Cardiovascular Medicine · Yale School of Medicine

Using AI to decode the hidden signals of cardiovascular and cardiometabolic disease

We develop and apply AI to multimodal data already generated in care, including echocardiography, cardiac CT, ECG, and longitudinal health records, to define cardiovascular phenotypes more precisely, detect disease earlier, and match therapies to the patients most likely to benefit.

Evangelos K. Oikonomou
Led by Evangelos K. Oikonomou, MD, DPhil
Assistant Professor of Medicine · Yale School of Medicine
Transthoracic echocardiogram, parasternal viewTTE
PanAdipo attention map over the same echocardiogramPanAdipo
Paired chest CT with segmented adipose depotsCT · adipose depots
Paired chest CT with segmented cardiac chambersCT · heart volume
PanAdipo reads routine echocardiograms and attends to the epicardium. Paired CT segmentation of adipose depots and heart volume serves as the reference. From our PanAdipo study (medRxiv 2026).

Our work follows a translational arc from computational method development to clinical deployment and discovery. Each project moves through the same five stages, and what we learn at the end starts the next cycle.

  1. 01DevelopCreate new computational techniques that align, reconstruct, and learn from multimodal clinical data.
  2. 02BuildTrain foundation models and task-specific models for defined cardiovascular questions.
  3. 03ValidateTest performance, transportability, and equity across institutions, populations, and prospective cohorts.
  4. 04DeployUse validated models to screen for disease, predict risk, and link patient phenotypes to therapies.
  5. 05DiscoverLearn from deployment to reveal disease biology and generate the next clinical and computational questions.

This cycle connects method development to measurable clinical use, then feeds new observations back into discovery.

Explore our research areas →

Program GitHub →

#Cardiometabolic-Intelligence

We are reinventing the study of heart–adipose–liver–kidney interactions through digital biomarkers. Each organ is usually measured on its own, in a different clinic and with a different test. We use AI to read these organs together from the imaging, ECGs, and health records that patients already have.

github.com/Cardiometabolic-Intelligence → Deep adiposity phenotyping →

Openings

Work with us

We welcome researchers working at the intersection of cardiovascular medicine, imaging, AI, and translational data science. Feel free to reach out.

Current openings → Contact us →

Principal Investigator

Evangelos K. Oikonomou

Evangelos K. Oikonomou MD, DPhil

  • Assistant Professor of MedicineSection of Cardiovascular Medicine, Department of Internal Medicine, Yale School of Medicine
  • Deputy EditorEuropean Heart Journal
  • Associate DirectorYale CarDS Lab
100 Church Street South, New Haven, CT 06519
Full bio → Yale Faculty Profile →