A continuously enrolling precision cardiovascular oncology cohort.

Prospective Longitudinal Cardiovascular Phenotyping Using Radiation Treatment Planning Computed Tomography, Artificial Intelligence, Radiomics, Biomarkers, Wearable Technologies, and Precision Survivorship in Patients Receiving Radiation Therapy.

Conducted by My Heart Opulence LLC dba Heart Innovation and Equity Research (HIER) Institute, in collaboration with My Heart Spark Foundation, Inc. dba Heart Spark Research & Innovation Institute, and Dartmouth Radiation Oncology.

The prospective phase of this research program represents the translational extension of the retrospective cardiovascular oncology imaging repository. Whereas the retrospective investigation establishes imaging methodologies, validates quantitative biomarkers, develops artificial intelligence algorithms, and characterizes historical associations between cardiovascular phenotype and clinical outcomes, the prospective investigation seeks to implement these discoveries in real time during routine cancer care.

The prospective study is designed as a longitudinal observational cardiovascular oncology cohort embedded within routine clinical workflows. Participants are enrolled before initiation of radiation therapy and followed prospectively throughout active cancer treatment and survivorship. The study emphasizes integration rather than disruption of clinical care. Research procedures are coordinated with existing oncology appointments whenever feasible, thereby minimizing participant burden while maximizing longitudinal scientific information.

Unlike conventional observational registries that primarily collect structured clinical variables, this protocol establishes a deeply phenotyped prospective cohort integrating imaging, radiation dosimetry, laboratory biomarkers, biospecimens, wearable technologies, patient-reported outcomes, digital health, cardiovascular imaging, artificial intelligence, radiomics, environmental factors, and long-term clinical outcomes. The resulting platform is intended to become one of the world's most comprehensive cardiovascular oncology prospective cohorts.

Cancer treatment increasingly requires balancing oncologic efficacy with preservation of long-term cardiovascular health. Current cardiovascular risk assessment before radiation therapy relies primarily upon traditional clinical risk factors, clinician judgment, and occasionally cardiovascular imaging performed for unrelated indications. These approaches incompletely characterize the biological heterogeneity that determines individual susceptibility to treatment-related cardiovascular injury.

The retrospective protocol established the feasibility of extracting quantitative cardiovascular phenotypes from planning CT examinations. The prospective study now seeks to evaluate whether those imaging-derived biomarkers can improve clinical risk prediction, personalize surveillance, identify patients requiring preventive intervention, and ultimately improve survivorship.

Prospective data collection further allows incorporation of biological information unavailable retrospectively, including serial blood biomarkers, biospecimens, wearable physiological monitoring, standardized patient-reported outcomes, functional testing, quality-of-life assessments, behavioral measures, and repeated cardiovascular imaging. These complementary data streams permit investigation of dynamic biological processes occurring before, during, and after cancer therapy, rather than relying solely upon isolated baseline measurements.

The overall objective of the prospective investigation is to establish a continuously enrolling precision cardiovascular oncology cohort capable of identifying biological mechanisms responsible for cardiovascular toxicity while simultaneously developing clinically deployable prediction models and preventive strategies. The study seeks to characterize cardiovascular health as a continuously evolving biological process rather than a collection of isolated diagnoses. Accordingly, longitudinal changes in imaging biomarkers, circulating biomarkers, physiology, symptoms, functional capacity, body composition, and cardiovascular outcomes all contribute to individualized cardiovascular trajectories.

To determine whether integration of quantitative planning CT biomarkers with longitudinal clinical, imaging, laboratory, physiological, and patient-reported information improves prediction of major adverse cardiovascular events following cancer therapy.

Secondary objectives include evaluation of:

• Longitudinal cardiovascular remodeling
• Changes in body composition
• Changes in biological aging
• Radiation dose-response relationships
• Early imaging biomarkers preceding overt cardiotoxicity
• Serial biomarker trajectories
• Wearable-derived physiology
• Quality-of-life trajectories
• Exercise patterns
• Sleep characteristics
• Digital health engagement
• Medication adherence
• Preventive cardiology implementation
• Shared decision-making
• Survivorship planning

Exploratory objectives include development of:

• Digital twins
• Foundation imaging models
• Multimodal artificial intelligence
• Precision survivorship algorithms
• Adaptive surveillance strategies
• Individualized cardiovascular screening intervals
• Clinical decision support systems
• Federated learning collaborations
• Multicenter implementation science
• Adaptive pragmatic clinical trials

The study utilizes a prospective longitudinal observational cohort design. Eligible participants are approached before initiation of radiation therapy. After informed consent, participants undergo standardized baseline cardiovascular phenotyping and subsequently undergo repeated assessments coordinated with routine oncology care whenever feasible.

The observational nature of the study permits evaluation of natural biological variability while minimizing participant burden. No investigational therapy is mandated by the protocol. Clinical care remains entirely under the direction of treating physicians. Research activities supplement rather than replace routine clinical care.

Enrollment occurs within Dartmouth Radiation Oncology clinics. Potential participants are generally identified after the treating radiation oncologist has determined that radiation therapy is clinically appropriate. Whenever possible, recruitment occurs before CT simulation. This timing allows collection of baseline physiological information before treatment planning.

