Assessing the efficacy of machine learning algorithms for syncope classification: A systematic review

Abstract

Syncope is a transient loss of consciousness with rapid onset. The aims of the study were to systematically evaluate available machine learning (ML) algorithm for supporting syncope diagnosis to determine their performance compared to existing point scoring protocols. We systematically searched IEEE Xplore, Web of Science, and Elsevier for English articles (Jan 2011 - Sep 2021) on individuals aged five and above, employing ML algorithms in syncope detection with Head-up titl table test (HUTT)-monitored hemodynamic parameters and reported metrics. Extracted data encompassed subject count, age range, syncope protocols, ML type, hemodynamic parameters, and performance metrics. Of the 6301 studies initially identified, 10 studies, involving 1205 participants aged 5 to 82 years, met the inclusion criteria, and formed the basis for it. Selected studies must use ML algorithms in syncope detection with hemodynamic parameters recorded throughout HUTT. The overall ML algorithm performance achieved a sensitivity of 88.8% (95% CI: 79.4-96.1%), specificity of 81.5% (95% CI: 69.8-92.8%) and accuracy of 85.8% (95% CI: 78.6-92.8%). Machine learning improves syncope diagnosis compared to traditional scoring, requiring fewer parameters. Future enhancements with larger databases are anticipated. Integrating ML can curb needless admissions, refine diagnostics, and enhance the quality of life for syncope patients.

Keywords: Classification; Machine learning; Methodology for conducting a systematic literature review; Syncope diagnosis; Systematic review.

Graphical abstract

Fig. 1

Flowchart of the Study Selection. A flowchart that employing the PRISMA diagram to visualize the systematic search process. A total of 7815 articles were identified from three major databases (5648 from Web of Science, 1141 from Elsevier, and 1026 from IEEE Explorer). Subsequently, ten articles were selected, with an additional three from different sources.

Fig. 2

Proportion of the different answers in the high- and low-priority items. The comparison of checklist scores between low-priority and high-priority parameters, employing a 50-point scale and a four-tier evaluation system (NA - not applicable, OK - adequately addresses, mR - minor revisions needed, and MR - major revisions needed). The figure showcases the distribution of scores across six checklist parameters, emphasizing the impact of double weighting on high-priority items and classifying the overall study quality as low, medium, or high.

Fig. 3

Forest Plot of Performance Metrics. It is showcasing (a) Sensitivity, (b) Specificity, and (c) Accuracy for machine learning algorithms in Syncope classification across selected studies. The comprehensive forest plot provides a visual overview of performance metrics, including estimates and corresponding 95% confidence intervals. This graphical representation enables a quick assessment of variability and precision across multiple studies.

Fig. 4

Forest Plot of Different Performance Metrics Estimate from the Studies Using Scoring Method in Syncope Classification. The forest plots, accompanied by 95% confidence intervals, visually depict the performance metrics—Sensitivity, Specificity, and Accuracy—of Machine Learning Algorithms employing scoring methods. This figure provides a comprehensive view of algorithmic efficacy in syncope classification.