#0673
Algorithmic Prediction of Gleason Score Upgrading in Biopsy 3+3 Prostate Cancer: Insights from a Retrospective Cohort
T. Chiu1, K. Lu1, H. Lee1, Y. Juan1, C. Li1, W. Wu1
1Kaohsiung Medical University Hospital, Department of Urology, Kaohsiung city, Taiwan
Introduction:
Patients with biopsy-confirmed Gleason score (GS) 3+3 prostate cancer (PCa) (Grade Group 1) are typically regarded as having low-risk, indolent disease. However, a significant subset may experience Gleason score upgrading (GSU) after radical prostatectomy (RP), which is associated with a higher risk of adverse outcomes. This study aimed to develop and validate an algorithm-based model using clinical characteristics to predict GSU in patients undergoing RP for GS 3+3 prostate cancer.
Material and methods:
This retrospective cohort analysis included patients who underwent RP for GS 3+3 PCa at Kaohsiung Medical University Hospital between 2010 and 2023. Preoperative clinical data, including Prostate-Specific Antigen (PSA) levels, PSA density, prostate volume, number of positive biopsy cores, and multiparametric MRI (mpMRI) findings (PI-RADS scores), were incorporated into a multivariate predictive model. Model performance was evaluated using area under the receiver operating characteristic curve (AUC), accuracy, sensitivity, and specificity.
Results:
A total of 121 patients (median age 66 years, mean PSA 8.6 ng/mL) met inclusion criteria. Among these, 78 patients (64.5%) exhibited GSU on final pathology. Compared to those without upgrading, patients with GSU had a median age of 67.5 years (p=0.0594) and a higher, though not statistically significant, median pre-biopsy PSA of 9.53 ng/mL (p=0.2219). PSA density was notably elevated in the GSU group (median 0.3 ng/mL/cm³ vs. 0.18 ng/mL/cm³, p=0.0033), while prostate volume was lower (median 36.4 mL vs. 41 mL, p=0.0145). Perineural invasion (PNI) was more prevalent in the GSU group (10.53% vs. 2.33%, p=0.0838), and lymphovascular invasion (LVI) was significantly higher (53.25% vs. 20.93%, p=0.0004). The algorithm demonstrated strong predictive performance, with an AUC of 0.85, 80% accuracy, 82% sensitivity, and 77% specificity. Independent predictors of GSU included PSA density, mpMRI PI-RADS score, and PNI