Is nonenhanced magnetic resonance imaging comparable with contrast enhanced computed tomography for assessing severity and predicting outcome in patients with acute pancreatitis?

Nonenhanced magnetic resonance imaging (NEMRI) is a reasonable alternative to contrast enhanced computed tomography (CECT) to assess complications of acute pancreatitis, and provides the advantage of detailed biliary tract imaging. Although MRI is more time-consuming and may have higher costs than CECT, accurate imaging while avoiding intravenous contrast will be particularly useful for patients with renal disease or contrast allergy. (LOE = 1b)

Stimac D, Miletic D, Radic M, et al. The role of nonenhanced magnetic resonance imaging in the early assessment of acute pancreatitis. Am J Gastroenterol 2007;102:997-1004.  [PMID:17378903]

Diagnostic test evaluation

Self-funded or unfunded

Inpatient (any location)

Consecutive patients admitted with acute pancreatitis were included. Between the third and fifth hospital days, patients underwent spiral computed tomography with intravenous contrast and NEMRI, including MR cholangiopancreatography (MRCP). Fourteen patients were excluded because of contrast allergy, metal implants, and inability to undergo MRI because of mechanical ventilation. Two radiologists who were masked to patients' clinical data independently reviewed both imaging studies. Disease severity on imaging was graded by the Balthazar grade, a scoring system based on radiographic findings of pancreatic enlargement, inflammation, fluid collections, and percent necrosis. This was compared with clinical disease severity, as assessed primarily with the Atlanta criteria, and secondarily with several other prognostic measures including the Ranson score, the Acute Physiology and Chronic Health Outcomes (APACHE) II score, and C-reactive protein levels. A total of 101 patients were studied. The patients' median age was 62 years and the causes of acute pancreatitis were primarily biliary (63%) and alcoholic (18%). Pancreatitis was classified as severe according to the Atlanta score in approximately 33% of patients, and by the Ranson score in 64% (Ransom score > 2). CECT and NEMRI showed significant correlation for Balthazar grade, assessment of pancreatic necrosis, and combined severity indices (Balthazar grade plus points for percent necrosis). Magnetic resonance severity indices also correlated with the Ranson score, systemic complications, and length of hospital stay. Using CECT as the gold standard, NEMRI had a sensitivity of 79% and specificity 92% for detection of severe pancreatitis (positive likelihood ratio = 10; negative likelihood ratio = 0.23).