Using Artificial Intelligence to Better Predict and Develop Biomarkers

Beyond adjusting treatments based on symptoms and physical examination findings, clinicians must rely on adjunctive testing to support their judgment; this includes imaging and measurement of circulating biomarkers. Most notably, among biomarkers measured, natriuretic peptides are considered the standard of care for treating patients with HF. ^, However, HF is a complicated condition, which cannot be summarized by one biomarker alone. ^, Although natriuretic peptides indicate cardiomyocyte stretch, they have limited ability to capture other changes associated with HF remodeling included inflammation, oxidative stress, and fibrosis. ^,

For this reason, multiple markers are needed to recognize the pathophysiology of HF and tailor care to optimize patient outcomes.

The introduction of omics, including genomics, transcriptomics, proteomics, and metabolomics, has enhanced biomarker discovery.

Omics allows for analysis of biological pathways involved in aspects of the central dogma and biological modifications of products. Using high-throughput and untargeted omics technologies provides better phenotyping and unveils previously unknown HF pathophysiology. Furthermore, integration of findings of multiple omics studies gives a comprehensive understanding of pathophysiology from the level of genes to final metabolites to recognize where disease processes arise. Discovery of markers enables researchers and clinicians to recognize potential risks for developing HF, serve as prognostic indicators, or ascertain druggable targets for future therapeutic interventions.