This systematic review adopts a cross-domain analytical perspective to evaluate the diagnostic performance and clinical applicability of data-driven and artificial intelligence-based methods across multiple biomedical domains, with particular emphasis on ophthalmic imaging. Rather than focusing on a single disease category, the review compares how methodological characteristics influence diagnostic accuracy and translational utility across heterogeneous clinical contexts. A literature search was conducted in accordance with Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines, including studies published since 2015. Study selection was guided by predefined eligibility criteria requiring original research reporting primary data, with a clearly defined study design, population or dataset, analytical methods, and measurable diagnostic outcomes. Studies lacking sufficient methodological detail or quantitative outcome reporting, as well as non-primary research studies, were not considered. A total of 11 studies were included, spanning ophthalmic imaging, biomarker-based diagnosis, behavioural and cognitive assessment, vascular disease, and experimental applications. High diagnostic performance was consistently observed in imaging-based screening domains, with sensitivity and specificity frequently exceeding 90%, while moderate but clinically meaningful performance was reported in biomarker-driven and cognitive domains. These findings indicate that analytical approaches provide the greatest clinical utility when aligned with specific healthcare objectives and implementation contexts. Integration of quantitative performance with translational relevance offers a structured framework to support clinical adoption and future methodological development.
Shah et al. (Thu,) studied this question.