Ectopic pancreas in upper gastrointestinal tract: MRI findings with emphasis on differentiation from submucosal tumor

Background: Ectopic pancreas can frequently be mistaken for other submucosal masses. Ectopic pancreas may follow the signal intensity of mother pancreas on various magnetic resonance (MR) sequences, which might be helpful for differentiation between ectopic pancreas and other submucosal tumors in upper gastrointestinal (UGI) tract. Purpose: To evaluate the value of MR imaging (MRI) in differentiating ectopic pancreases from submucosal tumors in UGI tract. Material and Methods: Fifteen patients with ectopic pancreas and 26 patients with UGI submucosal tumors were included. All patients underwent abdominal MRI with diffusion-weighted imaging (DWI) and gadoxetic acid-enhanced MRI. Qualitative (location, contour, growth pattern, lesion border, and presence of intralesional cystic portion and duct-like structure) and quantitative (long diameter [LD], short diameter [SD], LD/SD ratio, signal intensities and apparent diffusion coefficients [ADC], and signal intensity - and ADC ratios of UGI submucosal lesions to pancreas) parameters were compared between ectopic pancreases and UGI submucosal tumors using Fisher's exact test, the Mann-Whitney U test, and receiver-operating characteristic (ROC) analysis. Results: Duodenum was the most common location for ectopic pancreas (12/15, 79.9%), and the gastric body for UGI submucosal tumors (15/26, 57.7%) (P = 0.005). Round shape was an imaging feature more common in UGI submucosal tumors (12/26, 46.2%) than in ectopic pancreas (1/15, 6.7%; P = 0.021). On all of the various pulse sequences of MR images, ectopic pancreas showed isointensity comparable to that of pancreas more frequently than did sequences of UGI submucosal tumors (P < 0.01). The means (0.95 +/- 0.09, 0.99 +/- 0.06, 1.07 +/- 0.08) of the signal intensity ratio of ectopic pancreas to pancreas on fat-suppressed unenhanced T1-weighted, arterial, and portal phase images were significantly higher than those (0.60 +/- 0.09, 0.62 +/- 0.28, 0.86 +/- 0.27) of UGI submucosal tumors (P < 0.05). In contrast, the means (1.05 +/- 0.12, 0.93 +/- 0.18) of the signal intensity ratios of ectopic pancreas on T2-weighted images and DW images (b = 800 s/mm(2)) were significantly lower than those (1.82 +/- 0.39, 2.35 +/- 0.94) of UGI submucosal tumors (P < 0.001). On ROC analysis, if absolute value of difference between 1.0 and signal intensity ratio of submucosal lesions to the pancreas on T1-weighted images is <0.21, the sensitivity and specificity reach 100.0% for diagnosis ectopic pancreas from submucosal tumors. Conclusion: Abdominal MRI with DWI can be a valuable tool for differentiating ectopic pancreases from UGI submucosal tumors.