Prospective evaluation of the value of magnetic resonance imaging in suspected acute sigmoid diverticulitis.

PURPOSE: The purpose of this study was to prospectively investigate patients with suspected acute colonic diverticulitis and to provide sensitivity, specificity, and interobserver agreement in a blinded trial.
METHODS: Fifty-five patients (29 men; 59 +/- 13 (range, 29-76) years) who reported to the emergency room with clinically suspected acute colonic diverticulitis were prospectively included in the study. All patients underwent magnetic resonance imaging scans of their abdomen before and after contrast agent administration. Two assessors blinded to all clinical, laboratory, and radiologic results evaluated the images separately.
RESULTS: The assessors reported colonic wall thickening, segmental narrowing of the colon, presence of diverticula, pericolic fatty infiltration, ascites, and abscesses. The assessors had to diagnose or rule out acute colonic diverticulitis. Sensitivities, specificities, positive, and negative likelihood ratios were derived. To determine interobserver agreement, a Cohen's kappa coefficient was calculated. The two assessors exhibited sensitivities of more than 94 percent, specificities of 88 percent, positive likelihood ratios of more than 7.5, and negative likelihood ratios of less than 0.07. The kappa coefficient showed a significant, strong correlation between both assessors (kappa = 0.68).
CONCLUSIONS: Magnetic resonance imaging is investigator independent and provides high sensitivity and specificity for the diagnosis of acute colonic diverticulitis.