Plasma proteomic profiles predict individual future health risk

Developing a single-domain assay to identify individuals at high risk of future events is a priority for multi-disease and mortality prevention. This study developed a disease/mortality-specific proteomic risk score (ProRS) based on 1461 Olink plasma proteins measured in 52,006 UK Biobank participants using a neural network. ProRS effectively stratified risk for 45 common conditions and achieved high accuracy in predicting 10 endpoints (e.g., cancer, dementia, death). ProRS often outperformed established clinical indicators. While incorporating clinical predictors enhanced predictive power, the improvement over ProRS alone was limited. Proteins like GDF15 showed strong discriminative value across various diseases. The study demonstrated potential clinical utility, though independent external validation is needed before clinical application.

Jia You, Yu Guo, Yi Zhang, Ju-Jiao Kang, Lin-Bo Wang, Jian-Feng Feng, Wei Cheng, Jin-Tai Yu