Potential participants receive written and verbal information describing the study. Sufficient time is provided for questions. Participation is entirely voluntary. Declining participation does not influence clinical care. Participants may withdraw at any time without affecting their medical treatment.

• Adults eighteen years of age or older.
• Diagnosis of malignancy requiring radiation therapy.
• Planning CT scheduled.
• Ability to provide informed consent.
• Anticipated longitudinal follow-up within the health system.
• Willingness to allow research use of imaging and clinical information.

Optional study components additionally require:

• Willingness to provide biospecimens
• Willingness to wear physiological monitoring devices
• Willingness to complete electronic questionnaires
• Willingness to participate in future follow-up

• Individuals unable to provide informed consent.
• Patients whose anticipated survival precludes baseline data collection.
• Patients declining research authorization.
• Individuals for whom participation would impose excessive burden according to treating clinicians.

Importantly, comorbid cardiovascular disease is not an exclusion criterion. The study intentionally seeks broad representation of cardiovascular health.

Baseline evaluation represents one of the most scientifically valuable components of the prospective protocol. Whenever feasible, the baseline visit coincides with CT simulation or another scheduled oncology appointment. Baseline evaluation may include:

• Comprehensive medical history
• Cancer history
• Cardiovascular history
• Medication review
• Family history
• Lifestyle assessment
• Nutrition assessment
• Physical activity assessment
• Sleep assessment
• Psychosocial assessment
• Anthropometric measurements
• Vital signs
• Frailty assessment
• Functional capacity assessment
• Patient-reported outcomes
• Quality-of-life questionnaires
• Blood collection
• Planning CT acquisition
• Radiation treatment planning
• Electrocardiography if clinically obtained
• Echocardiography if clinically indicated
• Additional cardiovascular imaging if clinically obtained

No clinically unnecessary testing is mandated solely for research unless specifically approved through future protocol amendments.

The prospective cohort establishes a longitudinal biospecimen repository supporting future mechanistic investigation. Whenever participants provide separate consent, biospecimens may include:

• Whole blood
• Serum
• Plasma
• Peripheral blood mononuclear cells
• DNA
• RNA
• Cell-free DNA
• Cell-free RNA
• Extracellular vesicles
• Metabolomic specimens
• Proteomic specimens
• Inflammatory biomarker specimens
• Immune profiling specimens
• Future omics analyses

Whenever possible, specimen collection coincides with clinically indicated phlebotomy to minimize additional venipuncture. Aliquots are processed using standardized operating procedures, and temperature monitoring, chain of custody, time to processing, freeze-thaw cycles, storage location, inventory tracking, and quality assurance are documented prospectively.

Participants undergo longitudinal follow-up throughout active treatment and survivorship. Representative follow-up intervals include baseline, mid-radiation, end of radiation, three months, six months, one year, and annually thereafter. The exact schedule may vary according to disease type, treatment intensity, cardiovascular risk, and future protocol amendments.

Each follow-up encounter updates:

• Cardiovascular events
• Cancer status
• Medication changes
• Hospitalizations
• Laboratory values
• Imaging studies
• Symptoms
• Quality of life
• Exercise
• Nutrition
• Sleep
• Wearable physiology
• Body weight
• Functional status
• Patient goals
• Survivorship planning

Participants may optionally use wearable physiological devices. Potential measurements include:

• Heart rate
• Heart rate variability
• Physical activity
• Walking distance
• Step count
• Exercise intensity
• Sedentary time
• Sleep duration
• Sleep stages
• Respiratory rate
• Temperature
• Cardiac rhythm notifications
• Blood pressure when available
• Continuous glucose monitoring for selected substudies

These data permit continuous characterization of cardiovascular physiology between clinic visits. Importantly, wearable devices complement rather than replace clinical evaluation.

One of the principal advantages of the prospective investigation is the opportunity to characterize cardiovascular structure and function as dynamic biological processes rather than static baseline observations. Cardiovascular remodeling associated with cancer therapy unfolds over months to years through complex interactions among radiation exposure, systemic therapy, inflammation, vascular injury, myocardial adaptation, metabolic changes, aging, and recovery. Cross-sectional imaging captures only a single point along this trajectory; serial imaging permits direct observation of biological evolution.

Every clinically obtained cardiovascular imaging examination during follow-up becomes an integral component of the prospective repository. The study does not mandate imaging beyond accepted clinical practice; instead, it capitalizes on imaging already performed as part of patient care. This pragmatic approach minimizes participant burden while preserving ecological validity.

Serial echocardiography permits evaluation of ventricular remodeling, chamber dimensions, systolic and diastolic function, myocardial strain, valvular function, pulmonary artery pressures, right ventricular performance, and pericardial abnormalities. Global longitudinal strain frequently identifies subclinical myocardial dysfunction before measurable reductions in ejection fraction.

Cardiac magnetic resonance imaging obtained during routine care provides detailed assessment of ventricular function, myocardial fibrosis, edema, inflammation, extracellular volume fraction, tissue characterization, perfusion, scar formation, and infiltrative disease. These studies are harmonized with planning CT biomarkers whenever feasible to investigate relationships between baseline anatomical phenotype and subsequent myocardial remodeling. Coronary CT angiography, nuclear cardiology, stress testing, invasive angiography, and vascular imaging likewise contribute to comprehensive longitudinal phenotyping under standardized extraction procedures defined by the imaging manual.

