The use of magnetic resonance elastography in differentiating autoimmune pancreatitis from pancreatic ductal adenocarcinoma: A preliminary study

Abstract

PurposeTo assess the value of magnetic resonance elastography (MRE) in patients with autoimmune pancreatitis (AIP) and in the differentiation of AIP from pancreatic ductal adenocarcinoma (PDAC). Method and materialsThis prospective study included 14 AIP patients, 26 PDAC patients, and 14 healthy volunteers. All participants underwent pancreatic MRE (40-Hz; 3 T scanner) at enrollment, and 7 AIP patients underwent a second MRE after initiation of steroid therapy. Pancreatic stiffness values were obtained by MRE and a new logistic regression model (the calculated Rad score) was used to combine pancreatic stiffness and the distribution and shape of high-stiffness areas for differentiation of AIP and PDAC. The area under the curve (AUC) was calculated for all parameters using receiver operating characteristic (ROC) analysis. ResultsPancreatic stiffness was significantly higher (2.67 kPa [interquartile range, 2.24–3.56 kPa]) in AIP than in healthy pancreas (1.24 kPa [1.18–1.24 kPa]) and significantly lower in AIP than in PDAC (3.78 kPa [3.22–5.11 kPa]; both P < 0.05). Diffuse (n = 4 vs 1; P = 0.043) and multiple (n = 3 vs 0; P = 0.037) lesions were more common in AIP, while solitary (n = 25 vs 7; P = 0.001) and nodular lesions (n = 18 vs 2; P = 0.002) were more frequent in PDAC. Rad scores outperformed individual imaging parameters in distinguishing AIP from PDAC (AUC, 0.948 vs 0.607 to 0.782; all P < 0.05), with 84.6% specificity and 92.9% sensitivity. Pancreatic stiffness in AIP decreased significantly, from 2.66 kPa [2.29 to 3.05 kPa] to 1.55 kPa [1.43 to 1.67 kPa] (P = 0.016), during treatment. ConclusionsMRE shows promise as a quantitative imaging method for differentiating AIP from PDAC and for monitoring the treatment response in AIP.