Combining biomarkers to improve diagnostic accuracy using the overlap coefficient

Coolen-Maturi, Tahani

Abstract

Measuring the accuracy of diagnostic tests is crucial in many application areas, including medicine, machine learning, and credit scoring. In practice, multiple diagnostic tests or biomarkers are combined to improve diagnostic accuracy. The area under the receiver operating characteristic curve (AUC) is a common measure of diagnostic test performance and can be used as an objective function to maximise when combining multiple biomarkers. Another useful measure is the overlap coefficient, which quantifies the similarity between two independent distributions by their overlapping area. The smaller the overlapping area, the better the biomarker is at discrimination. The aim of this paper is to combine biomarkers to improve diagnostic accuracy by minimising the overlap coefficient. We approach this parametrically and non-parametrically using Kernel-based methods. We also present a probabilistic interpretation of the overlap coefficient, which gives more insight into this measure. The proposed methods are evaluated through a simulation study and illustrated via examples.

Citation

Coolen-Maturi, T. (online). Combining biomarkers to improve diagnostic accuracy using the overlap coefficient. Communications in Statistics - Theory and Methods, https://doi.org/10.1080/03610926.2025.2460095

• Journal Article Type: Article
• Acceptance Date: Jan 23, 2025
• Online Publication Date: Feb 27, 2025
• Deposit Date: Feb 19, 2025
• Publicly Available Date: Mar 19, 2025
• Journal: Communications in Statistics – Theory and Methods
• Print ISSN: 0361-0926
• Electronic ISSN: 1532-415X
• Publisher: Taylor and Francis Group
• Peer Reviewed: Peer Reviewed
• DOI: https://doi.org/10.1080/03610926.2025.2460095
• Keywords: Diagnostic accuracy; combining biomarkers; overlap coefficient; ROC; AUC
• Public URL: https://durham-repository.worktribe.com/output/3490181