Underestimation of Gleason score at prostate biopsy reflects sampling error in lower volume tumours

Abstract

Objective: • To determine the influence of tumour and prostate gland volumes on the underestimation of prostate cancer Gleason score in diagnostic core biopsies.

Patients and methods: • Patients undergoing radical prostatectomy with matched diagnostic biopsies were identified from a prospectively recorded database. • Tumour volumes were measured in serial whole-mount sections with image analysis software as part of routine histological assessment. • Differences in various metrics of tumour and prostate volume between upgraded tumours and tumours concordant for the lower or higher grade were analysed.

Results: • In all, 684 consecutive patients with Gleason score 6 or 7 prostate cancer on diagnostic biopsy were identified. • Of 298 patients diagnosed with Gleason 6 tumour on biopsy, 201 (67.4%) were upgraded to Gleason 7 or higher on final pathology. Similarly, of 262 patients diagnosed with Gleason 3 + 4 = 7 prostate cancer on initial biopsy, 60 (22.9%) were upgraded to Gleason score 4 + 3 = 7 or higher. • Tumours upgraded from Gleason 6 to 7 had a significantly lower index tumour volume (1.73 vs 2 mL, P= 0.029), higher calculated prostate volume (41.6 vs 39 mL, P= 0.017) and lower relative percentage of tumour to benign glandular tissue (4.3% vs 5.9%, P= 0.001) than tumours concordant for the higher grade. • Similarly, tumours that were Gleason score 3 + 4 on biopsy and upgraded on final pathology to 4 + 3 were significantly smaller as measured by both total tumour volume (2.3 vs 3.3 mL, P= 0.005) and index tumour volume (2.2 vs 3, P= 0.027) and occupied a smaller percentage of the gland volume (6.3% vs 8.9%, P= 0.017) compared with tumours concordant for the higher grade. • On multivariate analysis, lower prostate weight (hazard ratio 0.97, 95% confidence interval 0.96-0.99, P < 0.001) and larger total tumour volume (hazard ratio 1.87, 95% confidence interval 1.4-2.6, P < 0.001) independently predicted an upgrade in Gleason score from 6 to 7. In tumours upgraded from biopsy Gleason 3 + 4, only higher index tumour volume (hazard ratio 3.1, 95% confidence interval 1.01-9.3, P= 0.048) was a significant predictor of upgrading on multivariate analysis.

Conclusions: • Under-graded tumours are significantly smaller than tumours concordant for the higher grade, indicating that incomplete tumour sampling plays a significant role in Gleason score assignment error. • Surrogate measures of tumour volume may predict those at greatest risk of Gleason score upgrade.