Pancreatic neuroendocrine tumour (PNET): Staging accuracy of MDCT and its diagnostic performance for the differentiation of PNET with uncommon CT findings from pancreatic adenocarcinoma.

Abstract

To investigate staging accuracy of multidetector CT (MDCT) for pancreatic neuroendocrine tumour (PNET) and diagnostic performance for differentiation of PNET from pancreatic adenocarcinoma. We included 109 patients with surgically proven PNET (NETG1 = 66, NETG2 = 31, NEC = 12) who underwent MDCT. Two reviewers assessed stage and presence of predefined CT findings. We analysed the relationship between CT findings and tumour grade. Using PNETs with uncommon findings, we also estimated the possibility of PNET or adenocarcinoma. Accuracy for T stage was 85-88% and N-metastasis was 83-89%. Common findings included well circumscribed, homogeneously enhanced, hypervascular mass, common in lower grade tumours (p < 0.05). Uncommon findings included ill-defined, heterogeneously enhanced, hypovascular mass and duct dilation, common in higher grade tumours (p < 0.05). Using 31 PNETs with uncommon findings, diagnostic performance for differentiation from adenocarcinoma was 0.760-0.806. Duct dilatation was an independent predictor for adenocarcinoma (Exp(B) = 4.569). PNETs with uncommon findings were associated with significantly worse survival versus PNET with common findings (62.7 vs. 95.7 months, p < 0.001). MDCT is useful for preoperative evaluation of PNET; it not only accurately depicts the tumour stage but also prediction of tumour grade, because uncommon findings were more common in higher grade tumours. • CT accurately depicts the T stage and node metastasis of PNET. • Uncommon findings were more common in higher grade tumours. • CT information may be beneficial for optimal therapeutic planning.