Predictive Modeling

Fillmore has developed and validated sophisticated predictive models to improve risk stratification for cancer-associated complications and disease progression. His work employs deep learning, transformer models, and longitudinal electronic health record data to predict venous thromboembolism, cardiovascular risk, and mortality across various malignancies. By integrating multimodal data, including circulating tumor DNA and plasma protein panels, he has created tools that outperform traditional static scores to enable more precise, personalized preventative interventions.