Reproducibility in Contouring the Neurovascular Bundle for Prostate Cancer Radiotherapy Treatment Planning

Abstract

Erectile dysfunction is a significant long-term toxicity of prostatic radiation. Our institution has previously reported on the dosimetric feasibility of MRI-guided neurovascular bundle (NVB) sparing during definitive prostate radiotherapy (RT), a structure that is spared during prostatectomy but not traditionally spared during RT. The purpose of this study is to determine the reproducibility and reliability of contouring the NVB on MRI. One hundred twenty total data points on ten consecutive 3-Tesla pelvic MRI’s for patients with prostate cancer without extra prostatic extension were reviewed. One pelvic radiologist, with 15 years’ experience, served as the expert in contouring the right and left NVB on T2-320 sequences for twenty total structures. Five radiation oncologists, with varying levels of experience, delineated the right and left NVB on the same scans for one hundred total structures. The intraclass correlation coefficient (ICC), Pearson’s correlation coefficient (PCC), and the Dice similarity coefficient (DSC) were calculated to evaluate the reproducibility of the MRI-based NVB contours. The overall ICC was 0.89 (95% CI: 0.81-0.95, P<0.01). The PCC for each rater versus the expert were estimated (see Table 1). The average DSC was reported for each rater: rater 1 was 0.72 (standard deviation [SD]: 0.07), rater 2 was 0.73 (SD: 0.06), rater 3 was 0.73 (SD: 0.09), rater 4 was 0.74 (SD: 0.09), and rater 5 was 0.68 (SD: 0.13). Overall across all raters, the average DSC was 0.72 (SD: 0.09).Abstract 237; Table 1The Pearson correlation coefficient for each rater compared with the expertRaterCorrelation Coefficient (95% CI)10.95 (0.86-0.98)20.98 (0.95-0.99)30.94 (0.86-0.98)40.98 (0.95-0.99)50.84 (0.63-0.93) Open table in a new tab Our series is the first to quantify the reliability and reproducibility of NVB contouring on MRI, with overall excellent agreement among radiation oncologists and the expert radiologist. Additionally, the observed high DSC indicates good overlap of the NVB volumes. Given these results, as well as our institution’s previous publication showing the dosimetric feasibility of sparing the NVB during RT, we have begun investigation into a prospective trial evaluation NVB-sparing RT for prostate cancer.