Performance of transrectal prostate biopsies in detecting tumours and implications for focal therapy

Abstract

OBJECTIVE: The aim of this study was to investigate the performance of transrectal biopsies in predicting pathological outcomes after radical prostatectomy (RP) and in estimating possible candidates for focal therapies. MATERIAL AND METHODS: The study was a retrospective analysis of 96 prostate cancer patients treated by robot-assisted laparoscopic RP at Helsinki University Central Hospital in 2009-2010. Data from diagnostic biopsies were compared with data from reassessment of RP slides. At reanalysis, an index tumour was chosen for all patients and was determined as being the most dedifferentiated tumour or the largest tumour with Gleason pattern 3 in case Gleason patterns 4 or 5 were absent. The performance of prostate biopsies in predicting cancer laterality, tumour size and tumour location was analysed. Statistical methods included Spearman's correlation, linear regression analysis and Pearson's chi-squared test. Suitability for focal therapies was assessed based on tumour morphology and laterality. RESULTS: The extent of cancer in biopsies correlated with tumour size in the apex and middle of the prostate [standard coefficients in linear regression for the apex 2.479-2.553, 95% confidence interval (CI) 1.952-3.056, p < 0.001-0.007; and for the middle 1.936-2.388, 95% CI 1.504-2.861, p < 0.001]. Prostate biopsies performed moderately in predicting tumour location in RP slides (positive predictive value 34.1-90.9%). Thirty-six patients (37.5%) would possibly have been candidates for focal therapy and thirty-nine (40.6%) patients for hemiablation. CONCLUSIONS: Contemporary transrectal prostate biopsies are a suboptimal tool for predicting pathological findings at RP. Approximately 40% of patients would possibly have been suitable candidates for focal or hemiablative therapies.