Gleason Pattern 4 Burden and Adverse Pathology in ISUP Grade Group 2 Prostate Cancer: Implications for Active...

Active surveillance (AS) is increasingly considered for selected patients with International Society of Urological Pathology (ISUP) Grade Group (GG) 2 prostate cancer. However, candidate selection remains controversial due to histological heterogeneity and the risk of harboring adverse pathology (AP). We aimed to identify preoperative clinical, multiparametric magnetic resonance imaging (mpMRI), and histological predictors of AP, biochemical recurrence (BCR), and postoperative PSMA PET/CT positivity in men with biopsy-confirmed ISUP GG 2 disease undergoing robotic-assisted radical prostatectomy (RARP).

We retrospectively analyzed 127 consecutive patients with biopsy-proven ISUP GG 2 prostate adenocarcinoma from a prospective single-center RARP database (2012-2024). All patients underwent pre-biopsy 3T mpMRI and systematic plus software-fusion targeted biopsies, with diagnostic specimens strictly lacking cribriform architecture or intraductal carcinoma. AP was defined as pathological stage ≥ pT3 and/or ISUP GG upgrading to ≥3 in the final surgical specimen. BCR was defined as two consecutive postoperative PSA measurements ≥ 0.2 ng/mL. Patients with BCR underwent PSMA PET/CT restaging. Univariable and multivariable logistic regression analyses evaluated independent preoperative predictors of AP, incorporating separate models for ≥10% and ≥15 biopsy Gleason pattern 4 thresholds to prevent collinearity. Model performance was evaluated using the area under the receiver operating characteristic curve (AUC) and Hosmer-Lemeshow calibration tests.

Median patient age was 63.2 years, median PSA was 7.6 ng/mL, and median PSA density (PSAD) was 0.18 ng/mL. PI-RADS ≥ 4 lesions were identified in 71.6% of patients, and median biopsy pattern 4 proportion was 21.1%. Overall, AP occurred in 39 of 127 patients (30.7%), driven by ISUP upgrading to GG ≥ 3 in 21.3% (n = 27) and upstaging to ≥pT3 in 17.3% (n = 22). No undergrading was observed. On univariable analysis, PSA (p = 0.028), PSAD (p = 0.011), and biopsy pattern 4 proportion ≥ 10% (p = 0.042) and ≥15% (p = 0.027) were significantly associated with AP. On multivariable analysis, PSAD (OR 6.75, 95% CI 3.51-12.87, p = 0.005) and biopsy pattern 4 proportion ≥ 10% (OR 2.54, 95% CI 1.15-6.54, p = 0.036) or ≥15% (OR 2.74, 95% CI 1.06-6.06, p = 0.023) remained independent predictors of AP. The final multivariable model demonstrated solid discrimination (AUC = 0.78, 95% CI 0.71-0.85) and satisfactory calibration (p = 0.42). At a mean follow-up of 32.3 months, BCR occurred in 23 patients (18.1%; 3-year recurrence-free survival: 81.3%). Restaging PSMA PET/CT was positive in 8 of 23 recurrences (34.8%; three local, four regional nodal, one solitary bone). Neither clinical nor radiological parameters independently predicted BCR or PSMA PET/CT positivity.

Biopsy Gleason pattern 4 extent (<10% and <15%) and PSAD are the strongest independent preoperative predictors of organ-confined, non-upgraded pathology in ISUP GG 2 prostate cancer. Combining a low pattern 4 burden (<10%, or up to 15% in highly selected cases) with a low PSAD provides objective parameters to refine "favorable intermediate-risk" stratification by identifying patients with a lower probability of adverse pathological features.