New Grant to Support Stroke Research

The National Institutes of Health (NIH) has awarded $3.7 million to researchers at Columbia University Irving Medical Center (CUIMC) and Weill Cornell Medicine to confront one of stroke medicine’s most pressing unsolved problems: the one in five ischemic strokes that occur without any identifiable cause, and the dangerous arterial plaques that may be quietly behind them.

The five-year project, “Carotid Atherosclerotic Risk Assessment Through Imaging and Prediction” (CAROTID-MAP), will be led at CUIMC by Ajay Gupta, MD, MS, James Picker Professor of Radiology and chair of the Department of Radiology. Hediyeh Baradaran, MD, MS, associate professor of radiology and chief of the Division of Neuroradiology will serve as co-principal investigator.

Rigorous Research Towards Stroke Prevention

Unexplained or “cryptogenic” strokes leave patients and clinicians without a clear prevention strategy. Mounting evidence points to a culprit that standard imaging tests routinely miss: high-risk plaques within the carotid arteries that do not significantly narrow the vessel but can still rupture, shed clots, and cause stroke. The new study is designed to identify these plaques, characterize their risk, and predict which patients will suffer a future stroke—enabling prevention before catastrophe strikes.

“Many strokes still occur without a clear explanation, which limits our ability to prevent future events,” said Baradaran. “By moving beyond traditional measures like stenosis and harnessing advanced CT-based plaque characterization, we aim to reveal previously unrecognized sources of stroke risk and equip treating clinicians with more precise, data-driven insights to guide care.”

The study’s foundation is CT angiography (CTA)—a fast, widely available imaging tool that is already a routine part of the stroke workup at most major hospitals.

Rather than requiring a new or specialized scan, CAROTID-MAP aims to extract risk information from images already being obtained. The approach mirrors a transformation that reshaped preventive cardiology: in the coronary arteries, CT imaging has long enabled cardiologists to identify high-risk plaque features and predict heart attack risk before a clinical event, driving earlier and more targeted prevention. Despite the parallel anatomy and biology, a comparable CT-based risk framework has never been established for the carotid arteries.

“For too long, a significant fraction of our patients have left the hospital after a stroke without a clear explanation and without a targeted strategy to prevent the next one,” said Gupta. “CAROTID-MAP gives us the opportunity to change that—doing for stroke what has already been done so powerfully in the coronary world.”

Advanced Imaging Technology

The timing for this study is particularly good: a new generation of CT technology, photon counting CT, promises substantially higher spatial and contrast resolution and improved plaque tissue characterization, meaning the imaging foundation being laid by this study will only grow more powerful as the technology becomes more widely deployed.

Critically, CAROTID-MAP will pair that imaging foundation with artificial intelligence. Machine learning algorithms—including deep neural networks trained directly on CT pixel data—will be developed to automatically characterize high-risk plaque features and stratify stroke risk at a scale and speed that manual expert review alone cannot achieve. The algorithms will ultimately be tested in a prospective cohort of patients followed for recurrent stroke, with the goal of producing a validated, clinically deployable tool ready for multicenter trials.

Cross-campus Collaboration

The project is led by a multidisciplinary team of principal investigators spanning both medical school campuses of the NewYork-Presbyterian health system, with Gupta serving as contact PI. Baradaran will lead the imaging interpretation core. At Weill Cornell Medicine, Mert Sabuncu, PhD, professor of electrical engineering and radiology, will direct the AI and machine learning component, and Hooman Kamel, MD, vice-chair of research for neurology and a nationally recognized expert in stroke prevention, will lead the prospective clinical outcomes study.

“We are particularly excited to be conducting this study across our broad health system and collaborating with our colleagues at Weill Cornell Medicine,” said Gupta, who also serves as radiologist-in-chief for NewYork-Presbyterian’s CUIMC campus. “This is exactly the kind of cross-campus science that NewYork-Presbyterian makes possible, and we believe it will translate into real benefits for patients.”

CAROTID-MAP is funded through NIH grant 1R01HL174863 from the National Heart, Lung, and Blood Institute and runs through May 2030.