KCR 2012








Notification of abstract acceptance announced via E-mail on July 31st.

Abstract view

Oct. 18 (Thu) 203 Room - SS12 CV

17:00-17:10 [SS 12 CV-06]
Diagnostic Performance of Automatic Computer-assisted Algorithm for the Detection of Significant Coronary Artery Disease in Patients with Acute Chest Pain at Emergency Department


● Speaker	Sunyoung Lee (Samsung Medical Center)
● Authors	Sunyoung Lee,Yeon Hyeon Choe,Sung Mok Kim,Ji Hye Min
● Affiliation	Samsung Medical Center
● PDF View	SS 12 CV-06.pdf

PURPOSE:
The purpose of our study was to evaluate the performance of computer-aided algorithm for automated stenosis detection at coronary CT angiography (CCTA).

MATERIALS AND METHODS:
We investigated 131 consecutive patients (87 male, mean age 65±12 years) who had acute chest pain and underwent 128-slice dual-source CCTA and coronary angiography (CAG) between June 2009 and August 2011 in the emergency department (ED). All CCTA data were analyzed using a software algorithm for automated, without human interaction, detection of coronary artery stenosis. The performance of the automatic computer-assisted detection (auto-CAD) for evaluation of stenosis of 50% or more was compared with CAG. In addition, the accuracy of the semi-quantitative assessment of CCTA by two experienced radiologists was compared with CAG.

RESULTS:
In 111 of 131 patients (20 were excluded due to failure of data processing or previous history of stent insertion / CABG), CAG demonstrated significant coronary artery stenosis in 65 of 111 patients (58.6%) of which the auto-CAD algorithm correctly identified 64 (98.5%). For the detection of the 50% or more coronary artery stenosis, per-patient analysis of auto-CAD revealed the following: 98.5% sensitivity, 34.8% specificity, 68.1% positive predictive value (PPV), 94.1% negative predictive value (NPV). Per-vessel analysis of auto-CAD showed 88.4% sensitivity, 58.4% specificity, 45.9% PPV, 92.7% NPV. The visual inspection of CCTA had 96.8%/98.5 sensitivity, 94.5%/82.6% specificity, 87.6% /88.9% PPV, and 98.7%/97.4 NPV for diagnosing stenosis of >50% on per-vessel/per-patient analysis, respectively.

CONCLUSION:
Auto-CAD algorithm showed the high NPV for the detection of >50% CAD on CCTA in the setting of acute chest pain. If used as a second reader, the auto-CAD algorithm can be used to exclude significant stenosis and facilitate the decision-making process in the ED.

CLINICAL RELEVANCE/APPLICATION:
Auto-CAD algorithm may enhance the confidence of human interpreters for excluding significant stenosis in coronary arteries at coronary CT angiography.