Clinical, Laboratory and Multi-Omics Data to Leverage Machine Learning for Personalized Diagnostics

Early and accurate disease diagnosis is crucial for stopping disease progression and identifying appropriate treatments, but precise diagnosis of autoimmune disease is notoriously challenging due to the non-specific symptoms: it can take 4 to 5 years and multiple visits to various physicians. However, an early and precise disease diagnosis is crucial for patients and helps doctors define personalized therapeutic strategies. Clinical decision support systems (CDSS) are software tools that support clinicians in assigning patients to a disease category. Although individual hospitals use their own CDSS, their use is often limited to certain data types and preset rules reflecting medical questionnaires. To address this unmet need, in collaboration with clinicians at the university hospitals in Strasburg, Mainz and Freiburg, we developed the Personalis platform. Personalis is a CDSS aimed at demultiplexing complex autoimmune diseases by applying a range of machine learning models to different types of integrated medical datasets such as clinical, laboratory and multi-omics data. Machine learning models – such as neural networks, support vector machines and random forests – predict autoimmune diseases with more than 95% accuracy, and that accuracy increases when laboratory results such as cytokine concentration, as well as genetics, immunomics and metabolomics data, are added.

Contact

Prof. Dr. Enkelejda Miho

Expert in Medical Informatics, Immunoinformatics, Machine Learning, Drug Discovery, Digital Health, Diagnostics