Autoimmune diseases, where the immune system mistakenly attacks the body’s own cells, often have a preclinical stage marked by mild symptoms or specific antibodies in the blood. This stage presents a crucial opportunity for early intervention. A team of researchers from Penn State College of Medicine has developed an innovative method that uses artificial intelligence (AI) to predict the progression of autoimmune diseases in patients with early symptoms.

Leveraging electronic health records and genetic studies, the team designed a tool called the Genetic Progression Score (GPS). This methodology, based on machine learning and transfer learning techniques, has proven to be between 25% and 1,000% more accurate than traditional models in predicting who will develop the full disease.

The GPS integrates data from genome-wide association studies and clinical biobanks, providing valuable insights into genetic and clinical factors related to disease progression. According to Dajiang Liu, director of AI and biomedical informatics at the university, this technology not only improves early detection but also guides personalized therapeutic decisions and optimizes clinical trial design.

The model has already been successfully tested on diseases like rheumatoid arthritis and lupus using data from biobanks such as Vanderbilt University and the All of Us program. Beyond its impact on autoimmune diseases, researchers believe this approach could be applied to other conditions, highlighting the transformative role of AI in reducing health disparities and advancing medical research.

Article written by Christine Yu

07/01/2025

Source:

Penssilvania university

https://www.psu.edu/news/research/story/predicting-progression-autoimmune-disease-ai