The prospective cohort provides a unique opportunity to investigate circulating biomarkers reflecting cardiovascular injury, inflammation, thrombosis, endothelial dysfunction, fibrosis, oxidative stress, immune activation, metabolic health, and biological aging. Whenever participants provide optional biospecimen consent, serial blood samples may be obtained in coordination with clinically indicated phlebotomy; the protocol is intentionally designed to avoid unnecessary venipuncture.

The cardiovascular biomarker program may include:

• High-sensitivity cardiac troponin I or T
• B-type natriuretic peptide and NT-proBNP
• High-sensitivity C-reactive protein
• Interleukin-6 and tumor necrosis factor alpha
• Soluble ST2 and galectin-3
• Growth differentiation factor-15
• Myeloperoxidase and oxidized LDL
• Lipoprotein(a) and apolipoproteins
• Fibrinogen and D-dimer
• Cystatin C and markers of renal function
• Endothelial injury biomarkers
• Circulating extracellular vesicles
• Metabolomic and proteomic profiles
• Emerging biomarkers as scientific knowledge evolves

Serial biomarker trajectories permit investigation of temporal relationships between circulating biological changes and evolving imaging phenotypes. Rather than treating biomarkers as isolated laboratory values, the study characterizes coordinated biological responses occurring throughout the course of cancer treatment.

Participants providing separate consent may contribute biospecimens supporting future germline genomic sequencing, somatic mutation profiling from clinical oncology testing where available, transcriptomics, epigenomics, methylation profiling, proteomics, metabolomics, lipidomics, microbiome analyses, extracellular vesicle characterization, immune profiling, and systems biology integration. Biospecimens are processed and stored using procedures preserving flexibility for emerging analytical methodologies not yet available at study initiation. All future molecular analyses beyond the scope of the present protocol will undergo additional scientific review and regulatory oversight as required.

Clinical outcomes represent only one dimension of survivorship. Patients frequently experience symptoms, functional limitations, emotional challenges, cognitive changes, fatigue, financial toxicity, and quality-of-life alterations that profoundly influence long-term health despite remaining incompletely reflected within routine medical records. The prospective study therefore incorporates comprehensive electronic patient-reported outcomes.

Domains include:

• Physical functioning and exercise tolerance
• Fatigue, dyspnea, chest discomfort, palpitations
• Sleep quality
• Emotional well-being, anxiety, depression, resilience
• Social functioning
• Financial burden and treatment satisfaction
• Symptom burden and health literacy
• Medication adherence, nutrition, physical activity
• Cognitive function and sexual health
• Survivorship priorities

Validated instruments are selected whenever available. The specific battery may evolve through amendments. Electronic collection permits completion before clinic visits using secure web platforms optimized for desktop, tablet, and mobile devices.

Digital phenotyping refers to continuous characterization of health using information generated through interactions with digital technologies — wearable physiology, smartphone activity, sleep behavior, communication rhythms, mobility, environmental exposures, geospatial movement, and patient-generated health data. Participation in digital phenotyping components remains entirely optional.

Collected information may include daily step count, exercise intensity, sedentary behavior, sleep duration, sleep efficiency, heart rate variability, resting and recovery heart rate, respiratory rate, mobility indices, geospatial environmental exposures, ambient temperature, air quality, and additional physiological variables available through approved digital platforms. The objective is not continuous surveillance but rather development of richer longitudinal representations of cardiovascular physiology between clinic visits.

The prospective cohort utilizes a secure participant-facing digital platform supporting communication, study engagement, electronic consent when appropriate, questionnaire completion, educational resources, appointment reminders, biospecimen scheduling, wearable integration, and dissemination of general study updates.

The platform emphasizes simplicity, accessibility, and inclusiveness. Participants complete study activities from smartphones, tablets, desktop computers, or institutional devices without requiring advanced technical expertise. Accessibility features support individuals with visual impairment, hearing impairment, reduced dexterity, multilingual needs in future expansions, and varying digital literacy. Offline functionality may be incorporated for selected components to improve participation among individuals with inconsistent internet connectivity.

A distinguishing characteristic of the protocol is explicit incorporation of implementation science from the earliest stages of investigation. Many scientifically valid cardiovascular biomarkers ultimately fail to influence patient care because implementation barriers remain unaddressed. This study seeks not only to discover imaging biomarkers but also to understand how those biomarkers may eventually become integrated into routine radiation oncology workflows.

Implementation outcomes include:

• Feasibility, acceptability, appropriateness
• Adoption, fidelity, sustainability
• Workflow integration and clinician satisfaction
• Patient satisfaction
• Time requirements and computational efficiency
• Cost considerations and resource utilization
• Organizational readiness

Mixed-methods investigations incorporating quantitative metrics and qualitative interviews may be performed under future amendments to understand barriers and facilitators affecting clinical implementation.

The prospective cohort is intentionally designed to support future embedded pragmatic investigations, including studies evaluating automated coronary artery calcium reporting, AI-assisted cardiovascular referral pathways, digital survivorship interventions, exercise promotion strategies, nutritional counseling, hypertension management programs, statin implementation, cardio-oncology consultation pathways, remote monitoring, educational interventions, and survivorship navigation.

These studies would occur under separate IRB amendments or nested protocols while leveraging the infrastructure established through the present observational cohort. This flexible umbrella design substantially reduces redundant infrastructure development while accelerating translation from discovery science to clinical implementation.

