To assess the value of computed tomography (CT) and magnetic resonance imaging (MRI) in the diagnosis of pancreatic neuroendocrine neoplasms (PNEN) and to analyze the factors influencing the preoperative imaging diagnosis of PNEN.

From January 2016 to November 2016, patients with PNEN diagnosed by surgery and biopsy were collected. CT and MRI data of them were analyzed. The CT values or signal intensity of the lesions and the pancreatic parenchyma were measured and the contrast-to-noise ratio (CNR) of the lesion was calculated. Detecting sensitivity and diagnosis accuracy of CT and MRI were compared. Detecting sensitivity of different MRI sequences was also analyzed. Diagnosis accuracy of non-functional PNEN and functional PNEN was compared and analyzed. Lesion CNR was compared between arterial phase and portal venous phase of the contrast enhanced CT. The sensitivity, accuracy and constituent ratio were compared by nonparametric analysis. Independent sample t test and one-way analysis of variance were performed for the quantitative parameters comparison.

A total of 54 patients with 56 lesions of PNEN were included for two of whom had two lesions each. CT and MRI were both performed in 44 patients (46 lesions). Detecting sensitivity and diagnosis accuracy of CT were 97.8% (45/46) and 87.0% (40/46), respectively. Detecting sensitivity of MRI were 97.8% (45/46) and 89.1% (41/46), respectively. There was no significant difference in detecting sensitivity and diagnosis accuracy between CT and MRI (both P>0.05). The CNR of lesion in arterial phase was higher than that of portal venous phase(4.7±3.8 vs 3.4±2.5), and the difference was statistically significant (t=2.949, P<0.05). Detecting rates of T1 weighted imaging with fat suppression (T1WI-FS) image, T2 weighted imaging with fat suppression (T2WI-FS) image, diffusion weighted imaging and dynamic contrast enhanced T1WI-FS image were 90.0% (45/50), 88.0%(44/50), 86.0%(43/50), and 91.7% (44/48), respectively. There was no significant difference in detecting rate among these images sequences (Q=2.526, P=0.510). Tumor diameter in non-functional PNEN was significantly larger than that in functional PNEN ((2.9±1.6) cm vs (1.7±0.7) cm) (t=3.479, P<0.05). The overall diagnosis rate of non-functional PNEN with CT and MRI before operation was 70.8% (17/24), which was significantly lower than that of functional PNEN (100.0%, 31/31) (χ²=10.360, P=0.002).

CT and MRI are both sensitive in detecting PNEN, and they were two complementary modalities. CT image in arterial phase delineated the lesion better than that in portal venous phase. MRI images with different sequences can be complementary and there is no significant difference in detecting sensitivity for PNEN among different sequences. CT and MRI play an equal role in the diagnosis of PNEN before operation. Because of atypical CT and MRI findings, the diagnosis of non-functional PNEN is more difficult than functional PNEN.

Pancreatic neuroendocrine neoplasms; Tomography, X-ray computed; Magnetic resonance imaging; Diagnosis