Timo Soeterik

70 CHAPTER 4 What our study adds is that we have validated their previously reported findings in a large multi-institutional cohort of newly diagnosed prostate cancer patients, including a far larger number of low-risk patients ( N = 447, 27%). Our data confirms that additional use mpMRI for risk stratification results in stage migration to higher risk groups in one- third of the patients, and also assignment of patients to intensified treatment regimens. Furthermore, our analysis showed that the subsequent stage migration of patients to higher risk groups resulted in a more favorable prognosis of the high-risk subgroup as a whole, in terms of prolonged recurrence-free survival (Figure 1). This observed evolvement of cancer prognosis, due to change of tumour stage as a consequence of use of newly developed diagnostic modalities, is described as the “Will Rogers phenomenon.” 16 Our additional findings regarding the diagnostic accuracy of mpMRI and DRE for the detection of non-organ-confined disease may further guide clinicians in the way they should incorporate mpMRI information for risk group classification. As shown, the positive predictive value of mpMRI for stage ≥T3a increases in case of a higher PSA and ISUP biopsy grade, reaching 82% in patients with either PSA above 20 ng/mL or ISUP > grade 3. Opposite findings were observed regarding the NPV, as the NPV declined inversely proportional to the risk classification from low to locally advanced high-risk, respectively from 83% to 46%. Our study shows that solely relying on mpMRI potentially leads to overtreatment given the relatively high false positive rates for non-organ-confined disease compared with DRE, especially in patients with a PSA below 10 ng/mL and biopsy ISUP grade 1. Presence of non-organ-confined disease on mpMRI is extremely rare in ISUP grade 1 prostate cancer, with a reported prevalence of 7 of 7817 (0.3%) in a series of patients undergoing radical prostatectomy. 17 Thus, if stage ≥T3a is present on mpMRI in patients with ISUP 1 on biopsy, this may either indicate a false positive test result or biopsy sampling error. In addition, our study revealed that lower specificity of mpMRI for the detection of stage ≥T3a is countered by a far higher sensitivity, compared with DRE. Among patients undergoing RARP with a serum PSA <20 ng/mL and ISUP grade ≤ 3, sensitivity and specificity for detection of stage ≥T3a were respectively 8 of 131 (6%) and 213 of 218 (98%) for DRE versus 65 of 131 (50%) and 180 of 218 (83%) for mpMRI. Use of mpMRI thus resulted in correctly classifying an additional 57 of 131 (44%) patients as high-risk, countered by additionally over staging solely 38 of 218 (17%) patients with organ-confined disease as high-risk. These results confirm that mpMRI outperforms DRE with regard to local staging, and reduces the risk of understaging and potential undertreating those who harbor non- palpable stage ≥T3a prostate cancer. However, the lower specificity of mpMRI for the detection of stage ≥T3a tumours, and thus its increased risk overstaging, shows the technique is still not perfect. Combining mpMRI information with other baseline factors including detailed (target) biopsy information and baseline PSA can further improve prediction of unfavorable histopathology. 18 At present, several nomograms have been developed which can be used for this purpose. 19,20