Predicting Patient Genotypes Using Electronic Health Records and Multimodal Data Through AI-Based Models

What this Trial Is About

Researchers are exploring how electronic health records (EHR) and various types of clinical data can be used to predict a patient's genotype without direct genetic testing. This observational study evaluates whether an AI-based model can accurately infer genetic information by analyzing health data such as lab results, imaging studies, and medical history. The goal is to develop a non-invasive and cost-effective method for genotype prediction that could support personalized medicine and early disease diagnosis. The study involves analyzing historical health data from participants, including EHR, clinical lab results, and multimodal imaging like X-rays, MRIs, and CT scans. An AI predictive model is trained to combine and interpret these data sources to estimate genetic profiles. Participants do not receive any active treatment or genetic testing during the study; their existing health data are used solely for AI analysis. The accuracy of the AI predictions will be compared to actual genetic testing results when available. Participants contribute by allowing access to their medical records and imaging data for AI evaluation. The study measures the AI model's performance using metrics such as the Area Under the Curve (AUC) and F1 Score over one year, along with sensitivity and specificity. There are no study-related procedures beyond data collection, and no direct interventions or genetic tests are performed. The study aims to advance AI applications in genotype prediction through careful analysis of existing clinical data.

AI-Driven Genotype Prediction Using EHR and Multimodal Data