Research

Five themes. Six diseases. One imaging platform.

The initiative is organized into five overarching scientific themes and six disease-specific programs — each grounded in the same underlying resource: the routine treatment planning CT.

Scientific Themes

The five research themes.

Each theme is designed to sustain a durable portfolio of investigations spanning retrospective analysis, prospective observation, AI validation, and pragmatic implementation.

01

Opportunistic Cardiovascular Screening

Every planning CT becomes a cardiovascular assessment.

The overwhelming majority of information contained within a treatment planning CT is never extracted for cardiovascular purposes. This theme systematically quantifies a comprehensive panel of imaging biomarkers already visible on every study.

Coronary artery calciumThoracic aortic calciumMitral annular calcificationAortic valve calcificationCardiac chamber enlargementPericardial calcificationEpicardial fatPericardial fatAscending aortic diameterPulmonary artery diameterThoracic muscle massFrailty markersVertebral bone densityBody compositionSarcopeniaVisceral adipositySubcutaneous adiposityLiver attenuationEmphysemaAirway diseasePulmonary fibrosis
02

Radiation Dose–Cardiac Structure Atlas

Map delivered dose to every cardiac substructure — and link it to lifetime events.

We create individualized dose maps to twenty or more cardiac and thoracic structures and associate delivered dose with future cardiovascular events, imaging findings, arrhythmias, heart failure, and mortality.

Whole heartLeft ventricleRight ventricleLeft atriumRight atriumLeft anterior descending arteryRight coronary arteryCircumflex arteryAortic valveMitral valvePulmonary valveTricuspid valvePericardiumConduction systemAscending aortaDescending aortaPulmonary arteriesEsophagusGreat vessels
03

AI Cardiovascular Phenotyping

Deep learning that segments, quantifies, and phenotypes at scale.

Automated pipelines detect and measure coronary calcium, vascular calcification, cardiomegaly, atrial enlargement, epicardial fat, thoracic aortic aneurysm, pulmonary hypertension, body composition, sarcopenia, muscle quality, bone density, vertebral fractures, and frailty across entire cohorts.

Automated segmentationAutomated biomarker extractionDeep learning predictionClinical decision support
04

Cardiovascular Toxicity Prediction

Multi-outcome models integrating imaging, dosimetry, EHR and therapy.

We develop and validate prediction models for the full clinical spectrum of cardiovascular disease among patients receiving radiation and systemic cancer therapies.

Heart failureAtrial fibrillationCoronary diseaseMyocardial infarctionStrokePulmonary hypertensionValvular diseaseCardiomyopathyMyocarditisPericarditisArrhythmiasConduction diseaseSudden cardiac deathOverall survival
05

Precision Radiation Planning

Baseline cardiovascular phenotype informs how we treat the cancer.

This theme investigates whether pretreatment cardiovascular phenotype should modify radiation planning: beam selection, dose constraints, heart avoidance, coronary avoidance, proton therapy selection, and adaptive planning strategies.

Beam selectionDose constraintsHeart avoidanceCoronary avoidanceProton therapy selectionAdaptive radiation planning
Disease-Specific Programs

Where the imaging meets the clinic.

Each disease program applies the platform's imaging, dosimetric, and AI capabilities to a distinct clinical population.

Breast Cancer

  • Left- vs right-sided radiation
  • Coronary artery calcium
  • LAD dose
  • Cardiac aging
  • Cardiotoxicity prediction

Lung Cancer

  • Baseline emphysema
  • Pulmonary artery enlargement
  • Coronary disease
  • Cardiovascular mortality
  • Pulmonary hypertension
  • Cachexia

Esophageal Cancer

  • Cardiac dose
  • Pericardial disease
  • Myocarditis
  • Heart failure
  • Arrhythmias
  • Survival

Lymphoma

  • Young survivors
  • Lifetime radiation exposure
  • Accelerated vascular aging
  • Valve disease

Head & Neck Cancer

  • Carotid calcification
  • Stroke prediction
  • Vascular aging

Mediastinal Tumors

  • High-dose cardiac exposure
  • Late cardiomyopathy
  • Pericardial disease
  • Coronary disease
Five-Year Strategic Roadmap

From foundation to reference center.

Years 1–2

Foundation

Establish governance, imaging repository, and prospective registry. Build automated segmentation and quantitative imaging pipelines. Complete retrospective feasibility studies and initial external funding applications.

Years 2–3

Validation

Develop and validate AI algorithms across multiple cardiovascular biomarkers. Launch prospective imaging-based cardiovascular screening workflows. Publish foundational methodology and validation studies.

Years 3–5

Expansion

Expand to multicenter collaborations. Integrate decision-support tools into radiation oncology workflows. Initiate embedded implementation and interventional trials. Establish the program as a national and international reference center.