Pre-biopsy 3-Tesla MRI and targeted biopsy of the index prostate cancer: correlation with robot-assisted radical prostatectomy

Objective To study whether pre-biopsy 3-Tesla prostate magnetic resonance imaging (MRI) with targeted biopsy allows accurate anatomical and oncological characterization of the index prostate tumour, and whether this translates into improved positive surgical margin (PSM) rates after radical prostatectomy. Patients and Methods We conducted a retrospective analysis of all men (n = 201) who underwent robot-assisted radical prostatectomy (RARP) between July 2012 and July 2014. Patients were divided into a study group (n = 63) who had undergone pre-biopsy 3-Tesla MRI, followed by visual targeted and systematic prostate biopsy, and a control group (n = 138) who had undergone systematic biopsy alone. The two groups were well matched regarding patient and cancer characteristics. The primary study objective was to assess the accuracy of pre-biopsy MRI for localizing the index tumour. Secondary study objectives were to assess the accuracy of MRI in assessing the maximum tumour diameter (MTD) of the index tumour focus and accuracy of the targeted biopsy in determining the Gleason score and primary Gleason grade of the index tumour focus and whether PSMs were improved after RARP. The reference standard was whole-gland pathology of the resected prostate gland. Continuous variables and proportions were compared using the t-test and Mann-Whitney test or contingency tables, respectively. Pearson's correlation coefficient and Bland-Altman plots were used to compare measurement of MTD. Results The MRI accurately located the index tumour focus in 73% of patients. Accuracies, stratified according to use of the Prostate Imaging Reporting and Data System (PI-RADS) categories 5, 4 and 3, were 94, 75 and 60% respectively. Accuracies stratified according to MTD of <= 0.7, <= 1 and >1 cm were 50, 57 and 79%, respectively. There was a positive linear correlation between MRI and histological MTD (r = 0.42, 95% confidence interval [CI] 0.16-0.63; P = 0.002), but MRI generally underestimated the MTD: the mean MRI-measured MTD was 1.51 cm (95% CI 1.29-1.72) vs a mean pathological MTD of 2.15 cm (95% CI 1.86-2.43). Targeted biopsy identified 37% more cancer per core than non-targeted biopsy. The mean maximum core length was 8.9 mm (95% CI 7.8-10) vs 6.5 mm (95% CI 5.8-7.2) for the study vs the control group (P = 0.0002; non-paired t-test). Gleason scoring was significantly more predictive after targeted biopsies, with unchanged scores in 40/63 men (63%) vs 62/138 men (45%) in the study and control groups, respectively (P = 0.001; Fisher's test). The odds of Gleason upgrading were 2.5 times greater (P = 0.028) in the control group. The primary Gleason grade was not significantly different in the two groups [45/63 men (71%) vs 91/138 men (66%); study vs control group respectively (P = 0.51, Fisher's test)]. Overall PSMs were nonsignificantly lower in the study group (15.8 vs 18.8%; P = 0.84, Fisher's test); and the MRI location of the index tumour focus correlated with the site of PSM in 70% of men in the study group. Conclusions Pre-biopsy MRI can accurately identify the index prostate tumour, especially in those with higher PI-RADS grades and tumour diameter. Targeted biopsy of this focus retrieves significantly more cancerous tissue per core, and is more accurate regarding Gleason scores, but not primary Gleason grade. MRI underestimated the MTD, and PSMs were not significantly improved in the present study.