Timo Soeterik

133 Development of a novel EPE nomogram as demonstrated by the net benefit over a range of risk thresholds within a suitable range for clinical decision-making. Besides the potential of improving quality of care in terms of patient experience, our nomogram may also improve quality of care in terms of clinical outcomes. Using the nomogram with a risk threshold set at 20%, accurate patient selection for nerve sparing RP is possibly leading to the relatively highest clinical benefit. With a 20% risk threshold, 2170/3740 (58%) of the prostatic lobes of the development cohort would fall below the cut-off and nerve sparing would be advised for these. Of these however, 190 cases (9%) would have EPE. Although nerve sparing can be safely performed in the majority of patients with risk of EPE below this threshold, it remains critical to relate these risks to the patient’s preferences and willingness to trade-off between the potential quality of life benefit of nerve sparing and the increased risk of positive surgical margins. In addition to optimising preoperative staging, surgeons could consider other tools, such as intraoperative frozen section technology (NeuroSafe), to further optimise surgical outcomes. 30 Moreover, since NeuroSafe is a time-consuming and costly procedure, our nomogram as a triage test for NeuroSafe could contribute to the cost-effective deployment of NeuroSafe. Although this study has a number of strengths, such as a large number of cases and external validation in two separate patient cohorts, some limitations have to be acknowledged. Firstly, the majority of the study data was derived from daily clinical practice and there was no central histopathologic or radiologic review. However, fact real-world clinical data were used for model development and validation could also be a potential strength since these features reflect the real-world clinical situation, for which this nomogram is designed. Secondly, although accounted for using multiple imputations, the percentage of prostatic lobes with one or more missing covariates (27%) in this study is a limitation. However, results from our additional analysis performed using complete- case data (data not shown) would not alter the study’s main conclusions. Lastly, model development and external validation were both performed by the same study group and solely concerned Dutch patients. We therefore encourage other (international) study groups to also externally validate our nomogram as well. CONCLUSIONS We developed a simple and robust nomogram, including mpMRI information and readily available clinical parameters, for the prediction of side-specific EPE. This nomogram can be safely used to optimise patient selection for nerve sparing radical prostatectomy. 7