Timo Soeterik

69 Impact of MRI on Risk Classification FIGURE 1. Impact of DRE and mpMRI based risk classification on biochemical-free survival patterns in 293 patients undergoing RARP. (A) DRE-based classification, Log-rank test: p < 0.001. (B) mpMRI-based classification, Log-rank test: p < 0.001. A. B. DISCUSSION Our study demonstrates that use of mpMRI for local staging has a major impact on prostate cancer risk group classification, as stage migration to a higher disease stage occurred in one-third of the evaluated patients. In addition, mpMRI information impacts treatment assignment, as upstaging due to mpMRI findings was associated with a 3 times higher odds for treatment intensification in patients treated with EBRT. A comparative analysis of DRE and mpMRI for the detection of tumour extension outside the prostate showed that with mpMRI more cases of non-organ-confined disease were detected (sensitivity: 51% vs 12%, p < 0.001), countered by higher rates of false positive test results (specificity: 82% vs 97%, p < 0.001). Since mpMRI outperformed DRE in terms of cumulative rates of true negative and positive test results for non-organ-confined disease (67% vs 58%, p < 0.001), it should be preferred over DRE for local staging. The present findings results are consistent with those reported by Draulans et al, who evaluated the impact of mpMRI on the EAU risk classification in 180 patients undergoing RARP. 11 The authors reported comparable upstaging (31%) and downstaging (4%) of risk grouping, as well as the need for treatment intensification due to mpMRI information (27%). 11 Marcus et al also previously evaluated the impact of mpMRI on risk stratification, adhering to the NCCN risk classification. 12 In their study including 71 patients, they observed upstaging in 17% of the patients, as well as significant differences between pre-MRI and post-MRI risk group classifications. Impact on change of management as a direct result of mpMRI was observed in 13 of 71 (18%) patients. Main limitations of both studies were their single-centre nature, including a relatively small sample of patients undergoing RARP and a relatively low number of patients with low-risk disease (4% and 16%, respectively). 4