Endoscopic Ultrasound and Computed Tomography Predictors of Pancreatic Cancer Resectability

Introduction This study investigates the ability of endoscopic ultrasound (EUS) and computed tomography (CT) to predict a margin negative (R0) resection and the need for venous resection in patients undergoing pancreaticoduodenectomy (PD). Methods Patients with pancreatic head adenocarcinoma undergoing surgery with intent to resect during the last 5 years were identified. EUS and CT data on vascular involvement were collected. Preoperative imaging was compared to intraoperative findings and final pathology. Contingency table analysis using Fisher’s exact test identified imaging features of EUS and CT associated with unresectability and positive margins. Results Seventy-six patients met study criteria. Forty-seven (62%) underwent potentially curative PD. The R0 resection rate was 70%. There were 16 unresectable patients because of locally advanced disease. Venous involvement >180° and arterial involvement >90° by CT had 100% positive predictive value for failure to achieve R0 resection (p < .01). If patients with prestudy biliary stents were excluded, EUS venous abutment or invasion also predicted R0 failure (p = .02). Combined but not individual EUS and CT findings were predictive of need for vein resection. Conclusions Pancreas protocol CT imaging appears to be a better predictor of resectability compared to EUS. EUS accuracy is affected by the presence of biliary stents. This article was presented at SSAT, Washington DC, May 2007.