (lost to follow-up (n = 2), missing CT data (n = 1)). The median follow-up time was 22.5 (range: 2.1-62.3) months. 272/301 (90.4%) were VISION-PET-E, while 29/301 (9.6%) were VISION-PET-SF: 8/301 (2.7%) and 21/301 (7.0%) patients with low PSMA expression and PSMA-negative lesions, respectively. The VISION-PET-SF patients had worse PSA50RR (21% vs 50%; p = 0.005), anyPSARR (41% vs 71%; p = 0.003), median PSA PFS (2.1 months (95%CI: 1.1-3.3) vs. 4.1 (4.0-5.8); p = 0.023)), and tended to have a shorter OS (9.6 months (95%CI: 4.7-14.0) vs. 14.2 (12.6-15.9); p = 0.16) than the VISION-PET-E patients. Conclusions: VISION-PET-SF patients had worse outcomes than the VISION-PET-E patients. Our cohort only included patients who were eligible by each local site criteria. It did not include those 10-15% of patients who were excluded by local site assessment. Thus, 20-25% of the patients may be screen failures in unselected populations. Refinements in patient selection for 177Lu-PSMA are needed to optimize outcomes. Not characterizing target expression prior to PSMA-targeted treatment appears now non-ethical as a predictive whole-body imaging biomarker for response to PSMA-targeted therapies is available.