Researchers at the Dana-Farber Cancer Institute and Massachusetts General Hospital have developed an artificial intelligence (AI) tool capable of predicting the risk of more than 300 diseases by analyzing electronic health records (EHRs). The tool, named Aladynoulli, assesses routine medical data along with genetic test results when available, offering predictions on conditions including various cancers and cardiovascular diseases.
Described in a study published Wednesday in the journal Nature, the AI model demonstrated performance superior to commonly used online risk calculators for cardiovascular disease within a 10-year period and for breast cancer within one year. According to Giovanni Parmigiani, a senior researcher at Dana-Farber and a co-author of the paper, the AI effectively navigates the complex and voluminous data from medical records that would be difficult for clinicians to process manually.
“Medical records contain a wealth of longitudinal information across different disease areas, which is challenging to interpret without computational assistance,” Parmigiani said. The tool’s ability to integrate diverse types of patient data represents an advance in personalized risk assessment.
While AI applications in healthcare have previously focused on identifying undiagnosed rare diseases or early signs of Alzheimer’s, this technology extends the predictive scope to a broad array of health conditions. The researchers aim to integrate Aladynoulli into the EHR systems at Dana-Farber and Massachusetts General Hospital in the near future, though no specific timeline has been announced for implementation or wider distribution.
Parmigiani noted that despite the tool being “ready for prime time,” the adoption of AI in clinical settings involves navigating complex regulatory, technical, and practical challenges before it can be broadly deployed. The development marks a step forward in leveraging machine learning for disease risk prediction, potentially aiding earlier intervention and personalized patient management.