Artificial intelligence models developed using retrospective and prospective data may ultimately support clinical decision-support tools integrated into routine radiation oncology workflows. Potential applications include automated identification of coronary artery calcification, individualized cardiovascular risk estimation, cardio-oncology referral recommendation, prioritization of preventive cardiology evaluation, automated generation of structured cardiovascular imaging reports, and individualized survivorship surveillance recommendations. All clinical decision- support tools undergo extensive validation before any clinical deployment. During the observational phase of the present protocol, investigational outputs do not influence patient management.

Long-term follow-up represents one of the greatest scientific assets of the prospective cohort. Retention strategies emphasize respect, flexibility, communication, convenience, and ongoing engagement. Study activities are coordinated with existing oncology appointments whenever possible. Electronic questionnaires permit remote completion. Periodic educational newsletters describing overall study progress may be distributed without disclosing individual research findings.

Participants receive clear explanations regarding the scientific importance of continued follow-up. Contact information is updated regularly, and multiple approved communication methods are utilized according to participant preference. Retention metrics are monitored continuously, and reasons for withdrawal or loss to follow-up are documented when voluntarily provided.

Recognizing the rapid pace of scientific innovation within cardiovascular oncology, the protocol is intentionally designed as a living framework rather than a static document. Future amendments may incorporate new imaging biomarkers, emerging artificial intelligence methodologies, additional biospecimen analyses, novel wearable technologies, environmental exposure assessment, digital therapeutics, implementation strategies, pragmatic clinical studies, multicenter expansion, international collaborations, and evolving regulatory guidance. This adaptive design preserves methodological consistency while allowing the research program to evolve responsibly alongside advances in science and technology.

The ultimate vision extends beyond completion of a single observational study. The long-term aspiration is creation of a continuously learning cardiovascular oncology ecosystem in which routine clinical imaging, longitudinal physiological measurements, biospecimens, artificial intelligence, patient-reported outcomes, radiation dosimetry, and survivorship experiences collectively inform continuous refinement of cardiovascular prevention strategies.

Every newly enrolled participant enriches the repository. Every additional imaging examination improves artificial intelligence models. Every longitudinal outcome enhances prediction algorithms. Every validated discovery contributes to progressively more individualized survivorship care. In this manner, the research program transforms routine clinical care into an engine of continuous scientific discovery while simultaneously returning that knowledge to improve future patient care.

The statistical framework is intentionally designed to support both immediate observational investigations and future expansion into multicenter implementation studies, pragmatic clinical trials, adaptive platform investigations, AI validation, radiomics discovery, and precision survivorship research. Rather than designing the study around a single endpoint or one hypothesis, the protocol establishes a comprehensive scientific infrastructure capable of supporting multiple independent but interconnected investigations over many years. Planning emphasizes scalability, reproducibility, flexibility, and continuous refinement, prioritizing estimation of clinically meaningful effect sizes, individualized risk prediction, calibration, biological discovery, external validation, and translational implementation over exclusive reliance upon null hypothesis significance testing.

Unlike conventional randomized clinical trials in which enrollment concludes after reaching a predetermined sample size, the present protocol establishes a continuously expanding observational cohort. Sample size is therefore dynamic rather than fixed. Initial enrollment targets are determined collaboratively with Dartmouth Radiation Oncology based upon anticipated patient volume, available research personnel, imaging processing capacity, institutional infrastructure, and funding. Every additional participant contributes incremental scientific value by improving precision of imaging biomarker estimates, strengthening AI models, increasing phenotypic diversity, enhancing external validity, and expanding opportunities for nested investigations. As longitudinal follow-up increases, the statistical value of each participant similarly increases because repeated observations contribute additional biological information over time.

Pilot Phase

The pilot phase emphasizes workflow optimization, imaging pipeline validation, biospecimen processing, data harmonization, quality assurance, and feasibility. Approximately several hundred participants may be sufficient for operational validation and preliminary biomarker analyses.

Expansion Phase

Following successful pilot implementation, enrollment expands to several thousand participants representing diverse malignancies, treatment approaches, demographic groups, cardiovascular phenotypes, and survivorship trajectories. This phase supports robust multivariable analyses, AI development, radiomics investigations, subgroup analyses, and disease-specific cardiovascular toxicity models.

Mature Registry

The long-term vision encompasses tens of thousands of participants through institutional growth, multicenter collaboration, and national expansion. At this scale, the repository becomes capable of supporting rare cardiovascular outcomes, uncommon cancer types, individualized treatment response analyses, deep learning development, federated learning initiatives, and precision medicine investigations not feasible within smaller cohorts.

Formal power calculations are performed separately for each predefined primary analysis. Because future nested investigations evaluate diverse endpoints with varying event rates, a single universal power calculation would not appropriately characterize the entire research program. Each major manuscript, grant application, or embedded study includes endpoint-specific sample size justification based upon anticipated event frequency, clinically meaningful effect size, expected covariate structure, longitudinal follow-up duration, missing data assumptions, competing risks, and planned statistical methodology. Simulation-based power estimation is used for complex AI models and high-dimensional radiomics analyses.

