Diagnostic Performance of Computed Tomography in Predicting a Positive Circumferential Resection Margin for Resectable and Borderline Resectable Pancreatic Head Cancer

Abstract

Introduction: To measure the diagnostic performance of computed tomography (CT) for the circumferential resection margin (CRM) status in patients with resectable and borderline resectable pancreatic head cancer. Method: We included 75 patients (37 females, mean age, 71 ± 11 years) who had a resectable and borderline resectable pancreatic head cancer and underwent upfront surgery. For the analysis, the CRM was subdivided into superior mesenteric artery (SMA), superior mesenteric vein/portal vein (SMV/PV) and posterior margins. Three abdominal radiologists retrospectively reviewed the preoperative CT images and evaluated CT resection margin status. The reference standard for the CRM status was pathologically made. A positive resection margin was defined as the presence of tumor cells within 1 mm of the CRM. The diagnostic performance of CT was measured for the overall CRM (per-patient analysis) and each of resection margin (per-margin analysis). We pooled the data across the radiologists using the generalized linear mixed model. Result: Thirty-eight patients were pathologically confirmed to have involved CRM. In per-patient analysis, the pooled sensitivity and specificity was 13.7% and 98.1%, respectively. In per-margin analysis, the pooled sensitivity and specificity for SMA margin was 12.4% and 98.1%, respectively; for posterior margin, 10.3% and 99.4%, respectively; for SMV/PV, 26.9% and 96.3%, respectively. Conclusion: Very low sensitivity and high specificity of CT in predicting a positive CRM suggests that CT has poor performance to select patients who would benefit from neoadjuvant therapy rather than upfront surgery in patients with resectable and borderline resectable pancreatic head cancer.