Chest CT Foundation Model: public study overview

Researchers at University College London are developing and evaluating a research artificial intelligence (AI) foundation model for chest computed tomography (CT) scans. The aim is to support future research into lung diseases such as lung cancer, chronic obstructive pulmonary disease (COPD), and lung fibrosis.

Lung diseases affect many people and are often diagnosed late, when treatment options may be more limited. Chest CT scans can show detailed information about the lungs, and the number of scans is increasing through routine care and screening programmes. This creates an opportunity to learn more from imaging data, while making sure that privacy and information governance safeguards are in place.

The study will use pseudonymised NHS Scotland chest CT scans from the Scottish Medical Imaging archive. Where approved, these data may be linked with routinely collected health information, such as age, sex, radiology reports, relevant medical history, prescribing information, cancer registry information, and mortality information. These linked data help researchers describe the study population, assess fairness across subgroups, and evaluate whether research model outputs relate to known clinical information.

All NHS Scotland patient-level data will remain within the National Safe Haven, a secure research environment used for approved analysis of health data. The UCL research team will not receive names, addresses, CHI numbers, or other direct identifiers. No patient-level data, CT scans, linked records, or patient-level predictions will leave the Safe Haven. Public outputs will be limited to aggregate, disclosure-checked research findings.

This project is research-only. The model will not be used to diagnose, treat, triage, or make decisions about individual patients. Any future clinical use would require separate clinical validation, information governance review, regulatory assessment, and further approval.

For more information about this study, please contact:

• Prof Joseph Jacob, Professor of Thoracic Imaging, University College London: j.jacob@ucl.ac.uk
• Niccolò McConnell, Doctoral Researcher, University College London: niccolo.mcconnell.17@ucl.ac.uk