Detectability of pancreas divisum in patients with acute pancreatitis on multi-detector row computed tomography

Abstract

The purpose of this study was to retrospectively assess the diagnostic performance of multi-detector row computed tomography (MDCT) in an evaluation of pancreas divisum using endoscopic retrograde pancreatography (ERP) as the reference standard. We analyzed 41 consecutive patients (14 cases of pancreas divisum and 27 cases of standard anatomy) who had undergone both MDCT and ERP for the evaluation of clinically diagnosed acute pancreatitis between November 2004 and June 2007. The CT reconstruction thickness and interval were both 3 mm. Two radiologists independently reviewed CT data, and the diagnostic confidence in determining the pancreatic ductal anatomy was scored using a five-point scale. CT detectability was correlated with the severity of pancreatitis and the degree of pancreatic necrosis based on the Balthazar index. With consensus, 16 of 41 cases (39.0%) were evaluated as indeterminate. Ductal anatomy was correctly diagnosed in 23 of 41 cases (56.1%). Eight of 14 cases (57.1%) were correctly diagnosed as pancreas divisum. Standard anatomy was identified in 15 of 27 cases (55.6%). The inter-observer agreement was substantial (κ = 0.71). Grade B or more pancreatitis and the presence of pancreatic necrosis significantly influenced the evaluation of ductal anatomy (p = 0.01 and p < 0.01, respectively). Pancreas divisum was correctly diagnosed in the case of grade A acute pancreatitis. The CT detectability of pancreas divisum in patients with grade B or more pancreatitis is still relatively low even in the MDCT era.