Deep learning-based atherosclerotic coronary plaque segmentation on coronary CT angiography.
Computed tomography angiography
Coronary artery disease
Deep learning
Journal
European radiology
ISSN: 1432-1084
Titre abrégé: Eur Radiol
Pays: Germany
ID NLM: 9114774
Informations de publication
Date de publication:
Oct 2022
Oct 2022
Historique:
received:
17
01
2022
accepted:
03
04
2022
revised:
31
03
2022
pubmed:
8
5
2022
medline:
17
9
2022
entrez:
7
5
2022
Statut:
ppublish
Résumé
Volumetric evaluation of coronary artery disease (CAD) allows better prediction of cardiac events. However, CAD segmentation is labor intensive. Our objective was to create an open-source deep learning (DL) model to segment coronary plaques on coronary CT angiography (CCTA). Three hundred eight individuals' 894 CCTA scans with 3035 manually segmented plaques by an expert reader (considered as ground truth) were used to train (186/308, 60%), validate (tune, 61/308, 20%), and test (61/308, 20%) a 3D U-net model. We also evaluated the model on an external test set of 50 individuals with vulnerable plaques acquired at a different site. Furthermore, we applied transfer learning on 77 individuals' data and re-evaluated the model's performance using intra-class correlation coefficient (ICC). On the test set, DL outperformed the currently used minimum cost approach method to quantify total: ICC: 0.88 [CI: 0.85-0.91] vs. 0.63 [CI: 0.42-0.76], noncalcified: 0.84 [CI: 0.80-0.88] vs. 0.45 [CI: 0.26-0.59], calcified: 0.99 [CI: 0.98-0.99] vs. 0.96 [CI: 0.94-0.97], and low attenuation noncalcified: 0.25 [CI: 0.13-0.37] vs. -0.01 [CI: -0.13 to 0.11] plaque volumes. On the external dataset, substantial improvement was observed in DL model performance after transfer learning, total: 0.62 [CI: 0.01-0.84] vs. 0.94 [CI: 0.87-0.97], noncalcified: 0.54 [CI: -0.04 to 0.80] vs. 0.93 [CI: 0.86-0.96], calcified: 0.91 [CI:0.85-0.95] vs. 0.95 [CI: 0.91-0.97], and low attenuation noncalcified 0.48 [CI: 0.18-0.69] vs. 0.86 [CI: 0.76-0.92]. Our open-source DL algorithm achieved excellent agreement with expert CAD segmentations. However, transfer learning may be required to achieve accurate segmentations in the case of different plaque characteristics or machinery. • Deep learning 3D U-net model for coronary segmentation achieves comparable results with expert readers' volumetric plaque quantification. • Transfer learning may be needed to achieve similar results for other scanner and plaque characteristics. • The developed deep learning algorithm is open-source and may be implemented in any CT analysis software.
Identifiants
pubmed: 35524783
doi: 10.1007/s00330-022-08801-8
pii: 10.1007/s00330-022-08801-8
doi:
Types de publication
Journal Article
Langues
eng
Sous-ensembles de citation
IM
Pagination
7217-7226Subventions
Organisme : NIDA NIH HHS
ID : R01DA12777
Pays : United States
Organisme : NIDA NIH HHS
ID : R01DA15020
Pays : United States
Organisme : NIDA NIH HHS
ID : R01DA25524
Pays : United States
Organisme : NIDA NIH HHS
ID : R21DA048780
Pays : United States
Organisme : NIDA NIH HHS
ID : U01DA040325
Pays : United States
Organisme : NIDA NIH HHS
ID : R01DA12777
Pays : United States
Organisme : NIDA NIH HHS
ID : R01DA15020
Pays : United States
Organisme : NIDA NIH HHS
ID : R01DA25524
Pays : United States
Organisme : NIDA NIH HHS
ID : R21DA048780
Pays : United States
Organisme : NIDA NIH HHS
ID : U01DA040325
Pays : United States
Informations de copyright
© 2022. The Author(s), under exclusive licence to European Society of Radiology.
