Exploration of a novel quantitative model based on PRIMARY score in the diagnosis of clinically significant prostate...

To evaluate the feasibility and efficacy of the Quantitative-PRIMARY (Q-PRIMARY) model based on the PRIMARY score for diagnosing clinically significant prostate cancer (csPCa) using [18F]PSMA-1007 PET/CT.

A total of 324 consecutive patients were prospectively enrolled and retrospectively analyzed. Cohort assignment was based on the availability of in-house multiparametric MRI: a training cohort (n = 222, no in-house MRI) and an internal validation cohort (n = 102, with in-house MRI). Histopathology from systematic 12-core biopsy or radical-prostatectomy specimens served as the reference standard. We developed the Q-PRIMARY model by integrating visual PRIMARY scores with quantitative SUV-based thresholds (ΔSUVpeak, SUVmax, SUVR) for each PRIMARY category (scores 2-4).

There were 27, 41, 33, and 46 patients in the four groups, with positive lesions found in 62%, 62%, 53.53%, and 65.08%, respectively. Optimal diagnostic parameters for csPCa included ΔSUVpeak, SUVmax, SUVratio (SUVR), and SUVmax. The areas under the curve (AUC) ranged from 0.888 to 0.944, with high sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV). The Q-PRIMARY demonstrates superior diagnostic accuracy and clinical benefits compared to MRI PI-RADS, with more stable performance in AUC (0.978 vs. 0.768), sensitivity (97.18% vs. 92.96%), specificity (96.77% vs. 54.84%), PPV (98.60% vs. 82.50%), NPV (93.70% vs. 77.30%), and clinical decision curve analysis (DCA).

In this selected high-risk cohort, the [18F]PSMA-1007 PET/CT Q-PRIMARY framework delivered high diagnostic performance and complemented multiparametric MRI. External prospective validation is warranted before clinical adoption.