How AI Revolutionizes Cardiovascular Risk Detection and Emergency Care at UTMB

At the University of Texas Medical Branch (UTMB), artificial intelligence is transforming cardiovascular risk detection and emergency care by automatically analyzing every CT scan performed, regardless of the initial reason for the scan. This innovative approach leverages AI to identify incidental coronary artery calcification (iCAC), a key predictor of heart disease, enabling early intervention for patients who might otherwise go unnoticed.

UTMB’s AI system uses convolutional neural networks (CNNs) to calculate an Agatston score from CT images, quantifying arterial plaque buildup. Patients with scores above 100 are categorized into risk tiers using additional clinical data extracted from electronic health records (EHRs), including free-text notes processed by advanced language models like GPT-4o. Those flagged as high risk receive automated notifications and follow-up communications, ensuring timely clinical attention.

Beyond cardiovascular risk, UTMB employs AI algorithms to rapidly detect strokes and pulmonary embolisms by analyzing CT scans for critical indicators such as blood flow obstructions. These algorithms instantly alert care teams with annotated images, accelerating emergency interventions where every minute counts. This automation has already resulted in measurable time savings and improved patient outcomes.

UTMB rigorously validates AI models pre- and post-deployment by comparing AI-generated scores with radiologist assessments to monitor accuracy, sensitivity, and bias. The institution also employs peer learning techniques to mitigate anchoring bias, ensuring that AI assistance enhances rather than narrows diagnostic perspectives. Continuous evaluation helps maintain trust and reliability in AI-driven clinical decisions.

Additional AI applications at UTMB include automated inpatient admission assessments and identification of care gaps by analyzing comprehensive patient data from EHRs. These tools act as co-pilots for medical staff, reducing the intellectual bottleneck caused by overwhelming data volume and enabling proactive, data-driven healthcare management.

UTMB’s experience highlights the transformative potential of AI in healthcare: automating routine yet critical analyses, enhancing early detection of disease, and streamlining clinical workflows. By integrating AI into everyday medical imaging and patient data review, healthcare providers can unlock valuable insights hidden in vast datasets, ultimately improving patient outcomes and operational efficiency.