Feasibility of an automatic computer-assisted algorithm for the detection of significant coronary artery disease in patients presenting with acute chest pain

Ki Woon Kang, Hyuk Jae Chang, Hackjoon Shim, Young Jin Kim, Byoung Wook Choi, Woo In Yang, Jee Young Shim, Jongwon Ha, Namsik Chung

Research output: Contribution to journalArticle

15 Citations (Scopus)

Abstract

Automatic computer-assisted detection (auto-CAD) of significant coronary artery disease (CAD) in coronary computed tomography angiography (cCTA) has been shown to have relatively high accuracy. However, to date, scarce data are available regarding the performance of auto-CAD in the setting of acute chest pain. This study sought to demonstrate the feasibility of an auto-CAD algorithm for cCTA in patients presenting with acute chest pain. We retrospectively investigated 398 consecutive patients (229 male, mean age 50 ± 21 years) who had acute chest pain and underwent cCTA between Apr 2007 and Jan 2011 in the emergency department (ED). All cCTA data were analyzed using an auto-CAD algorithm for the detection of >50% CAD on cCTA. The accuracy of auto-CAD was compared with the formal radiology report. In 380 of 398 patients (18 were excluded due to failure of data processing), per-patient analysis of auto-CAD revealed the following: sensitivity 94%, specificity 63%, positive predictive value (PPV) 76%, and negative predictive value (NPV) 89%. After the exclusion of 37 cases that were interpreted as invalid by the auto-CAD algorithm, the NPV was further increased up to 97%, considering the false-negative cases in the formal radiology report, and was confirmed by subsequent invasive angiogram during the index visit. We successfully demonstrated the high accuracy of an auto-CAD algorithm, compared with the formal radiology report, for the detection of >50% CAD on cCTA in the setting of acute chest pain. The auto-CAD algorithm can be used to facilitate the decision-making process in the ED.

Original languageEnglish
Pages (from-to)e640-e646
JournalEuropean Journal of Radiology
Volume81
Issue number4
DOIs
Publication statusPublished - 2012 Apr 1

Fingerprint

Acute Pain
Chest Pain
Coronary Artery Disease
Radiology
Hospital Emergency Service
Computed Tomography Angiography
Decision Making
Angiography
Sensitivity and Specificity

All Science Journal Classification (ASJC) codes

  • Radiology Nuclear Medicine and imaging

Cite this

@article{16e086ac63db4f5b912fa69f20b90f14,
title = "Feasibility of an automatic computer-assisted algorithm for the detection of significant coronary artery disease in patients presenting with acute chest pain",
abstract = "Automatic computer-assisted detection (auto-CAD) of significant coronary artery disease (CAD) in coronary computed tomography angiography (cCTA) has been shown to have relatively high accuracy. However, to date, scarce data are available regarding the performance of auto-CAD in the setting of acute chest pain. This study sought to demonstrate the feasibility of an auto-CAD algorithm for cCTA in patients presenting with acute chest pain. We retrospectively investigated 398 consecutive patients (229 male, mean age 50 ± 21 years) who had acute chest pain and underwent cCTA between Apr 2007 and Jan 2011 in the emergency department (ED). All cCTA data were analyzed using an auto-CAD algorithm for the detection of >50{\%} CAD on cCTA. The accuracy of auto-CAD was compared with the formal radiology report. In 380 of 398 patients (18 were excluded due to failure of data processing), per-patient analysis of auto-CAD revealed the following: sensitivity 94{\%}, specificity 63{\%}, positive predictive value (PPV) 76{\%}, and negative predictive value (NPV) 89{\%}. After the exclusion of 37 cases that were interpreted as invalid by the auto-CAD algorithm, the NPV was further increased up to 97{\%}, considering the false-negative cases in the formal radiology report, and was confirmed by subsequent invasive angiogram during the index visit. We successfully demonstrated the high accuracy of an auto-CAD algorithm, compared with the formal radiology report, for the detection of >50{\%} CAD on cCTA in the setting of acute chest pain. The auto-CAD algorithm can be used to facilitate the decision-making process in the ED.",
author = "Kang, {Ki Woon} and Chang, {Hyuk Jae} and Hackjoon Shim and Kim, {Young Jin} and Choi, {Byoung Wook} and Yang, {Woo In} and Shim, {Jee Young} and Jongwon Ha and Namsik Chung",
year = "2012",
month = "4",
day = "1",
doi = "10.1016/j.ejrad.2012.01.017",
language = "English",
volume = "81",
pages = "e640--e646",
journal = "European Journal of Radiology",
issn = "0720-048X",
publisher = "Elsevier Ireland Ltd",
number = "4",

}

Feasibility of an automatic computer-assisted algorithm for the detection of significant coronary artery disease in patients presenting with acute chest pain. / Kang, Ki Woon; Chang, Hyuk Jae; Shim, Hackjoon; Kim, Young Jin; Choi, Byoung Wook; Yang, Woo In; Shim, Jee Young; Ha, Jongwon; Chung, Namsik.

