Preoperative CT-based detection of extrapancreatic perineural invasion in pancreatic cancer

Abstract

Accuracy for computed tomography (CT) diagnosis of extrapancreatic perineural invasion (EPNI) in pancreatic ductal adenocarcinoma (PDAC), which is a significant cause of recurrence, has not been established. The aim of the study was to evaluate the diagnostic accuracy of CT in detecting EPNI preoperatively in resectable PDAC of the pancreatic head. Retrospective study design was approved by institutional review board. Preoperative CT-series of 46 patients with resectable PDAC were evaluated by two independent observers. Plexus Pancreaticus Capitalis-II (PPC-II) was assessed as this area is more susceptible for EPNI. All patients underwent surgery with dedicated histopathology, which served as the reference standard. Histologically EPNI was confirmed in 63.1%. Sensitivity of MDCT was 93.1% (95% confidence interval (CI) 77.23% to 99.15%), specificity 64.7% (95% CI 38.33% to 85.79%) with area under the curve (AUC) 0.789 for the first observer. Positive predictive value (PPV) was 81.82% (95% CI 70.12% to 89.62%), negative predictive value (NPV—84.62% (95% CI 57.98% to 95.64%) with diagnostic accuracy of 82.61% (95% CI 68.58% to 92.18%). Interobserver agreement showed k-value of 0.893 ({p} < 0.001), which represents very good agreement between observers. Median actual survival in patients without EPNI was 30 months (95% CI 18.284–41.716), in patients with EPNI—13 months (95% CI 12.115–13.885). CT provides sufficient diagnostic information to detect PPC-II invasion in patients with resectable PDAC of the pancreatic head. Preoperative detection of EPNI might be an additional argument to perform neoadjuvant chemotherapy in patients with resectable PDAC. It should be included in preoperative evaluation form of CT-findings.