Because this investigation is observational, interim analyses evaluate scientific progress rather than treatment efficacy. Periodic analyses assess recruitment, participant diversity, follow-up completeness, biospecimen quality, imaging processing completion, AI performance, radiomics reproducibility, data completeness, endpoint accrual, publication productivity, and grant development. These analyses inform operational improvements while preserving integrity of future confirmatory investigations.

The operational workflow is intentionally designed to integrate seamlessly into routine oncology practice. The participant pathway begins when a patient is scheduled for radiation oncology consultation. Potentially eligible patients are identified electronically through scheduling systems or radiation oncology information systems. Following confirmation of clinical eligibility by treating physicians, research personnel review study eligibility.

When appropriate, patients receive study information before CT simulation. Following informed consent, baseline data collection occurs in coordination with routine clinical care. Planning CT acquisition proceeds according to standard clinical protocols, and clinical treatment remains entirely under the direction of treating physicians. Research personnel coordinate biospecimen collection, questionnaires, wearable device enrollment, longitudinal follow-up, imaging processing, and data integration. Throughout follow-up, clinical information generated during routine care continuously enriches the research repository.

The centralized Research Coordination Center is responsible for day-to-day operational management, including participant enrollment, study scheduling, biospecimen coordination, data management, quality assurance, regulatory documentation, training, monitoring, meeting coordination, communication, budget management, report preparation, grant support, and publication tracking. The Coordination Center serves as the operational heart of the research enterprise.

The Imaging Core Laboratory oversees all quantitative imaging analyses, including image import, quality assessment, preprocessing, segmentation, manual review, AI validation, radiomics extraction, image registration, dose registration, feature generation, database integration, image archiving, software validation, and quality assurance. The Imaging Core also develops standardized operating procedures for future collaborating institutions.

The Artificial Intelligence Core coordinates model development, model validation, code management, software version control, computational infrastructure, cloud computing, high-performance computing, deep learning, multimodal learning, explainable AI, algorithm documentation, federated learning preparation, regulatory documentation, algorithm auditing, performance monitoring, and deployment readiness. The AI Core works closely with the Imaging Core and Statistical Core throughout the project.

The Statistical Core oversees study design, power calculations, randomization procedures for future trials, database review, quality control, missing data, statistical programming, interim analyses, manuscript statistics, grant statistics, AI evaluation, prediction modeling, calibration, validation, visualization, and regulatory statistical reports. The Statistical Core maintains all primary analysis code using reproducible computational workflows.

The Biospecimen Core oversees sample collection, labeling, chain of custody, processing, centrifugation, aliquoting, cryopreservation, inventory, quality monitoring, shipping, freezer monitoring, temperature documentation, and future specimen requests. All biospecimens undergo standardized processing using validated procedures.

The Regulatory Core coordinates IRB submissions, continuing review, protocol amendments, HIPAA documentation, data use agreements, material transfer agreements, business associate agreements, investigator training, Good Clinical Practice documentation, monitoring reports, audit preparation, regulatory correspondence, and adverse event reporting. The Regulatory Core maintains complete electronic regulatory binders.

The study operates under a comprehensive Quality Management System designed to promote reproducibility and continuous improvement.

• Written standard operating procedures
• Training manuals and competency assessments
• Annual retraining
• Document control and version control
• Change management and deviation reporting
• Corrective and preventive actions
• Risk assessments and quality indicators
• Internal and external audits
• Continuous process improvement

Because this observational investigation involves minimal physical risk, a traditional pharmaceutical-style Data Safety Monitoring Board is not anticipated to be required during the observational phase. Scientific oversight nevertheless remains essential and includes regulatory compliance, participant protection, privacy protection, data quality, protocol adherence, quality assurance, biospecimen integrity, AI governance, statistical oversight, publication review, budget oversight, and committee reporting. Should future embedded interventional studies be introduced, an independent DSMB may subsequently be established.

Every investigator joining the study completes standardized onboarding including human subjects protection, HIPAA, Good Clinical Practice, research ethics, data management, image processing, quality assurance, biospecimen handling, radiation oncology orientation, cardiovascular oncology fundamentals, AI governance, publication policy, and conflict-of-interest management. Annual competency review maintains consistency across the investigative team.

Phase I

Infrastructure development, regulatory approvals, database development, software validation, training, and pilot enrollment.

Phase II

Full enrollment, imaging processing, artificial intelligence development, radiomics analyses, publication initiation, and grant expansion.

Phase III

Longitudinal follow-up, external validation, multicenter collaboration, clinical implementation, decision-support development, national expansion, international collaboration, and learning health system integration.

Long-term sustainability depends upon continuous integration of scientific discovery, clinical collaboration, education, mentorship, technology development, external funding, philanthropy, industry partnerships consistent with institutional policies, and dissemination of high-impact research. Rather than functioning as a single grant-supported project, the program is envisioned as a permanent academic cardiovascular oncology research ecosystem that continually generates new scientific questions, develops innovative analytical methodologies, trains future investigators, improves patient care, and serves as a national resource for precision cardiovascular oncology.

