Timo Soeterik

21 General Introduction and Thesis Outline Part II Impact of MRI on Prostate Cancer Risk Classification Performing MRI to assist prostate cancer detection and staging is becoming standard of care. However, to what extent MRI local staging information can be incorporated into existing risk classification systems remains unclear. Due to the potentially improved local staging of the tumour by MRI, compared with DRE, cT of the tumour could be established more accurately. It remains unclear how this impacts the currently used D’Amico risk classification system. To evaluate whether MRI can replace DRE for determining cT, the accuracy of MRI as a local staging tool, as well as its impact on prostate cancer risk classification in terms of stage migration, needs further exploration. We assumed that use of MRI would lead to upstaging when used for determining cT, due to higher detections rates of EPE and seminal vesicle invasion, compared with DRE. To determine the correctness of the established cT by MRI and DRE, we will compare cT assessed by both modalities with definite surgical pathology. Comparison of the cT assessed by both strategies will provide insight into which entity should be used to determine the cT and prostate cancer risk classification (Chapter 4) . Part III Association Between Nerve Sparing and Positive Surgical Margins Preservation of the neurovascular bundles during radical prostatectomy can be performed to optimise the probability of retaining postoperative erectile function and urinary continence. It is generally assumed that nerve sparing increases the risk of positive surgical margins. However, the majority of prior studies have resulted in contra-intuitive results, showing that nerve sparing is not associated with an increased risk of positive margins compared to non-nerve sparing surgery. These findings could be consequential to methodological limitations of prior studies, not exempt of selection bias. We assumed that by performing a study including an analysis with a side-specific approach, adjusting for a large number of covariates (including MRI staging information), would lead to more reliable results. Our findings regarding the association between side-specific nerve sparing and ipsilateral positive surgical margins are presented in Chapter 5 . Part IV Incorporation of MRI into Clinical Prediction Models Accurate determination of the local extent of the prostate cancer tumour is regarded to be crucial for determining the eligibility for preservation of the neurovascular bundles during radical prostatectomy. However, available tools that can be used for side- specific prediction of EPE are limited. The majority of these tools also do not include MRI local staging information. As MRI may improve local staging, we hypothesized that a nomogram including side-specific MRI staging information can lead to accurate 1