AI-assisted CT for Risk Stratification in Coronary Artery Disease (ACTION)

The ACTION Registry (Artificial Intelligence-assisted CT for Risk Stratification in Coronary Artery Disease) is a prospective, single-centre, observational patient registry conducted at the Clinical Research Facility, University Hospital Galway.

This registry is designed to systematically collect and integrate multimodal data from adults undergoing clinically indicated cardiac computed tomography (CT) to support advanced cardiovascular risk assessment using artificial intelligence (AI)-based approaches.

Registry Design and Procedures Eligible participants are consecutively enrolled at the time of referral for clinically indicated cardiac CT. Following informed consent, data are collected during a single baseline visit and supplemented by routine clinical data and follow-up information where available.

Registry procedures include:

* Collection of demographic and clinical data from medical records * Blood sampling for cardiovascular and metabolic biomarkers * Resting electrocardiogram (ECG) * Structured questionnaires capturing lifestyle, cardiovascular history, and risk factors * Standard-of-care cardiac CT imaging acquisition and expert interpretation * Recording of medication use and clinically relevant interventions A subset of participants undergoing invasive imaging (e.g., optical coherence tomography) will have additional data collected for cross-modality comparison.

Data Collection and Registry Variables

The registry captures structured data across the following domains:

* Clinical and demographic characteristics * Laboratory biomarkers * Imaging-derived parameters from cardiac CT * Lifestyle and behavioural factors * Treatment and medication data

All variables are defined in a standardized data dictionary, which specifies:

* Variable definitions and formats * Source of data (e.g., imaging system, laboratory system, questionnaire) * Coding standards where applicable * Acceptable ranges and units Data Management and Quality Assurance All data are anonymised using unique participant identifiers and stored in secure, access-controlled electronic systems compliant with GDPR and institutional data governance policies.

A comprehensive quality assurance plan is implemented, including:

Data Validation and Entry Controls

* Electronic data capture systems incorporate predefined validation rules * Range checks and logical consistency checks are applied at data entry * Automated queries are generated for missing, inconsistent, or out-of-range values Data Monitoring and Auditing * Periodic internal data reviews are conducted to ensure completeness and accuracy * Selected variables undergo source data verification against original records (e.g., imaging reports, laboratory systems, case report forms) * On-site or remote monitoring may be conducted in accordance with institutional policies Source Data Verification * A predefined proportion of records will be cross-checked with source documents * Discrepancies will be documented, reviewed, and resolved בהתאם standard procedures Standard Operating Procedures (SOPs)

Registry operations are governed by standardized procedures covering:

* Participant recruitment and consent * Data collection and entry * Imaging acquisition and reporting * Laboratory data handling * Data management and security * Statistical analysis and reporting * Adverse event identification and reporting (if applicable) * Change management and protocol amendments Sample Size Considerations As an observational registry, the study aims to recruit a representative, real-world population of adults undergoing cardiac CT. The sample size is determined pragmatically based on recruitment feasibility and expected patient volume at the study site.

The anticipated sample size is sufficient to support:

* Multivariable modelling for risk prediction * Subgroup analyses * Development and validation of AI-based models Handling of Missing Data

Missing or incomplete data may arise due to non-response, unavailable records, or technical limitations. The registry implements the following approach:

* Real-time prompts to minimise missing data at entry * Documentation of reasons for missingness where available * Statistical handling using appropriate methods (e.g., multiple imputation, sensitivity analyses), depending on the extent and pattern of missing data Statistical Analysis Plan Analyses will be conducted in accordance with a predefined statistical analysis plan.

Approaches may include:

* Descriptive statistics for baseline characteristics * Correlation analyses between imaging, clinical, and biomarker variables * Regression modelling for risk prediction and association analyses * Subgroup analyses across demographic and clinical strata * Cross-modality comparisons for imaging validation * Longitudinal analyses where follow-up data are available Model performance may be evaluated using appropriate metrics (e.g., discrimination, calibration, and reclassification measures).

Data Use and Future Applications

The registry is designed to support:

* Clinical research and hypothesis generation * Development and validation of AI-based cardiovascular risk models * Integration of imaging and clinical data for precision medicine approaches An anonymised dataset will be maintained for future research, subject to governance and ethical approvals.

Ethical and Regulatory Considerations The registry is conducted in accordance with the Declaration of Helsinki and Good Clinical Practice guidelines. Ethical approval will be obtained from the appropriate research ethics committee, and all participants will provide written informed consent prior to inclusion

• Correlation between CT-derived plaque characteristics and haematological biomarkers [Time frame: Baseline]
• Concordance between coronary CT angiography (CCTA) and optical coherence tomography (OCT) [Time frame: Baseline (subset undergoing OCT)]
• Association between lifestyle factors and imaging-derived plaque risk features [Time frame: Baseline]
• Subgroup analyses of cardiovascular risk by demographic and clinical factors [Time frame: Baseline]
• Longitudinal changes in plaque composition in patients receiving statin therapy [Time frame: Baseline, 1 year, and annually up to 5 years]
• Feasibility and utility of CT-derived fractional flow reserve (CT-FFR) and AI-based risk models [Time frame: Baseline]