Researchers at Penn State College of Medicine have developed a new artificial intelligence (AI) model that can predict the progression of autoimmune diseases more accurately. Autoimmune diseases occur when the immune system mistakenly attacks the body’s own cells. These diseases often have a preclinical stage, characterized by mild symptoms or certain antibodies in the blood, before progressing to the full disease stage.

The AI model analyzes data from electronic health records and extensive genetic studies to create a risk prediction score. This method is significantly more accurate than existing models, improving prediction accuracy by up to 1,000%. The findings were published in Nature Communications.

“By targeting people with family history or early symptoms, we can use machine learning to identify those with the highest risk and suitable treatments,” said Dajiang Liu, a professor at Penn State and co-lead author of the study.

Around 8% of Americans live with autoimmune diseases, most of whom are women. Early detection and intervention are crucial because the damage caused by these diseases can be irreversible. For example, antibodies in the blood can be detected in rheumatoid arthritis patients five years before symptoms appear.

The research team developed a Genetic Progression Score (GPS) method to predict disease progression from preclinical stages. GPS leverages transfer learning, a technique that allows a model trained on one task to be adapted for another related task. This improves accuracy, even with smaller data samples.

GPS uses data from extensive genetic studies and electronic health records to identify high-risk individuals and characterize disease progression stages. This integrated approach improves prediction accuracy.

The AI model was tested using real-world data and validated with data from the National Institutes of Health’s All of Us biobank. GPS outperformed 20 other models, providing more accurate predictions.

Accurate prediction of disease progression enables early interventions, targeted monitoring, and personalized treatment, leading to better patient outcomes. This AI model could also improve clinical trials by identifying those who are most likely to benefit from new therapies.

The research team includes genetics, public health, dermatology, and rheumatology experts, reflecting Penn State’s comprehensive research program in autoimmune diseases.