Abdominal imaging studies: comparison of diagnostic accuracies resulting from ultrasound, computed tomography, and magnetic resonance imaging in the same individual.

We retrospectively compared the diagnostic accuracy of abdominal ultrasound (US), computed tomography (CT), and magnetic resonance imaging (MRI) studies performed on the same individual to determine the relative performance of these modalities in the evaluation of disease processes, arising from different intra-abdominal organ systems. We retrospectively reviewed all procedure codes accrued by our abdominal imaging section during a 1-year period to determine how many patients underwent all three imaging procedures in our institution within a 2-week interval. These cases were then further evaluated to determine: (1) the primary organ system of disease involvement, (2) the final diagnosis, and (3) the imaging modality that provided the most accurate information upon which appropriate medical management was based. Imaging findings were determined by review of diagnostic reports, and medical management was determined by chart review. Two thousand six hundred-ninety five patients underwent ultrasound, 4,394 patients underwent CT, and 872 patients underwent MRI for the investigation of abdominal disease. Among these 5,126 patients, 26 underwent sequential US, CT, and MRI evaluation within a two-week interval. Appropriate clinical management was based on the US findings in 12/26 cases (46%; 95% CI: 27-67%), CT findings in 16/26 cases (62%; 95% CI: 41-80%), and MRI findings in 24 cases (92%; 95% CI: 75-99%). Significant differences in directing appropriate clinical management were found when comparing the relative diagnostic accuracies of MRI to US (p = 0.0003) and MRI to CT (p = 0.009). However, differences were not significant when comparing CT to US (p = 0.27). All final diagnoses in a given patient were accurately determined by US in 7/26 cases (27%), by CT in 10/26 cases (38%), and by MRI in 23/26 cases (88%). When analyzed according to individual diagnoses (61 total), US detected 33 of 61 (54%) abnormalities, CT detected 41 (67%) abnormalities, and MRI detected 51 (84%) abnormalities. US correctly characterized 32 (52%) abnormalities, CT correctly characterized 36 (59%) abnormalities, and MRI correctly characterized 49 (80%) abnormalities. Ultrasound, CT, and MR imaging correctly diagnosed the disease process in 3 of 12 (25%), 3 of 12 (25%), and 10 of 12 (83%) patients, respectively, with liver disease; 2 of 9 (22%), 6 of 9 (66%), and 9 of 9 (100%) patients, respectively, with pancreatic disease; 17 of 17 (100%), 8 of 17 (47%), and 10 of 17 (58%) patients, respectively, with gallbladder and biliary disease; 2 of 8 (25%), 5 of 8 (63%), and 5 of 8 (63%) patients, respectively, with renal disease; 0 of 3 (0%), 2 of 3 (66%), and 3 of 3 (100%) patients, respectively, with adrenal disease; and 8 of 12 (75%), 12 of 12 (100%), and 12 of 12 (100%) patients with free intraperitoneal fluid. Our results provide new information, regarding the relative benefits of ultrasound, CT, and MRI for the investigation of abdominal diseases at our institution. This initial data suggests that ultrasound provides the most accurate diagnoses in the investigation of gallbladder disease; MRI provides the most accurate diagnoses in the investigation of hepatic, adrenal, and pancreatic disease; and either CT or MRI may be the most appropriate first imaging study for the detection of renal disease.