Digital Subtraction of Magnetic Resonance Images Improves Detection and Characterization of Pancreatic Neuroendocrine Neoplasms.

OBJECTIVE: The aim of this study was to evaluate the usefulness of digital image subtraction of contrast-enhanced magnetic resonance (MR) images for detection and characterization of pancreatic neuroendocrine neoplasms (PanNENs). METHODS: Magnetic resonance examinations of 50 histologically verified PanNENs were retrospectively evaluated by 2 radiologists; 50 ductal adenocarcinomas were included as a control group. Late arterial phase images and correspondent subtracted images were analyzed. Tumor detectability on a subjective 3-point scale and contrast-to-noise ratios were compared across sequences using paired Student t tests. Tumor signal intensity was compared between sequences using χ or Fisher exact tests. RESULTS: Subjective conspicuity and contrast-to-noise ratios of PanNENs were significantly higher on subtracted images compared with correspondent late arterial phase images (P < 0.001 and P = 0.002). The rate of clearly hyperenhancing PanNENs was higher on subtracted images compared with arterial phase images (76% vs 36%). CONCLUSIONS: Digital image subtraction improves tumor conspicuity and allows better characterization of PanNENs compared with late arterial phase images.