Can post-hoc video review of robotic pancreaticoduodenectomy predict portal/superior mesenteric vein margin status in pancreatic adenocarcinoma?

Background: Achieving margin negative resection is a significant determinant of outcome in pancreatic adenocarcinoma (PDA). However, because of the fibrotic nature of PDA, it can be difficult to discriminate fibrosis from active disease intra-operatively. We sought to determine if post-hoc video review of robotic pancreatico-duodenectomy (RPD) could predict the portal/superior mesenteric vein (PV/SMV) margin status on final pathology. Methods: Experienced pancreatic surgeons, blinded to patient and operative variables, reviewed the PV/SMV margin for available RPD videos of consecutive PDA patients from 9/2012 through 6/2017. Results: 107 RPD videos were reviewed. Of 76 patients (71%) predicted to have a negative vein margin on video review, 20 patients (26%) had a pathologic positive margin. 25 of 31 patients (81%) predicted to have positive margin on video review were positive on pathology. The specificity of video prediction was 90.3% with a sensitivity of 55.6% and an accuracy of 75.7%. Conclusion: Post-hoc video review prediction is unable to reliably predict a positive (R1) margin at the portal vein/SMV, suggesting that intra-operative clinical assessment may be suboptimal in determining the need for more extensive resections.