Références
Mathers CD, Stevens GA, Boerma T, White RA, Tobias MI (2015) Causes of international increases in older age life expectancy. Lancet 385:540–548
doi: 10.1016/S0140-6736(14)60569-9
GBD 2017 Causes of Death Collaborators (2018) Global, regional, and national age-sex-specific mortality for 282 causes of death in 195 countries and territories, 1980-2017: a systematic analysis for the Global Burden of Disease Study 2017
Knuuti J, Wijns W, Saraste A et al (2020) 2019 ESC Guidelines for the diagnosis and management of chronic coronary syndromes. Eur Heart J 41:407–477
doi: 10.1093/eurheartj/ehz425
Moss AJ, Williams MC, Newby DE, Nicol ED (2017) The updated NICE guidelines: cardiac CT as the first-line test for coronary artery disease. Curr Cardiovasc Imaging Rep 10:15
doi: 10.1007/s12410-017-9412-6
Leipsic J, Abbara S, Achenbach S et al (2014) SCCT guidelines for the interpretation and reporting of coronary CT angiography: a report of the Society of Cardiovascular Computed Tomography Guidelines Committee. J Cardiovasc Comput Tomogr 8:342–358
doi: 10.1016/j.jcct.2014.07.003
Kolossvary M, Szilveszter B, Merkely B, Maurovich-Horvat P (2017) Plaque imaging with CT-a comprehensive review on coronary CT angiography based risk assessment. Cardiovasc Diagn Ther 7:489–506
doi: 10.21037/cdt.2016.11.06
Williams MC, Earls JP, Hecht H (2021) Quantitative assessment of atherosclerotic plaque, recent progress and current limitations. J Cardiovasc Comput Tomogr. https://doi.org/10.1016/j.jcct.2021.07.001
Lee SE, Sung JM, Rizvi A et al (2018) Quantification of coronary atherosclerosis in the assessment of coronary artery disease. Circ Cardiovasc Imaging 11:e007562
doi: 10.1161/CIRCIMAGING.117.007562
Ferencik M, Mayrhofer T, Puchner SB et al (2015) Computed tomography-based high-risk coronary plaque score to predict acute coronary syndrome among patients with acute chest pain--results from the ROMICAT II trial. J Cardiovasc Comput Tomogr 9:538–545
doi: 10.1016/j.jcct.2015.07.003
Williams MC, Kwiecinski J, Doris M et al (2020) Low-attenuation noncalcified plaque on coronary computed tomography angiography predicts myocardial infarction: results from the multicenter SCOT-HEART trial (Scottish Computed Tomography of the HEART). Circulation 141:1452–1462
doi: 10.1161/CIRCULATIONAHA.119.044720
Kolossváry M, Karády J, Szilveszter B et al (2017) Radiomic features are superior to conventional quantitative computed tomographic metrics to identify coronary plaques with napkin-ring sign. Circ Cardiovasc Imaging 10:e006843
doi: 10.1161/CIRCIMAGING.117.006843
Kolossvary M, Gerstenblith G, Bluemke DA et al (2021) Contribution of risk factors to the development of coronary atherosclerosis as confirmed via coronary CT angiography: a longitudinal radiomics-based study. Radiology 299:97–106
doi: 10.1148/radiol.2021203179
Boogers MJ, Broersen A, van Velzen JE et al (2012) Automated quantification of coronary plaque with computed tomography: comparison with intravascular ultrasound using a dedicated registration algorithm for fusion-based quantification. Eur Heart J 33:1007–1016
doi: 10.1093/eurheartj/ehr465
Kolossvary M, Javorszky N, Karady J et al (2021) Effect of vessel wall segmentation on volumetric and radiomic parameters of coronary plaques with adverse characteristics. J Cardiovasc Comput Tomogr 15:137–145
doi: 10.1016/j.jcct.2020.08.001
Chen C, Qin C, Qiu H et al (2020) Deep learning for cardiac image segmentation: a review. Front Cardiovasc Med 7:25
doi: 10.3389/fcvm.2020.00025
Zeleznik R, Foldyna B, Eslami P et al (2021) Deep convolutional neural networks to predict cardiovascular risk from computed tomography. Nat Commun 12:715
doi: 10.1038/s41467-021-20966-2
van Velzen SGM, Lessmann N, Velthuis BK et al (2020) Deep learning for automatic calcium scoring in CT: validation using multiple cardiac CT and chest CT protocols. Radiology 295:66–79
doi: 10.1148/radiol.2020191621
Lai S, Fishman EK, Lai H et al (2008) Long-term cocaine use and antiretroviral therapy are associated with silent coronary artery disease in African Americans with HIV infection who have no cardiovascular symptoms. Clin Infect Dis 46:600–610
