Improving the Accuracy of CT Scanning in the Diagnosis of Small Bowel Obstruction: To Scan or Not to Scan…Again?

Abstract

Introduction: Although small bowel obstruction (SBO) is a frequent occurrence, 80% resolve with conservative management. CT scanning is a helpful adjunct in diagnosing SBO but has variable sensitivity, ranging anywhere from 60-85% depending on scan technique, timing, and radiologist expertise. No standard practice exists at our institution for the radiographic evaluation of SBO. Instead, various contrasted CT protocols are used at varying intervals at the discretion of the ordering physician. We sought to determine how accurate our radiographic diagnosis of SBO was compared to the clinical diagnosis, as well as to determine the optimal modality and timing sequence of CT imaging for evaluating SBO. Methods: A retrospective chart review from July 2012 to July 2013 identified consecutive patients at our facility who underwent CT scans that resulted in a diagnosis of partial or full SBO. We analyzed their clinical outcomes to see if their clinical diagnosis was concordant with their radiographic diagnosis. Subsequently we determined the accuracy of our radiographic testing. Results: From July 2012 to July 2013, 136 patients underwent CT scans that resulted in a radiographic diagnosis of partial or full SBO. The positive predictive value of a CT scan matching the clinical diagnosis was 66.7%. Using Chi-squared analysis, there was no association between either enteral or intravenous contrast improving the diagnostic accuracy of the original CT scan. Undergoing delayed CT scanning with PO contrast after the original scan increased the positive predictive value to 88.5% for a CT scan matching the clinical diagnosis. The median time for delayed imaging was 8 hours after the original CT scan, while the interquartile range for repeat scanning was 9 hours. Undergoing more than one CT scan resulted in a change in diagnosis in 40% of the cases. Conclusion: CT scanning is recommended by several professional societies for the diagnosis and management of SBO. In our hands, it did not match the clinical diagnosis in 1/3 of the cases. Additional delayed imaging is helpful in improving the accuracy of the diagnosis; however, at our institution there is no standard protocol for when or how to obtain additional imaging. This data will be used at our institution to develop a protocol for management of SBO to reduce unnecessary scanning while improving diagnostic accuracy.