Gleason score has been identified as an important variable to predict disease extent and biologic behaviour of prostate cancer. However, the correlation between Gleason score of needle biopsy and surgical specimen is often poor. We studied 72 patients who underwent needle biopsy and radical prostatectomy to correlate Gleason score with PSA, clinical and pathological tumour stage. Only 47.2% of Gleason scores were identical in the biopsy and specimens, 37.5% were undergraded and 15.2% were overgraded. Correlations between clinical and pathological stage were identical in 30.5% of patients, 61.1% of patients were understaged and 8.3% overstaged. In conclusion, accuracy of clinical staging and grading of prostate cancer is low. Although the Gleason score on needle biopsy might be useful to predict the final stage and grade, correlation with surgical specimen is poor.