What is the clinical evidence from this study?

Study design: Systematic Review. Population: Sudden cardiac death. Intervention: Machine learning models vs. Traditional regression models. Primary outcome: Sudden cardiac death risk prediction (AUROC).

August 25, 2022Open Access

Machine learning in sudden cardiac death risk prediction: a systematic review

Key Result

Machine learning models for sudden cardiac death risk prediction yielded AUROCs ranging from 0.71 to 0.96 and improved performance over traditional regression models in 5 out of 6 studies.

Study Design

Type

Systematic Review

Structured PICO

Do machine learning models improve sudden cardiac death risk prediction compared to traditional regression models?

Population

11 studies with heterogeneous populations evaluating sudden cardiac death risk prediction, with study sizes ranging from 122 to 124,097 participants.

Intervention

Machine learning models for sudden cardiac death risk prediction using demographic, clinical, electrocardiogram, electrophysiological, imaging, and genetic data (4 to 72 variables per model).

Comparator

Traditional regression models (evaluated in 6 studies).

Outcome

Predictive performance for sudden cardiac death risk (most commonly measured by area under the receiver operator characteristic [AUROC]).

Machine learning shows potential incremental utility over traditional models for predicting sudden cardiac death, but current evidence is limited by poor reporting standards and high risk of bias.

Limitations

Heterogeneous populations, methods, and outcome measures preventing meta-analysis
No studies adhered to a reporting standard
Five of the papers were at high risk of bias

Abstract

AIMS: Most patients who receive implantable cardioverter defibrillators (ICDs) for primary prevention do not receive therapy during the lifespan of the ICD, whilst up to 50% of sudden cardiac death (SCD) occur in individuals who are considered low risk by conventional criteria. Machine learning offers a novel approach to risk stratification for ICD assignment. METHODS AND RESULTS: Systematic search was performed in MEDLINE, Embase, Emcare, CINAHL, Cochrane Library, OpenGrey, MedrXiv, arXiv, Scopus, and Web of Science. Studies modelling SCD risk prediction within days to years using machine learning were eligible for inclusion. Transparency and quality of reporting (TRIPOD) and risk of bias (PROBAST) were assessed. A total of 4356 studies were screened with 11 meeting the inclusion criteria with heterogeneous populations, methods, and outcome measures preventing meta-analysis. The study size ranged from 122 to 124 097 participants. Input data sources included demographic, clinical, electrocardiogram, electrophysiological, imaging, and genetic data ranging from 4 to 72 variables per model. The most common outcome metric reported was the area under the receiver operator characteristic (n = 7) ranging between 0.71 and 0.96. In six studies comparing machine learning models and regression, machine learning improved performance in five. No studies adhered to a reporting standard. Five of the papers were at high risk of bias. CONCLUSION: Machine learning for SCD prediction has been under-applied and incorrectly implemented but is ripe for future investigation. It may have some incremental utility in predicting SCD over traditional models. The development of reporting standards for machine learning is required to improve the quality of evidence reporting in the field.