The cardiovascular oncology imaging program is envisioned as a long-term institutional research enterprise rather than a single investigator-initiated project. The budget supports infrastructure necessary to establish, maintain, expand, and continuously improve a precision cardiovascular oncology research ecosystem integrating radiation oncology, cardiovascular medicine, artificial intelligence, radiomics, implementation science, digital health, translational biology, education, and future multicenter collaboration. Investment emphasizes durable infrastructure that will continue generating scientific value across multiple grants, manuscripts, trainees, collaborations, and future clinical trials. Although individual grant mechanisms may support subsets of the program, the complete budget reflects the mature program necessary to sustain a nationally recognized Cardiovascular Oncology Imaging Center.

• Personnel
• Research Coordination
• Imaging Core
• Artificial Intelligence Core
• Biostatistics Core
• Data Management Core
• Clinical Informatics Core
• Radiation Physics Support
• Biospecimen Core
• Laboratory Support
• Computing Infrastructure
• Software
• Cloud Computing
• REDCap and Database Development
• Regulatory Support
• Quality Assurance
• Participant Costs
• Scientific Meetings
• Publications
• Education
• Equipment
• Indirect Costs

Personnel represent the largest investment because multidisciplinary expertise is essential for success. Representative effort allocations include:

• Principal Investigator — 15–30% effort (scientific leadership, oversight, grants, mentorship)
• Radiation Oncology Co-Investigator — 10–20% (protocol oversight, cardiac contour review)
• Cardio-Oncology Co-Investigator — 15–20% (endpoint adjudication, imaging review, implementation)
• Medical Physics Faculty — 10–15% (dose reconstruction, cardiac dosimetry, planning analytics)
• Radiologist — 5–10% (imaging interpretation, segmentation oversight)
• Biomedical Informatics Faculty — 10–20% (EDW, Epic extraction, integration)
• Artificial Intelligence Faculty — 20% (deep learning, computer vision, foundation models)
• Senior Biostatistician — 25% (design, analysis, prediction modeling, validation)
• Research Program Director — 100% (daily operations, milestones, grants, budget)
• Clinical Research Coordinators — 2–5 FTE (recruitment, consent, follow-up, regulatory)
• Research Nurses — 1–3 FTE (biospecimens, participant coordination, education)
• Imaging Scientists — 2–4 FTE (processing, segmentation, QC, radiomics)
• Machine Learning Engineers — 2–5 FTE (models, pipelines, deployment, cloud)
• Software Engineers — 2–4 FTE (research platform, web, mobile, dashboards, APIs)
• Database Engineers — 1–2 FTE (REDCap, SQL, warehouse, ETL)
• Regulatory Specialists — 1–2 FTE (IRB, HIPAA, audits, documentation)
• Laboratory Technologists — 1–2 FTE (processing, freezer management, shipping)
• Graduate Students — 4–8 trainees (imaging, AI, statistics, radiomics, BME)
• Postdoctoral Fellows — 2–6 fellows (AI, cardiovascular imaging, radiomics, informatics)
• Medical Students & Undergraduates — variable effort (summer scholars, capstones, annotation, QA)

Imaging Core

High-performance workstations, DICOM storage, image archive, segmentation software, QA software, 3D visualization, image registration, radiomics software, and AI inference servers.

Artificial Intelligence Core

GPU clusters, deep learning servers, cloud GPUs, model storage, version control, experiment tracking, containerization, continuous integration, MLOps, and monitoring.

Cloud Computing

Azure, AWS, Google Cloud, and institutional HPC — secure object storage, GPU instances, backup, long-term archive, and disaster recovery.

Software

MATLAB, Python ecosystem, R, SAS, Stata, 3D Slicer, ITK/SimpleITK, MONAI, PyTorch, TensorFlow, Docker, GitHub Enterprise, REDCap, SQL Server, Power BI, Tableau, EndNote, NVivo, ArcGIS, and additional tools as required.

Database Infrastructure

REDCap development, electronic Case Report Forms, APIs, automated extraction, EDW integration, Epic integration, FHIR/HL7 support, audit logging, and version control.

Biospecimen Infrastructure

Ultra-low freezers, liquid nitrogen, cryovials, barcode systems, temperature monitoring, shipping, processing supplies, laboratory consumables, and quality assurance.

Participant, Meeting, Publication & Equipment Costs

Parking, travel reimbursement, gift cards where approved, wearable devices, questionnaire completion incentives, participant newsletters; meeting support for AHA, ASCO, ASTRO, SCMR, RSNA, SNMMI, ACC, IC-OS, ML conferences, and implementation science meetings; open-access fees, figure preparation, professional illustration, supplemental materials, repository deposition, data sharing; and high-performance GPU servers, storage arrays, backup systems, workstations, digital pathology displays, large monitors, network upgrades, and conference/video technology.

Personnel represent the primary investment because successful execution requires expertise spanning cardiovascular medicine, radiation oncology, medical physics, artificial intelligence, biomedical informatics, radiomics, epidemiology, biostatistics, laboratory science, software engineering, implementation science, clinical research operations, and regulatory management. Computing infrastructure is similarly essential because processing tens of thousands of planning CT examinations, extracting millions of radiomic features, training deep neural networks, storing DICOM datasets, and maintaining secure research environments require substantial computational resources.

Investment in centralized cores increases efficiency by allowing imaging scientists, statisticians, AI investigators, laboratory personnel, and research coordinators to support numerous projects simultaneously rather than duplicating effort across individual studies. The resulting infrastructure creates a sustainable scientific ecosystem capable of supporting multiple grants, hundreds of manuscripts, trainee education, investigator development, technology commercialization, and long-term institutional leadership in cardiovascular oncology.

