Pancreatic Cancer CT: Prediction of Resectability according to NCCN Criteria.

Purpose To evaluate the diagnostic performance of CT in the determination of pancreatic cancer resectability according to the National Comprehensive Cancer Network (NCCN) criteria to predict R0 resection. Materials and Methods Structured reports of pancreatic CT clinically prepared by board-certified abdominal radiologists from January 2014 to March 2017 were retrospectively reviewed to assess resectability (resectable, borderline resectable, or unresectable) according to NCCN criteria (version 1.2017) in 616 patients (369 men, 247 women; mean age, 63 years ± 10 [standard deviation]) with pancreatic cancer. Negative resection margin (R0) rates were assessed based on CT resectability status in patients who underwent upfront surgery. R0 resection-associated factors were identified by using logistic regression analysis. Results In 371 patients who underwent surgery, R0 resection rates were 73% (171 of 235), 55% (57 of 104), and 16% (five of 32) for resectable, borderline resectable, and unresectable disease, respectively (P < .001). At multivariable analysis, tumor diameter larger than 4 cm (P < .001) and abutment to the portomesenteric vein (P < .001) were significantly associated with margin-positive resection in patients with resectable disease at CT. R0 resection rates were 80% (123 of 154) for resectable disease without portomesenteric vein abutment, 59% (48 of 81) for resectable disease with portomesenteric vein abutment, 83% (57 of 69) for resectable disease 2 cm or smaller, and 29% (five of 17) for tumors larger than 4 cm. Conclusion CT resectability is used to stratify patients with pancreatic cancer according to the possibility of R0 resection. Larger tumor size and tumor abutment to the portomesenteric vein are associated with margin-positive resection in patients with resectable pancreatic cancer. © RSNA, 2018 Online supplemental material is available for this article. See also the editorial by Fowler in this issue.