OBJECTIVE: Biopsy grading at tertiary care centers may or may not be superior to biopsies performed at referral institutions. METHODS: Referral biopsy and tertiary care center biopsy Gleason sums were studied in 758 men treated with radical prostatectomy (RP) at a tertiary care center between 1992 and 2004. Grade agreement was calculated using the Cohen kappa (ê). Logistic regression models predicting high-grade prostate cancer at RP were fitted using either referral or tertiary care center biopsies. Comparison of bootstrap-corrected predictive accuracy estimates were performed using the Mantel-Haenszel test. RESULTS: Grade agreement between biopsy and RP Gleason sum was higher (P = 0.003) for tertiary care center biopsies v referral biopsies (55.5% v 47.9%; P = 0.003). Upgrading occurred in 39.8% of referral biopsies v 32.6% of tertiary care center biopsies (P = 0.03). Tertiary care center biopsies were more accurate in determining RP Gleason sum than referral biopsies (71.5% v 65.6%, P = 0.04). CONCLUSION: More accurate prediction of RP Gleason grade may be achieved if biopsy is performed and graded at tertiary care centers.
