The International Society of Urological Pathology (ISUP) grade group (GG) is a classification system derived from the Gleason score. This system stratifies prostate cancer risk based on pathological patterns on histology and plays a key role in treatment planning. However, the often multifocal nature of prostate cancer contributes to discrepancies between biopsy and final radical prostatectomy (RP) histology and limits accuracy of outcome prediction. While both index (highest-grade) and composite (combined ISUP grade of all positive cores) grading approaches are used in clinical practice, it remains unclear which approach better predicts final pathology and patient outcomes. This study aims to evaluate concordance, upgrading, and downgrading rates of index versus composite GG at biopsy (iGG, cGG) in predicting RP histology and biochemical recurrence (BCR) in clinically localised prostate cancer.

We performed a retrospective analysis of a prospective cohort of patients who underwent RP for clinically localised prostate cancer (cT1–T3, N0, M0) between 2009 and 2021 at a single institution. Biopsy specimens were assigned both index and composite GG. Concordance with index RP GG, rates of upgrading/downgrading, and BCR were compared between iGG and cGG. Subgroup analyses evaluated predictive value of MRI parameters and biopsy-derived parameters. BCR was defined as PSA >0.2 ng/mL confirmed on repeat testing.

2084 patients met inclusion criteria. iGG and cGG were concordant in 85% (n=1780) and discordant in 15% (n=304). In biopsy-concordant cases, biopsy GG predicted RP histology with 69.5% concordance; 22.1% were upgraded and 8.4% downgraded. In biopsy-discordant cases, iGG concordance with RP histology was 39.1%, with 10.9% upgraded and 50.5% downgraded. Using cGG in biopsy-discordant cases yielded 41.8% concordance with RP histology, with significantly higher rates of upgrading (53.3%) and fewer downgrades (4.9%). In patients with cGG 2, discordance with iGG was associated with increased BCR risk (HR 2.06, 95% CI 1.09–3.87; p=0.022). We evaluated patients with pre-operative MRIs: PIRADS ≥4, maximum cancer core length ≥6.7 mm, and ≥4 cores sampled from the index lesion region independently predicted iGG concordance with final histology (p<0.01 for all). iGG in patients meeting all three criteria had significantly higher concordance (79.2% vs 41.3%, p<0.01).

Biopsy GG only achieves moderate concordance with RP histology even in biopsy-concordant cases. Concordance with RP lowers in biopsy-discordant cases. cGG may underestimate final pathology and BCR risk. iGG may overestimate disease severity. iGG better reflects pathology in patients with PIRADS ≥4 lesions, longer cancer core length, and sufficient lesion-targeted sampling. Incorporating radiological and histological features can improve biopsy interpretation and guide optimal management in localised prostate cancer.