Comparative analysis of ultrasonographic characteristics between pancreatic neuroendocrine tumor and pancreatic ductal denocarcinoma

Abstract

Objective To compare the characteristics of ultrasonic imaging between pancreatic neuroendocrine tumor (pNET) and pancreatic ductal adenocarcinoma (pDAC), and to identify the ultrasonic characteristics in different pathological grades of pNETs. Methods The ultrasonic imaging data of 67 patients with pathologically confirmed pNETs and 82 patients with pathologically confirmed pDACs from the First Affiliated Hospital of Sun Yat-sen University from January 2010 to March 2019 were retrospectively analyzed. Differences in conventional ultrasonic characteristics and CEUS manifestations between the two groups were compared. Fifty pNET lesions were confirmed with pathological grades G1, G2 and G3.Ultrasonic characteristics of the 3 pathological grades were also compared. Results ①Comparison of ultrasonic imaging characteristics between pNET and pDAC showed that: there were statistically significant differences between the two groups in lesion distribution, lesion size, echo, boundary, signal of blood flow, calcification, dilatation of main pancreatic duct, liver metastasis, vascular involvement, CEUS enhancement in the three phases and non-enhancement area of necrosis (all P<0.05). The binary logistic regression model was built including the signal of blood flow, dilation of main pancreatic duct, liver metastasis and enhancementin arterial and venous phases. The diagnostic model for pNET had 0.988 in sensitivity, 0.791 in specificity, and the area under the ROC curve at 0.951, 95%CI being (0.920, 0.983). ②Comparison of the characteristics of ultrasonic imaging between the pNETs derived from the three pathological grades: there were statistically significant differences among the three groups with the liver metastasis and the enhancement in venous phase (both P<0.05). Only 10.5% (2/19) of G1 pNETs had liver metastasis, while 47.8% (11/23) of G2 and 62.5% (5/8) of G3 pNETs had liver metastasis. In venous phase, 78.9% (15/19) of G1 pNETs showed hyper- or iso- enhancement, while 100% (8/8) of G3 lesions showed hypo-enhancement. Conclusions The combination of features from multiple ultrasonic imaging may help to differentiate pNET from pDAC. There are certain differences in ultrasonic imaging features in pNETs at different pathological grades. Key words: Ultrasonography; Contrast-enhanced ultrasound; Pancreatic neuroendocrine tumor; Pancreatic ductal denocarcinoma