ISUP Grade Group Migration Between Prostate Biopsy and Radical Prostatectomy: A Systematic Review of Emerging...

Background/Objectives: International Society of Urological Pathology (ISUP) Grade Group discordance between prostate biopsy and radical prostatectomy (RP) affects 25-50% of men with localized prostate cancer and has direct implications for active surveillance, nerve-sparing, and adjuvant-treatment decisions. We synthesized contemporary evidence on the frequency of biopsy-to-RP grade migration and on emerging imaging, molecular/genomic, and artificial-intelligence (AI)-based predictors of migration. Methods: MEDLINE (PubMed) was searched from 1 January 2010 to 5 May 2026, supplemented by citation searching of recent systematic reviews. Eligible studies reported paired biopsy-RP pathology under the modified 2005 Gleason or the 2014/2019 ISUP Grade Group system, with ≥50 paired cases. Risk of bias was assessed with QUIPS, PROBAST, and QUADAS-2; certainty of evidence was rated with a GRADE framework adapted for prognostic research. Synthesis followed the SWiM reporting guidance. Results: Thirty-seven primary studies and eight prior systematic reviews were included. Concordance ranged from 44% to ~70%, upgrading from 14% to 67%, and downgrading from 5% to 26%. Biopsy GG1 disease showed the highest absolute upgrading risk (55-67%). Four predictors reached moderate GRADE certainty: PI-RADS category, biopsy approach (combined vs. systematic), PSMA-PET maximum standardized uptake value, and PSA density (PSAD). Cribriform/intraductal carcinoma at biopsy, the Decipher genomic classifier, the Prostate Health Index, and machine-learning and radiomics models reached very low certainty; clinical nomograms reached low certainty; germline alterations as direct predictors reached very low certainty. PROBAST flagged the analysis domain as high risk in five of eight prediction-model studies, and only one of eight models was externally validated. Conclusions: Despite a decade of refinements in grading, imaging, biopsy technique, and molecular profiling, biopsy-to-RP grade discordance remains substantial. Contemporary evidence supports incorporating PI-RADS, biopsy approach, and PSAD into shared decision-making; emerging molecular and AI-based predictors require prospective external validation in adequately powered, geographically representative cohorts before routine clinical adoption.