Ten studies were included in the review (627 patients, range 23 to 201): four prospective and six retrospective cohorts. Six studies reported consecutive patient recruitment and seven reported blinding of interpreters of MRI scans. Delay between MRI and radical prostatectomy ranged from one to 90 days.
The pooled sensitivity of MRI using a combination of diffusion and T2-weighting was 76% (95% CI 65 to 84), specificity was 82% (95% CI 77 to 87), LR+ was 4.31 (95% CI 3.12 to 5.92), LR- was 0.29 (95% CI 0.20 to 0.43) and the AUC was 0.84 (95% CI 0.80 to 0.87). Compared to T2-weighting alone, the diffusion and T2-weighting combination (seven studies), combination had a higher pooled sensitivity (72% versus 62%) and specificity (81% versus 77%).
Meta-regression showed patient recruitment and blinding of interpreters of MRI scans were the most important variables contributing to the observed heterogeneity. There was no evidence of a threshold effect of publication bias. Further results from subgroup and sensitivity analyses were reported.