PD60-03: The Impact Of Histological Variants On The Performance Characteristics Of 68Ga-PSMA PET/CT In The Primary And Recurrent Setting

Simone Scuderi*, Elio Mazzone, Milano, Italy, Giorgio Gandaglia, Milan, Italy, Roberta Lucianò, Nazario Pio Tenace, Milano, Italy, Alessia Cimadamore, Ancona, Italy, Andrea Necchi, Daniele Raggi, Laura Marandino, Riccardo Leni, Daniele Robesti, Lorenzo Toneatto, Donato Cannoletta, Leonardo Quarta, Milano, Italy, Mario De Angelis, Francesco Pellegrino, Gabriele Sorce, Milan, Italy, Francesco Barletta, Milano, Italy, Luigi Nocera, Milan, Italy, Giuseppe Rosiello, Milano, Italy, Armando Stabile, Milan, Italy, Francesco Montorsi, Alberto Briganti, Milano, Italy

Introduction: PSMA PET demonstrated good accuracy for the detection of metastases of prostate cancer (PCa) both in primary and recurrent settings. No data exists on the diagnostic accuracy of PSMA PET in men with histological variants (HV) of PCa. To address this void, we tested the diagnostic accuracy of PSMA PET in patients with HV of PCa compared to patients with pure adenocarcinoma of the prostate in men candidate for radical prostatectomy (RP) and in those with biochemical recurrence (BCR) after surgery.

Methods: We identified 42 high-risk men evaluated with 68Ga-PSMA PET/CT before RP and 44 high-risk men with BCR after RP between 2018 and 2021 at a single center. All specimens were evaluated by a single uro-pathologist. The following HV were considered: cribriform, glomeruloid, intraductal components of adenocarcinoma, fused poorly-formed, neuroendocrine and foamy. Sensitivity, specificity, positive (PPV) and negative predictive value (NPV) according to the presence of HV at pathology were assessed in men candidate for RP. In men with BCR, rates of positive PSMA PET in men with or without HV at pathology were compared, logistic regression predicted the probability of having a positive PSMA PET.

Results: HV was detected in 28 men (67%) candidate to RP and in 13 (29%) with BCR. No difference in ISUP grade or clinical stage were observed in both staging and re-staging cohorts (p>0.1). In the primary staging cohort, a positive PSMA PET was recorded in 9 (32%) and 5 (36%) men with or without HV. At final pathology, 10 (36%) and 7 (50%) men with and without HV had LNI. On a patient-level analysis, PSMA PET showed 71% sensitivity, 100% specificity, 82% NPV and 100% PPV in the detection of LNI in patients with pure aPCa, with an overall accuracy of 87%. In patients with HV, PSMA PET showed 50% sensitivity, 78% specificity, 74% NPV and 56% PPV in the detection of LNI, with lower overall accuracy (67%). In men with BCR, 28 (64%) had positive PSMA PET at the time of BCR. Among these, a positive PSMA PET was recorded in 23 (74%) men with pure aPCa at pathology. Conversely, only 5 patients with BCR and HV (38%) had a positive PSMA PET (p =0.046). At logistic regression, the presence of HV was associated with lower rates of positive PSMA PET (OR 0.21, p=0.028).

Conclusions: PSMA PET had substantial lower diagnostic accuracy if HV of PCa are present both in the primary and recurrent settings. Our findings suggest that the variability of PSMA PET performance may be correlated with the presence of HV of PCa. Therefore, HVs should be assessed to optimize indication for and interpretation of PSMA PET.

Source of Funding: None.