The research enterprise intentionally diversifies funding sources rather than relying upon a single grant mechanism.

• Federal research grants (R01, U01, P01, P50, SPORE)
• Program Project Grants and Cooperative Agreements
• Foundation funding and philanthropic gifts
• Cancer Center Support Grants and institutional investment
• Industry-sponsored research (consistent with institutional policies)
• Technology licensing and AI intellectual property
• Educational programs
• International collaborations

This diversified portfolio increases resilience and supports continuous scientific growth.

The ultimate goal extends well beyond creation of an imaging repository or publication of scientific manuscripts. The vision is to establish an enduring Cardiovascular Oncology Imaging Institute that serves as an international center for scientific discovery, clinical innovation, investigator training, artificial intelligence development, implementation science, and precision survivorship.

Over the coming decades, every radiation treatment planning CT obtained through participating institutions has the potential to contribute simultaneously to clinical care, scientific discovery, cardiovascular prevention, education, technology development, and continuous improvement of cancer survivorship. The collaboration among Dartmouth Radiation Oncology, HIER Institute, and Heart Spark RII provides a foundation upon which an internationally recognized program in precision cardiovascular oncology can be built, sustained, and continuously expanded — for the benefit of patients, clinicians, scientists, and future generations of investigators.

Forthcoming amendments to this prospective master protocol will expand the investigation to encompass serial cardiovascular imaging, comprehensive blood biomarker panels, standardized patient-reported outcome instruments, digital phenotyping, implementation science workflows, embedded pragmatic clinical trials, adaptive surveillance pathways, AI-driven clinical decision support, participant retention strategies, formal sample size justification, prespecified statistical analysis plans, and multi-institutional governance — building this cohort to a level suitable for major NIH U01 / P01, Program Project, and Cancer Center Support Grant applications.

The prospective cohort is architected from inception as a federated, multi-institutional consortium rather than a single-site study. Following maturation of the Dartmouth flagship site, additional academic radiation oncology programs, National Cancer Institute-designated cancer centers, integrated health systems, and international radiation oncology consortia will be onboarded through a phased expansion strategy. Site qualification incorporates imaging quality assessment, informatics readiness evaluation, human subjects protection review, investigator training, and executed data use agreements. Each site retains local custody of raw imaging and clinical data while contributing standardized, de-identified analytic packages to the coordinating center — a federated model consistent with contemporary consortium science and modern data sovereignty expectations.

Site Onboarding Framework

• Site Feasibility Questionnaire and imaging volume audit
• Local IRB reliance or single-IRB agreement (SMART IRB)
• Standardized site initiation visit with imaging phantom validation
• Federated informatics onboarding — no raw PHI transfer required
• Local principal investigator training and delegation logs
• Quarterly site performance dashboards with remediation pathways

Cardiovascular oncology outcomes are shaped by structural inequities that manifest in imaging access, radiation planning practices, cardio-oncology referral patterns, biomarker availability, and survivorship engagement. This protocol embeds equity science as a first-class scientific discipline rather than a compliance activity. Enrollment monitoring stratifies participation by race, ethnicity, preferred language, rurality, insurance status, and social determinants of health. Recruitment materials are professionally translated and community-reviewed. Algorithmic fairness auditing is a required deliverable for every predictive model, with subgroup calibration reports released alongside primary performance metrics.

• Enrollment equity dashboards with monthly reporting
• Community-based participatory research advisory structure
• Language-concordant recruitment and consent workflows
• Neighborhood-level SDoH linkage via validated indices
• Model fairness audits with subgroup calibration reports
• Transportation and digital access support for under-resourced participants

A dedicated protocol addendum will extend the platform to pediatric and adolescent/young adult (AYA) populations receiving thoracic, mediastinal, craniospinal, or total body irradiation. Pediatric radiation oncology introduces unique cardiovascular considerations including developmental cardiotoxicity, lifetime attributable risk, growth-related dosimetric variance, and multi-decade survivorship. The pediatric extension will incorporate age-appropriate consent and assent processes, family-centered communication, developmentally appropriate wearable devices, transition-of-care pathways, and integration with the Children's Oncology Group Long-Term Follow-Up Guidelines.

A significant fraction of the global cancer burden — and an even greater fraction of preventable cardiovascular mortality following cancer therapy — occurs in low- and middle-income countries where dedicated cardio-oncology infrastructure is limited. Because every course of radiation therapy already generates a planning CT, the platform can extend cardiovascular phenotyping to global settings without requiring additional imaging infrastructure. Partnerships with global radiation oncology networks, IAEA collaborating centers, and consortium members in sub-Saharan Africa, South Asia, Latin America, and the Middle East will support locally led scientific leadership, capacity building, and equitable knowledge exchange.

Strategic partnerships with medical device manufacturers, contrast and imaging vendors, artificial intelligence developers, wearable technology firms, and biopharmaceutical companies will accelerate translation while maintaining scientific independence. All industry engagements are governed by a written Industry Engagement Policy that preserves publication rights, guarantees investigator-initiated analyses, prohibits data exclusivity in scientific findings, and requires transparent disclosure. Sponsored analyses, in-kind imaging services, wearable device provision, and biomarker assay support are welcomed within these constraints.

