Potential of artificial intelligence in prediction of coronary arterial lesions
AbstractAim of the study was to analyze the possibility of using neural network analysis to predict the coronary bed lesion severity, as well as to compare the accuracy of the trained neural network model on the input structured data and ECG records with a cardiologist.
Material and methods. The study included 120 patients, who underwent elective or emergency coronary catherization.
Key inclusion criteria: male or female patients (older 18 years), performed coronary catherization, recorded electrocardiography one day before/or less before the intervention.
Key exclusion criteria: ECG identification of arrhythmias as the atrial fibrillation, AV nodal re-entrant tachycardia, ventricular tachycardia while recording, previous stenting and/or coronary artery bypass grafting.
The method of neural network analysis was used to predict the coronary bed lesion. The machine learning included clinical, laboratory, instrumental (ECG images) parameters (23 parameters in total). The neural network was used to solve the classification problems, receiving input as the structured data and images and providing output as a multifactorial classification of the main coronary arteries. The ratio for training and testing was 100/20. Supervised learning was used on the available data, where the outcomes were known (coronary angiography data), and neural network parameters were adjusted to minimize the error by backpropagation. For the experiment, based on the test sample, were prepared 20 cases to which the ECG images were attached. Five cardiologists daily supervising patients with acute coronary syndrome (ACS), had to individually analyze the coronary bed pathology for each major coronary artery and predict the necessity of revascularization.
Results. On a test sample including 20 patients, the neural network result was: AUC score - 0.74, where the accuracy - 80%, precision - 63%, recall - 55% and the fl score (harmonic mean between accuracy and recall) - 59%.
The average response rates of cardiologists were: accuracy - 76%, precision - 48%, recall - 55%, AUC score - 0.68 and f1 score - 49%. The best parameters among cardiologists were: accuracy - 76%, precision - 48%, recall - 67%, AUC score - 0.72, f1 score - 56%.
Conclusion. Neural network analysis of the available clinical, laboratory and instrumentation data allow to configure the network parameters for further accurate prediction of coronary artery disease. The results obtained in the form of an AUC score allow to consider this method to be effective in the coronary bed pathology diagnosis. On the test sample, the neural network is more effective than cardiologists on average. One in five specialists could roughly compare with the accuracy of trained neural network model.
Keywords:coronary arteries; neural networks; deep learning
Funding. The study had no sponsor support.
Conflict of interest.The authors declare no conflict of interest.
For citation: Abdualimov T.P., Obrezan A.G. Potential of artificial intelligence in prediction of coronary arterial lesions. Kardiologiya: novosti, mneniya, obuchenie [Cardiology: News, Opinions, Training]. 2022; 10 (1): 34-9. DOI: https://doi.org/10.33029/2309-1908-2022-10-1-34-39 (in Russian)
References
1. Kuo F.C., Mar B.G., Lindsley R.C., Lindeman N.I. The relative utilities of genome-wide, gene panel, and individual gene sequencing in clinical practice. Blood. 2017; 130: 433–9.
2. Muse E.D., Barrett P.M., Steinhubl S.R., Topol E.J. Towards a smart medical home. Lancet. 2017; 389: 358.
3. Steinhubl S.R., Muse E.D., Topol E.J. The emerging field of mobile health. Sci Transl Med. 2015; 7: 283rv3.
4. Shameer K., Badgeley M.A., Miotto R., Glicksberg B.S., Morgan J.W., Dudley J.T. Translational bioinformatics in the era of real-time biomedical, health care and wellness data streams. Brief Bioinform 2017; 18: 105–24.
5. Konstam M.A., Hill J.A., Kovacs R.J., et al. The academic medical system: reinvention to survive the revolution in health care. J Am Coll Cardiol. 2017; 69: 1305–12.
6. Steinhubl S.R., Topol E.J. Moving from digitalization to digitization in cardiovascular care: why is it important, and what could it mean for patients and providers? J Am Coll Cardiol. 2015; 66: 1489–96.
7. Boeldt D.L., Wineinger N.E., Waalen J., et al. How consumers and physicians view new medical technology: comparative survey. J Med Internet Res. 2015; 17: e215.
8. Wasserlauf J., You C., Patel R., et al. Smartwatch performance for the detection and quantification of atrial fibrillation. Circ Arrhythm Electrophysiol. 2019; 12: e006834.