Interrater agreement in contouring the neurovascular bundle and internal pudendal artery for MRgRT

Abstract

Purpose or Objective It is hypothesized that radiation damage to neural and vascular tissue, such as the neurovascular bundles (NVBs) and internal pudendal arteries (IPAs), during radiotherapy for prostate cancer (PCa) contributes to erectile dysfunction. Neurovascular sparing MR guided adaptive radiotherapy (MRgRT) aims to preserve erectile function after treatment. However, the NVBs and IPAs are not routinely contoured in current radiotherapy practice. Before neurovascular sparing MRgRT for PCa in the setting of an MR-Linac can be implemented, the agreement of the contouring of the NVBs and IPAs on pre-treatment MRI needs to be assessed. Materials and Methods The guidelines for reporting reliability and agreement studies (GRRAS) recommendations were followed. Four radiation oncologists independently contoured the prostate, the left and right NVB and the left and right IPA in an unselected consecutive series of 15 PCa patients treated with 5x7.25 Gy MRgRT on an MR-Linac. For each patient contouring was done on a single pre-treatment T2-weighted 1.5T MRI. Dice similarity coefficients (DSCs) for pairwise interrater agreement of contours were calculated. DSC = 0 indicating no spatial overlap and DSC = 1 indicating complete spatial overlap between contours. A subset of the caudal half of the contours was made, covering the midgland to apex part of the prostate. For this anatomic region the NVB lies in closest proximity to the prostate and conflict between dose coverage of the prostate and dose sparing of the NVB is highest. Results Median overall interrater DSC for the prostate was 0.91 (interquartile range (IQR): 0.88 – 0.92). For the left and right NVBs the median overall interrater DSC was 0.60 (IQR: 0.54 – 0.67) and 0.62 (IQR: 0.54 – 0.69) respectively (figure 1) and for the left and right IPAs 0.58 (IQR: 0.52 – 0.63) and 0.58 (IQR: 0.51 – 0.63) respectively (figure 2). Subset analysis of the caudal half of the NVBs resulted in a median overall interrater DSC of 0.67 (IQR: 0.59 – 0.73) for the left and 0.66 (IQR: 0.61 – 0.70) for the right side. Conclusion The interrater agreement of the contours of the NVBs and IPAs on MRI was clinically acceptable considering their small spatial volume. The agreement was highest in the subset of the caudal half of the NVB, where agreement is most relevant for neurovascular sparing MRgRT for PCa.