doi: 10.1086/526782
Lai H, Moore R, Celentano DD et al (2016) HIV infection itself may not be associated with subclinical coronary artery disease among African Americans without cardiovascular symptoms. J Am Heart Assoc 5:e002529
doi: 10.1161/JAHA.115.002529
Kolossvary M, Fishman EK, Gerstenblith G et al (2021) Cardiovascular risk factors and illicit drug use may have a more profound effect on coronary atherosclerosis progression in people living with HIV. Eur Radiol 31:2756–2767
doi: 10.1007/s00330-021-07755-7
Kolossvary M, Celentano D, Gerstenblith G et al (2021) HIV indirectly accelerates coronary artery disease by promoting the effects of risk factors: longitudinal observational study. Sci Rep 11:23110
doi: 10.1038/s41598-021-02556-w
Kolossvary M, Szilveszter B, Karady J, Drobni ZD, Merkely B, Maurovich-Horvat P (2019) Effect of image reconstruction algorithms on volumetric and radiomic parameters of coronary plaques. J Cardiovasc Comput Tomogr 13:325–330
doi: 10.1016/j.jcct.2018.11.004
Karady J, Panajotu A, Kolossvary M et al (2017) The effect of four-phasic versus three-phasic contrast media injection protocols on extravasation rate in coronary CT angiography: a randomized controlled trial. Eur Radiol 27:4538–4543
doi: 10.1007/s00330-017-4866-0
Newby DE, Adamson PD, Berry C et al (2018) Coronary CT angiography and 5-year risk of myocardial infarction. N Engl J Med 379:924–933
doi: 10.1056/NEJMoa1805971
Inoue K, Motoyama S, Sarai M et al (2010) Serial coronary CT angiography-verified changes in plaque characteristics as an end point: evaluation of effect of statin intervention. JACC Cardiovasc Imaging 3:691–698
doi: 10.1016/j.jcmg.2010.04.011
Oktay O, Schlemper J, Folgoc LL et al (2018) Attention U-Net: learning where to look for the pancreas. arXiv preprint arXiv:180403999
Pan LS, Li CW, Su SF, Tay SY, Tran QV, Chan WP (2021) Coronary artery segmentation under class imbalance using a U-Net based architecture on computed tomography angiography images. Sci Rep 11:14493
doi: 10.1038/s41598-021-93889-z
Siriapisith T, Kusakunniran W, Haddawy P (2021) A 3D deep learning approach to epicardial fat segmentation in non-contrast and post-contrast cardiac CT images. Peer J Comput Sci 7:e806
R Core Team (2019) R: a language and environment for statistical computing. R Foundation for Statistical Computing, 4.0.0
Bossuyt PM, Reitsma JB, Bruns DE et al (2015) STARD 2015: an updated list of essential items for reporting diagnostic accuracy studies. Radiology 277:826–832
doi: 10.1148/radiol.2015151516
Diaz-Zamudio M, Dey D, Schuhbaeck A et al (2015) Automated quantitative plaque burden from coronary CT angiography noninvasively predicts hemodynamic significance by using fractional flow reserve in intermediate coronary lesions. Radiology 276:408–415
doi: 10.1148/radiol.2015141648
Park HB, Heo R, B Óh et al (2015) Atherosclerotic plaque characteristics by CT angiography identify coronary lesions that cause ischemia: a direct comparison to fractional flow reserve. JACC Cardiovasc Imaging 8:1-10
Nakazato R, Shalev A, Doh JH et al (2013) Aggregate plaque volume by coronary computed tomography angiography is superior and incremental to luminal narrowing for diagnosis of ischemic lesions of intermediate stenosis severity. J Am Coll Cardiol 62:460–467
doi: 10.1016/j.jacc.2013.04.062
Hell MM, Motwani M, Otaki Y et al (2017) Quantitative global plaque characteristics from coronary computed tomography angiography for the prediction of future cardiac mortality during long-term follow-up. Eur Heart J Cardiovasc Imaging 18:1331–1339
doi: 10.1093/ehjci/jex183
Zreik M, van Hamersvelt RW, Wolterink JM, Leiner T, Viergever MA, Isgum I (2019) A recurrent CNN for automatic detection and classification of coronary artery plaque and stenosis in coronary CT angiography. IEEE Trans Med Imaging 38:1588–1598
doi: 10.1109/TMI.2018.2883807
Liu J, Jin C, Feng J, Du Y, Lu J, Zhou J (2019) A vessel-focused 3D convolutional network for automatic segmentation and classification of coronary artery plaques in cardiac CTA. Springer International Publishing, Cham, pp 131–141