At Aix Marseille University, a team of researchers is asking an urgent question: If we teach artificial intelligence to understand heart disease using data that mostly reflects how men's hearts work, what happens to the women whose hearts are fundamentally different? The answer could determine whether the next generation of personalized heart medicine actually works for everyone.
Digital twins—computer models that simulate a patient's unique biology—are reshaping cardiology. By combining medical imaging, clinical records, and biological data, researchers can now create a virtual version of a patient's heart and test treatment strategies before they're used in real life. The European Commission is backing this vision through the European Virtual Human Twin Initiative, while projects like SimCardioTest are already building patient-specific cardiovascular models to improve diagnosis and treatment planning. The promise is seductive: precision medicine tailored to each person's biology, rather than one-size-fits-all treatments based on population averages.
But this promise hinges on a critical blind spot in medical science. For decades, biomedical research has treated male biology as the default. A widely cited analysis in Nature found that male animals historically outnumbered females by roughly five to one in preclinical studies. This imbalance compounds across the entire medical pipeline, leaving digital twins potentially trained on data that doesn't fully represent how disease works in women's bodies.
In cardiovascular medicine, the stakes are particularly high. Heart disease kills nearly 18 million people annually worldwide, making it the leading cause of death globally. Yet it does not affect women and men equally. Myocarditis—inflammation of the heart muscle that can follow viral infections—strikes two to four times more frequently in men than women, especially among young adults, affecting around 1.8 million people globally each year. Research suggests these differences flow from variations in immune responses, hormonal influences, and cardiac tissue biology itself. When a digital twin model is trained primarily on male biology, it may miss the very mechanisms that explain disease progression in women.
The researchers working on the MYOCAR3 project, funded by Civis Alliance, are beginning to map these differences. Their work is revealing how variations in immune responses between women and men could fundamentally alter how inflammatory heart disease develops and manifests. If digital twins cannot capture this diversity, they risk perpetuating a centuries-old problem: medicine built on the assumption that the male body is universal.
As these technologies move from research labs toward clinical practice, ensuring they reflect the full spectrum of human biology is no longer optional—it's essential. The engineers, clinicians, and data scientists now building Europe's digital heart models have a choice: they can replicate the male-centered research traditions of the past, or they can intentionally construct datasets and algorithms that represent everyone. The future of truly personalized medicine depends on which path they choose.