Artificial Intelligence to Assist in Exclusion of Coronary Atherosclerosis During CCTA Evaluation of Chest Pain in the Emergency Department: Preparing an Application for Real-world Use.
Artificial intelligence
Chest pain
Coronary atherosclerosis
Coronary computed tomography angiography
Journal
Journal of digital imaging
ISSN: 1618-727X
Titre abrégé: J Digit Imaging
Pays: United States
ID NLM: 9100529
Informations de publication
Date de publication:
06 2021
06 2021
Historique:
received:
25
05
2020
accepted:
01
03
2021
revised:
09
11
2020
pubmed:
2
4
2021
medline:
28
9
2021
entrez:
1
4
2021
Statut:
ppublish
Résumé
Coronary computed tomography angiography (CCTA) evaluation of chest pain patients in an emergency department (ED) is considered appropriate. While a "negative" CCTA interpretation supports direct patient discharge from an ED, labor-intensive analyses are required, with accuracy in jeopardy from distractions. We describe the development of an artificial intelligence (AI) algorithm and workflow for assisting qualified interpreting physicians in CCTA screening for total absence of coronary atherosclerosis. The two-phase approach consisted of (1) phase 1-development and preliminary testing of an algorithm for vessel-centerline extraction classification in a balanced study population (n = 500 with 50% disease prevalence) derived by retrospective random case selection, and (2) phase 2-simulated clinical Trialing of developed algorithm on a per-case (entire coronary artery tree) basis in a more "real-world" study population (n = 100 with 28% disease prevalence) from an ED chest pain series. This allowed pre-deployment evaluation of the AI-based CCTA screening application which provides vessel-by-vessel graphic display of algorithm inference results integrated into a clinically capable viewer. Algorithm performance evaluation used area under the receiver operating characteristic curve (AUC-ROC); confusion matrices reflected ground truth vs AI determinations. The vessel-based algorithm demonstrated strong performance with AUC-ROC = 0.96. In both phase 1 and phase 2, independent of disease prevalence differences, negative predictive values at the case level were very high at 95%. The rate of completion of the algorithm workflow process (96% with inference results in 55-80 s) in phase 2 depended on adequate image quality. There is potential for this AI application to assist in CCTA interpretation to help extricate atherosclerosis from chest pain presentations.
Identifiants
pubmed: 33791909
doi: 10.1007/s10278-021-00441-6
pii: 10.1007/s10278-021-00441-6
pmc: PMC8329136
doi:
Types de publication
Journal Article
Research Support, Non-U.S. Gov't
Langues
eng
Sous-ensembles de citation
IM
Pagination
554-571Informations de copyright
© 2021. Society for Imaging Informatics in Medicine.
Références
BMC Med Genomics. 2011 Apr 08;4:31
pubmed: 21477282
J Am Coll Cardiol. 2013 Feb 26;61(8):880-92
pubmed: 23395069
Int J Cardiol. 2015;187:565-80
pubmed: 25863305
N Engl J Med. 2012 Apr 12;366(15):1393-403
pubmed: 22449295
Eur J Radiol. 2019 May;114:1-5
pubmed: 31005158
Comput Math Methods Med. 2016;2016:1835297
pubmed: 27721896
N Engl J Med. 2000 Apr 20;342(16):1163-70
pubmed: 10770981
Radiographics. 2015 Oct;35(6):1668-76
pubmed: 26466178
J Cardiovasc Comput Tomogr. 2018 May - Jun;12(3):204-209
pubmed: 29753765
Nat Med. 2019 Jan;25(1):44-56
pubmed: 30617339
Eur Heart J Cardiovasc Imaging. 2017 Oct 1;18(10):1170-1178
pubmed: 27679600
Sci Rep. 2019 Feb 4;9(1):1308
pubmed: 30718631
JACC Cardiovasc Imaging. 2013 Sep;6(9):963-72
pubmed: 23932641
Med Image Anal. 2009 Oct;13(5):701-14
pubmed: 19632885
J Am Coll Radiol. 2016 Feb;13(2):e1-e29
pubmed: 26810814
AJR Am J Roentgenol. 2013 Sep;201(3):611-7
pubmed: 23971454
J Med Imaging (Bellingham). 2020 Jan;7(1):016502
pubmed: 32064302
Acad Radiol. 2019 Jun;26(6):833-845
pubmed: 30559033
Radiology. 2019 Aug;292(2):354-362
pubmed: 31237495
Eur Heart J. 2015 Apr 14;36(15):906-14
pubmed: 25687351
Radiographics. 2017 Mar-Apr;37(2):505-515
pubmed: 28212054
Inhal Toxicol. 2014 Nov;26(13):811-28
pubmed: 25264934
Medicine (Baltimore). 2009 Sep;88(5):307-313
pubmed: 19745690
Radiol Artif Intell. 2019 Nov 27;1(6):e180095
pubmed: 33937804
Radiology. 2019 Mar;290(3):590-606
pubmed: 30694159
Radiology. 2018 Aug;288(2):318-328
pubmed: 29944078
J Am Coll Cardiol. 2018 Jun 5;71(22):2511-2522
pubmed: 29852975
J Cardiovasc Comput Tomogr. 2012 May-Jun;6(3):154-63
pubmed: 22551595
Contemp Clin Trials Commun. 2019 Aug 22;16:100434
pubmed: 31485545
Proc SPIE Int Soc Opt Eng. 2019 Feb;10949:
pubmed: 31762536
Eur Heart J Cardiovasc Imaging. 2020 May 1;21(5):479-488
pubmed: 32065624
Med Image Comput Comput Assist Interv. 2013;16(Pt 3):74-81
pubmed: 24505746
Am Heart J. 2013 Feb;165(2):154-63.e3
pubmed: 23351817
J Digit Imaging. 2020 Apr;33(2):431-438
pubmed: 31625028
J Cardiovasc Comput Tomogr. 2016 Nov - Dec;10(6):435-449
pubmed: 27780758
Curr Cardiovasc Imaging Rep. 2013 Apr 1;6(2):89-101
pubmed: 23524381
IEEE Trans Med Imaging. 2019 Jul;38(7):1588-1598
pubmed: 30507498
AJR Am J Roentgenol. 2018 Apr;210(4):816-820
pubmed: 29446681
J Cardiovasc Comput Tomogr. 2014 Sep-Oct;8(5):359-67
pubmed: 25301041
Eur Heart J Cardiovasc Imaging. 2020 May 1;21(5):489-490
pubmed: 32068781
Radiographics. 2017 Nov-Dec;37(7):2113-2131
pubmed: 29131760
N Engl J Med. 2012 Jul 26;367(4):299-308
pubmed: 22830462
Circulation. 2010 Oct 26;122(17):1756-76
pubmed: 20660809
IEEE Trans Biomed Eng. 2012 Jul;59(7):1850-60
pubmed: 22180504