Dr. Matthew Pease: “Artificial Intelligence and the Neurotrauma ICU”

At SDSC’s October Data Science Roundtable, Dr. Matthew Pease – a neurosurgeon and data scientist at Indiana University School of Medicine – discussed how artificial intelligence (AI) can transform care in the neurotrauma intensive care unit (ICU). Dr. Pease, who specializes in brain cancer and trauma, focuses on leveraging AI to predict outcomes, identify disease subtypes (phenotypes), and improve clinical decision making for traumatic brain injury (TBI) patients.

He emphasized that neurotrauma is a big data problem, with the electronic medical record housing vast multimodal data – from imaging and labs to electroencephalograms (EEG) and intracranial pressure monitoring. Current classification systems, such as the Glasgow Coma Scale, oversimplify TBI into “mild, moderate, and severe,” masking the true heterogeneity of the disease. Dr. Pease’s work uses machine learning and deep neural networks to analyze CT scans and patient data, outperforming traditional predictive models and even expert clinicians in forecasting outcomes.

A major theme of his talk was the gap between AI innovation and clinical application. He stressed that the ultimate challenge is not marginally improving models but bringing them safely and ethically into patient care. Dr. Pease also raised critical questions about how outcomes should be measured, noting that the complex, somber, and, in some cases, even premature decision-making process for the withdrawal of life-sustaining therapies often determines mortality, rather than the injury itself.

His research aims to move from reactive to predictive and personalized neurotrauma care, ensuring that AI helps clinicians make better, data-driven decisions to improve recovery and survival.

‍