In: European Journal of Radiology, Vol. 81, No. 4, 01.04.2012, p. e640-e646.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Feasibility of an automatic computer-assisted algorithm for the detection of significant coronary artery disease in patients presenting with acute chest pain

AU - Kang, Ki Woon

AU - Chang, Hyuk Jae

AU - Shim, Hackjoon

AU - Kim, Young Jin

AU - Choi, Byoung Wook

AU - Yang, Woo In

AU - Shim, Jee Young

AU - Ha, Jongwon

AU - Chung, Namsik

PY - 2012/4/1

Y1 - 2012/4/1

N2 - Automatic computer-assisted detection (auto-CAD) of significant coronary artery disease (CAD) in coronary computed tomography angiography (cCTA) has been shown to have relatively high accuracy. However, to date, scarce data are available regarding the performance of auto-CAD in the setting of acute chest pain. This study sought to demonstrate the feasibility of an auto-CAD algorithm for cCTA in patients presenting with acute chest pain. We retrospectively investigated 398 consecutive patients (229 male, mean age 50 ± 21 years) who had acute chest pain and underwent cCTA between Apr 2007 and Jan 2011 in the emergency department (ED). All cCTA data were analyzed using an auto-CAD algorithm for the detection of >50% CAD on cCTA. The accuracy of auto-CAD was compared with the formal radiology report. In 380 of 398 patients (18 were excluded due to failure of data processing), per-patient analysis of auto-CAD revealed the following: sensitivity 94%, specificity 63%, positive predictive value (PPV) 76%, and negative predictive value (NPV) 89%. After the exclusion of 37 cases that were interpreted as invalid by the auto-CAD algorithm, the NPV was further increased up to 97%, considering the false-negative cases in the formal radiology report, and was confirmed by subsequent invasive angiogram during the index visit. We successfully demonstrated the high accuracy of an auto-CAD algorithm, compared with the formal radiology report, for the detection of >50% CAD on cCTA in the setting of acute chest pain. The auto-CAD algorithm can be used to facilitate the decision-making process in the ED.

AB - Automatic computer-assisted detection (auto-CAD) of significant coronary artery disease (CAD) in coronary computed tomography angiography (cCTA) has been shown to have relatively high accuracy. However, to date, scarce data are available regarding the performance of auto-CAD in the setting of acute chest pain. This study sought to demonstrate the feasibility of an auto-CAD algorithm for cCTA in patients presenting with acute chest pain. We retrospectively investigated 398 consecutive patients (229 male, mean age 50 ± 21 years) who had acute chest pain and underwent cCTA between Apr 2007 and Jan 2011 in the emergency department (ED). All cCTA data were analyzed using an auto-CAD algorithm for the detection of >50% CAD on cCTA. The accuracy of auto-CAD was compared with the formal radiology report. In 380 of 398 patients (18 were excluded due to failure of data processing), per-patient analysis of auto-CAD revealed the following: sensitivity 94%, specificity 63%, positive predictive value (PPV) 76%, and negative predictive value (NPV) 89%. After the exclusion of 37 cases that were interpreted as invalid by the auto-CAD algorithm, the NPV was further increased up to 97%, considering the false-negative cases in the formal radiology report, and was confirmed by subsequent invasive angiogram during the index visit. We successfully demonstrated the high accuracy of an auto-CAD algorithm, compared with the formal radiology report, for the detection of >50% CAD on cCTA in the setting of acute chest pain. The auto-CAD algorithm can be used to facilitate the decision-making process in the ED.

UR - http://www.scopus.com/inward/record.url?scp=84862811911&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84862811911&partnerID=8YFLogxK

U2 - 10.1016/j.ejrad.2012.01.017

DO - 10.1016/j.ejrad.2012.01.017

M3 - Article

C2 - 22304980

AN - SCOPUS:84862811911

VL - 81

SP - e640-e646

JO - European Journal of Radiology

JF - European Journal of Radiology

SN - 0720-048X

IS - 4

ER -