Chester Chong MBBCh, MRCS, Presenter: Nothing to Disclose

Kiang Hiong Tay FRCR, Abstract Co-Author: Nothing to Disclose

Mark Christiaan Burgmans, Abstract Co-Author: Nothing to Disclose

Manish Taneja, Abstract Co-Author: Nothing to Disclose

Bien Soo Tan, Abstract Co-Author: Institutional research collaboration, Koninklijke Philips Electronics NV Institutional research collaboration, Siemens AG Institutional research collaboration, Toshiba Corporation

Purpose: To retrospectively evaluate the accuracy of multidetector row computed tomography angiography (MDCTA) for detection and localisation of acute gastrointestinal (GI) bleeding.

Methods: Between 2007 to 2010, we identified 44 patients (26 men, 18 women; mean age 63; range 27 to 88) from our Radiology Information System who underwent both MDCTA and digital subtraction angiography (DSA) for evaluation of acute GI bleeding. Five patients had repeat MDCTA and DSA. All MDCTAs were performed before DSAs. Fourteen patients underwent embolotherapy while 16 patients underwent laparotomy following MDCTA/DSA. Forty patients underwent endoscopy prior to MDCTA/DSA. Extravasation of contrast material into the lumen of GI tract was considered positive for acute GI bleeding. The site of contrast extravasation in each anatomical location was recorded. MDCTA were compared using DSA and combined endoscopy and laparotomy findings as reference standards. Accuracy for localisation of acute GI bleeding was assessed by comparing locations of active bleeding at both MDCTA and DSA patients who had positive scans.

Results: MDCTA depicted extravasation of contrast in 29 of 51 scans. The sensitivity, specificity, accuracy, and positive and negative predictive values of MDCTA for detection of GI bleeding using DSA as reference standard were 100%, 68%, 80%, 67% and 100% respectively, while the values using combined endoscopy and laparotomy findings as reference standard were 67%, 83%, 73%, 88% and 59% respectively. The sensitivity, specificity, accuracy, and positive and negative predictive values of DSA for detection of GI bleeding using combined endoscopy and laparotomy findings as reference standard were 43%, 93%, 63%, 90% and 52% respectively. The location of contrast material extravasation on MDCTA corresponded exactly to that of active bleeding on DSA in all patients.

Conclusion: MDCTA appears to be more sensitive and more accurate than DSA for detection and localization of acute GI bleed when referenced against laparotomy/endoscopy findings.

The study shows that MDCTA is more sensitive and accurate than DSA when referenced against endoscopy and laparotomy findings, thus supporting MDCTA as first line study in acute gastrointestinal bleed.

Chong, C, Tay, K, Burgmans, M, Taneja, M, Tan, B, Multidetector Row CT Angiography (MDCTA) vs Digital Subtraction Angiography (DSA) for Detection and Localization of Acute Gastrointestinal Bleed.  Radiological Society of North America 2011 Scientific Assembly and Annual Meeting, November 26 - December 2, 2011 ,Chicago IL. http://archive.rsna.org/2011/11007457.html