BACKGROUND AND AIMS: Computed tomography (CT), especially contrast-enhanced CT (CECT), provides important information on the severity and prognosis of acute pancreatitis (AP). Magnetic resonance imaging (MRI) has become a useful tool as an alternative to CT in the assessment of AP. The primary aim of our study was to determine the diagnostic value of nonenhanced MRI (NEMRI) to assess severity and predict outcome in patients with AP from the third to fifth day after admission. We also correlated MRI findings with CT and biochemical parameters. PATIENTS AND METHODS: The study included 101 patients (49 men, 52 women, median age 62 yr, range 20-82) with a diagnosis of AP admitted to our hospital between January 1, 2004 and June 31, 2005. The inclusion criteria consisted of a combination of clinical features, a typical case history, elevation of serum pancreatic enzymes, and diagnosis confirmed by imaging studies. Contrast-enhanced spiral CT exams were performed in all patients from the third to fifth day after admission, and Balthazar grade and CT severity index were calculated. All patients underwent NEMRI, and MR severity index (MRSI) was calculated. We also performed magnetic resonance cholangiopancreatography (MRCP) in all patients to detect bile duct lithiasis. RESULTS: Significant correlation between CECT and NEMRI was found for Balthazar grade (P < 0.001) and the assessment of pancreatic necrosis (P < 0.001), as well as between the combined severity indices (rho = 0.819, P < 0.001). MRSI correlated with Ranson score (rho = 0.656, P < 0.01), C-reactive protein (CRP) levels 48 h after admission (rho = 0.502, P < 0.01), appearance of systemic complications (rho = 0.576, P < 0.01), and length of hospital stay (rho = 0.484, P < 0.01). Considering the Atlanta criteria as the gold standard and the Ranson score, no difference in sensitivity, specificity, positive predictive value, negative predictive value, and accuracy of the two methods was observed. Comparing the group of patients with presumed acute pancreatic hemorrhage with the group of patients with severe AP, we found a significantly higher APACHE II score on the first day (P < 0.05), that the development of systemic complications was more frequent (P < 0.05), and that the hospital stay and ICU management of patients with MRI signs of pancreatic hemorrhage tended to be longer. CONCLUSION: NEMRI is comparable to CECT in the early assessment of the severity of AP, and both methods are equally efficient in predicting local and systemic complications of AP. MRI has a potential advantage over CT in detecting bile duct lithiasis and pancreatic hemorrhage.