The platform's intellectual property strategy balances open science with sustainable translation. Scientific findings, imaging biomarker definitions, and analytic protocols are released openly to advance the field. Novel algorithms, software implementations, clinical decision support workflows, and integrated cardiovascular reporting products may be protected through patents, copyrights, and trade secrets held by HIER Institute or jointly with participating academic partners under standard technology transfer agreements. Licensing revenue is directed back to program sustainability, trainee support, and equity initiatives.

Selected imaging biomarker readouts and AI-generated cardiovascular risk assessments derived from planning CT are anticipated to progress toward Food and Drug Administration clearance as Software as a Medical Device (SaMD). The regulatory strategy contemplates a staged approach: (1) analytical validation through the retrospective cohort, (2) clinical validation through the prospective cohort, (3) prospective observational deployment studies at pilot sites, (4) pre-submission engagement with the FDA Digital Health Center of Excellence, and (5) 510(k) or De Novo classification submissions for defined intended uses. Quality Management System, Good Machine Learning Practice, and Predetermined Change Control Plan documentation are maintained continuously.

• Analytical validation: locked model, sequestered test set, performance envelope
• Clinical validation: prospective multi-site accuracy and reproducibility
• Human factors and usability engineering (IEC 62366-1)
• Cybersecurity and software lifecycle documentation (IEC 62304)
• Good Machine Learning Practice (FDA/Health Canada/MHRA principles)
• Predetermined Change Control Plan for model updates
• Post-market performance monitoring and real-world evidence generation

The platform commits to Findable, Accessible, Interoperable, and Reusable (FAIR) data governance consistent with the NIH Data Management and Sharing Policy. A tiered access model supports scientific reuse while protecting participant privacy: (1) fully open aggregate statistics and code, (2) controlled-access de-identified derivative datasets via NIH-approved repositories such as TCIA, dbGaP, and NHLBI BioLINCC, and (3) federated analytic access for approved collaborators requiring image-level or genomic-level data. Every dataset is accompanied by machine-readable metadata, provenance records, and versioned analytic code.

A Patient and Community Advisory Board is convened as a standing governance body rather than a consultative afterthought. Members include cancer survivors, caregivers, cardio-oncology patients, community health workers, and lived- experience advocates representing the geographic and demographic breadth of the participating population. The Board reviews recruitment materials, informed consent language, retention strategies, dissemination products, and manuscripts of participant-facing significance. Members receive honoraria consistent with PCORI standards.

A layered dissemination strategy ensures scientific findings reach the audiences that can act upon them: peer-reviewed publications for the scientific community, plain-language summaries for participants and the public, clinician-facing briefings for cardio-oncology and radiation oncology practices, policy briefs for payers and regulators, and educational modules for trainees. Every peer-reviewed publication is paired with an accessible participant-facing summary released simultaneously.

The initiative operates a structured pipeline spanning undergraduate research experiences, medical student scholarly projects, resident and fellow research rotations, postdoctoral fellowships, junior faculty career development awards, and mid-career diversification opportunities. Dedicated slots prioritize trainees from historically underrepresented backgrounds in medicine, imaging science, and biomedical engineering. Mentorship is formalized through individual development plans and quarterly review.

The deployment of artificial intelligence into cardiovascular oncology practice raises substantive ethical questions regarding autonomy, informed consent for algorithmic care, distribution of benefits and burdens, opacity of complex models, and the potential for automation bias to displace clinical judgment. An AI Ethics Subcommittee reporting to the Executive Committee reviews all model deployment proposals against a written framework addressing transparency, contestability, accountability, subgroup performance, environmental cost, and the appropriateness of the clinical context.

Computational cardiovascular imaging carries measurable environmental costs through data center energy consumption and imaging archive storage. The platform commits to environmentally responsible research practices including efficient model architectures, batched inference scheduling, use of renewable-powered cloud regions, storage lifecycle management, and annual carbon accounting for computational research activities. Sustainability metrics are reported alongside scientific performance metrics.

A formal Research Continuity Plan governs response to interruptions ranging from cyberattacks and cloud outages to pandemics, natural disasters, and institutional transitions. The plan defines critical research functions, minimum viable operations, participant safety pathways, data protection procedures, communication protocols, and recovery time objectives. Tabletop exercises are conducted annually.

• Enrollment velocity and demographic representativeness vs. targets
• Imaging pipeline throughput and quality assurance pass rates
• Biospecimen collection completeness and long-term integrity metrics
• Publication output, citation impact, and altmetric reach
• Grant capture rate and portfolio diversification index
• Participant retention and patient-reported experience scores
• Model fairness metrics across pre-specified subgroups
• Environmental footprint per petabyte of imaging analyzed
• Trainee outputs and career transition outcomes

Even a continuously enrolling cohort requires a defined closeout and legacy framework so that scientific value persists beyond the operational lifespan of any single institution. Upon eventual programmatic closeout, all de-identified imaging, clinical, dosimetric, biomarker, genomic, patient-reported outcome, and wearable datasets — together with analytic code, model weights, documentation, and curated derivative datasets — are transferred to permanent public repositories (TCIA, dbGaP, NHLBI BioLINCC, or successor archives) under FAIR governance. Biospecimens are transferred to a designated academic biorepository. Participants receive a closeout communication describing where the research legacy resides and how they may continue to be informed of downstream discoveries.