Skip to main page content
U.S. flag

An official website of the United States government

Dot gov

The .gov means it’s official.
Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

Https

The site is secure.
The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
Editorial
. 2023 Jan 11;2(1):100166.
doi: 10.1016/j.jacadv.2022.100166. eCollection 2023 Jan.

Machine Learning Algorithms: Selection of Appropriate Validation Populations for Cardiology Research-Be Careful!

William E Sanders Jr  1 , Rola Khedraki  2 , Mark Rabbat  3 , Thomas R McMinn  4 , Tim Burton  5 , Ali Khosousi  5 , Farhad Fathieh  5 , Horace R Gillins  6 , Shyam Ramchandani  5 , Ian P Shadforth  6
Affiliations
Editorial

Machine Learning Algorithms: Selection of Appropriate Validation Populations for Cardiology Research-Be Careful!

William E Sanders Jr et al. JACC Adv. .
No abstract available

Keywords: artificial intelligence; coronary artery disease; digital health; machine learning; validation populations.

PubMed Disclaimer

Conflict of interest statement

CorVista Health funded the collection of subject data. Dr Sanders, Mr Burton, Dr Khosousi, Dr Fathieh, Dr Ramchandani, and Mr Shadforth are employees of CorVista Health. Dr Rabbat is a member of the Medical Advisory Board for CorVista Health. All other authors have reported that they have no relationships relevant to the contents of this paper to disclose.

Figures

Figure 1
Figure 1
Variance in Area Under the Receiver-Operating Characteristic Curve Using Different Validation Populations Three ROC curves generated by applying the same coronary artery disease (CAD) classification algorithm to the same positive population, with 3 distinct negative populations: Group 1 (green curve) = subjects shown to have no significant CAD by invasive coronary angiography (ICA) (AUC = 0.5944); group 2 (blue curve) = subjects shown to have no significant CAD by computed tomography angiography (CTA) (AUC = 0.7641); group 3 (orange curve) = subjects assumed to have no significant CAD as they have no known CAD, no symptoms of CAD, and no risk factors for cardiovascular disease (AUC = 0.9253).
See this image and copyright information in PMC

References

    1. Brenner H., Gefeller O. Variation of sensitivity, specificity, likelihood ratios and predictive values with disease prevalence. Stat Med. 1997;16:981–991. - PubMed
    1. Leeflang M.M.G., Rutjes A.W.S., Reitsma J.B., Hooft L., Bossuyt P.M.M. Variation of a test’s sensitivity and specificity with disease prevalence. CMAJ. 2013;185:E537–E544. - PMC - PubMed
    1. Stuckey T., Meine F., McMinn T., et al. Development and validation of a machine learned algorithm to IDENTIFY functionally significant coronary artery disease. Front Cardiovasc Med. 2022;9:956147. doi: 10.3389/fcvm.2022.956147. - DOI - PMC - PubMed
    1. Douglas P.S., Hoffmann U., Manesh R.P., et al. Outcomes of anatomical versus function testing for coronary artery disease. N Engl J Med. 2015;372:1291–1300. doi: 10.1056/NEJMoa1415516. - DOI - PMC - PubMed
    1. Park S.J., Kang S.J., Ahn J.M., et al. Visual-functional mismatch between coronary angiography and fractional flow reserve. J Am Coll Cardiol Intv. 2012;5:1029–1036. doi: 10.1016/j.jcin.2012.07.007. - DOI - PubMed

Publication types

LinkOut - more resources