Jin Woong Kim MD, Presenter: Nothing to Disclose

Yong Yeon Jeong, Abstract Co-Author: Nothing to Disclose

Woong Yoon, Abstract Co-Author: Nothing to Disclose

Sang Soo Shin MD, Abstract Co-Author: Nothing to Disclose

Suk Hee Heo MD, Abstract Co-Author: Nothing to Disclose

Heoung Keun Kang, Abstract Co-Author: Nothing to Disclose

To retrospectively compare the diagnostic accuracies of arterial and portal venous phase 64-slice multi-detector row CT (MDCT) for detection and localization of acute gastrointestinal (GI) bleeding.

Forty-six consecutive patients with suspicious acute GI bleeding (29 men, 17 women; age range, 20-93 years; mean age, 61) underwent 64-slice MDCT scan with arterial and portal venous phase, before angiography. Extravasated contrast medium within bowel lumen was considered as GI bleeding. Two abdominal radiologists independently reviewed arterial phase with nonenhanced image (session A), portal venous phase with nonenhanced image (session P), and arterial with portal venous phases and nonenhanced images (session AP). Arterial and portal venous phase 64-slice MDCT findings were compared with results of reference standard as conventional angiography. Sensitivities, specificities, and accuracies of three sessions on basis of a per-location analysis were calculated for each reader. Interobserver agreements (κ) of between reader I and II on each session were also assessed.

GI bleeding on angiography was depicted in 32 of 46 patients. For detection of acute GI bleeding, sensitivities, specificities, and accuracies of session A on basis of a per-location analysis were 81.3%, 98%, and 96.1% in reader I and 84.4%, 98.5%, and 96.1% in reader II, respectively. Sensitivities, specificities, and accuracies of session P were 81.3%, 99.5%, and 97% in reader I and 84.4%, 98.5%, and 96.5% in reader II, respectively. Sensitivities, specificities, and accuracies of session AP were 84.4%, 99.5%, and 97.4% in reader I and 84.4%, 97%, and 95.2% in reader II, respectively. Interobserver agreement (κ) between two readers on session A, session P and session AP were 0.91, 0.93 and 0.89, respectively.

The diagnostic accuracies of 64-slice MDCT for detection and localization of acute GI bleeding were not different from session A, session P and session AP. Interobserver agreements between two readers on all sessions were almost perfect agreement. Therefore, one phase 64-slice MDCT scan among arterial and portal venous phase scans may be effective to detect and localize acute GI bleeding.

Arterial and portal venous phases 64-slice MDCT is effective to detect and localize acute GI bleeding.

Kim, J, Jeong, Y, Yoon, W, Shin, S, Heo, S, Kang, H, Diagnostic Accuracy of Arterial and Portal Venous Phase 64-slice MDCT in Dectection and Localizatoin of Acute GI Bleeding.  Radiological Society of North America 2008 Scientific Assembly and Annual Meeting, February 18 - February 20, 2008 ,Chicago IL. http://archive.rsna.org/2008/6020085.html