Value of diffusion-weighted magnetic resonance imaging in predicting World Health Organization grade in G1/G2 pancreatic neuroendocrine tumors

The present study aimed to investigate the value of diffusion-weighted magnetic resonance imaging (DWI) in the grading of well-differentiated pancreatic neuroendocrine tumors (PanNETs). A total of 44 patients with histologically proven well-differentiated PanNET [grade 1 (G1) and grade 2 (G2) according to the World Health Organization (WHO) criteria] underwent pretreatment magnetic resonance imaging (MRI), which was retrospectively analyzed. The location, size, cystic or solid appearance, boundary, presence or absence of tumor contrast enhancement, and MRI signal of the tumor were assessed. Apparent diffusion coefficients (ADCs) within the primary tumor were measured on ADC maps. Receiver operating characteristic curves were used to determine ADC cut-off values and the sensitivity and specificity of the grade prediction. Spearman's rank correlation was utilized to probe the association between ADC value and PanNET grade or pathological parameters. G1 tumors exhibited a well-circumscribed border more often than G2 tumors. No marked differences were observed between PanNET G1 and PanNET G2 for cystic or solid appearance, enhancement, and T1- and T2-weighted imaging signals. Marked hyperintensities were more common in PanNET G2 tumors than in PanNET G1 tumors (P<0.01). The mean ADC values were statistically different between the normal pancreas and G1 and G2 tumors (P<0.01), and between PanNET G1 and PanNET G2 (P<0.05). Correlation analysis showed that ADC value was negatively correlated with PanNET grade, mitotic count and Ki-67 proliferation index (P<0.05). The cut-off ADC value was 0.930×10(−3) mm(2)/sec, which identified G2 tumors with 82.4% sensitivity and 79.5% selectivity. ADC value therefore represents a non-invasive and valuable imaging parameter in predicting the WHO grade of tumors in well-differentiated PanNET.