Research Square · Nov 18, 2026
Artificial Intelligence for Treatment-Response Prediction in GLP-1 Therapies: A Systematic Review and Translational Readiness Framework (AI-TRF)
By Al-Quraishi T, Albahri OS, Abawajy J et al.
Glucagon-like receptor agonists and incretin-based therapies improve glycaemic control and weight-related outcomes, but individual treatment responses vary. This review examines treatment-response prediction and the clinical evidence relevant to personalised treatment selection. The current evidence table contains 44 records identified through searches in PubMed, Scopus, and Web of Science. These records comprise a mixture of prediction and predictor analyses, clinical efficacy studies, and other evidence of response to treatment; they should not be interpreted as 44 independently validated AI prediction systems. The synthesis describes therapeutic targets, data sources, and analytical approaches. However, incomplete study-level extraction and unresolved classification issues prevent reliable estimates of prediction-model validation practices, explainability uptake, or comparative algorithm performance. We also propose the Artificial Intelligence Translational Readiness Framework (AI-TRF), comprising external validation, explainability, temporal modelling, prospective evaluation, and clinical implementation. AI-TRF has not been applied to individual studies or empirically validated. Reproducible searches, verified study-level extraction, and independent evaluation remain necessary to establish the clinical readiness of treatment-response prediction systems in this field.
doi.org/10.21203/rs.3.rs-11763328/v1Open accessInterventions