New Glasgow Study to Explore Whether AI Can Support Earlier Detection of Serious Heart and Lung Disease
Published: 4 June 2026
A new real-world study led by the University of Glasgow has launched to evaluate whether AI-supported chest X-ray interpretation in the Emergency Department (ED) can support earlier identification of serious cardiopulmonary disease within routine clinical care.
<Glasgow, 4 June 2026> A new real-world study led by the University of Glasgow has launched to evaluate whether AI-supported chest X-ray interpretation in the Emergency Department (ED) can support earlier identification of serious cardiopulmonary disease within routine clinical care. SEARCH-ED will assess how AI-enabled analysis of standard imaging can help accelerate investigation, referral, and treatment for patients presenting with previously undiagnosed heart failure, COPD, and lung cancer.
The study brings together the University of Glasgow’s HealthTech Innovation & Translation Lab, healthcare AI company Harrison.ai, NHS Greater Glasgow & Clyde, West of Scotland Innovation Hub, and AstraZeneca (which provided a grant towards the independent project but had no involvement in its creation or organisation).
As part of SEARCH-ED, Harrison.ai’s chest X-ray (CXR) solution will be integrated into real-world Emergency Department workflows at the Queen Elizabeth University Hospital. Capable of detecting up to 124 clinical findings in under 20 seconds, the CXR solution is designed to support identification of both emergent and incidental abnormalities, while prioritising urgent cases for timely clinical review.
More than 7 million chest X-rays are performed annually in UK emergency departments, making them one of the most common frontline investigations for both acute and long-term conditions. By embedding AI-supported chest X-ray interpretation directly into clinical care pathways, the study will assess its potential to inform clinical decision-making and, in turn, shorten time to diagnosis, accelerate treatment initiation, and improve patient and system-level outcomes.
SEARCH-ED focuses on three common cardiopulmonary conditions—heart failure, lung cancer, and chronic obstructive pulmonary disease (COPD)—where delayed diagnosis remains widespread in the UK and internationally, and where outcomes are strongly influenced by how early intervention begins.
Heart failure, in particular, remains one of the leading causes of emergency hospital admission in the UK and is associated with poor long-term outcomes. Around 80% of patients receive their first diagnosis through unscheduled care rather than in community settings. Much of the burden on patients and the NHS is driven by delayed diagnosis and repeated hospitalisation, underlining the potential value of earlier identification to support timely treatment, improve outcomes, and reduce avoidable admissions.
Professor David Lowe, Consultant in Emergency Medicine at NHS Greater Glasgow & Clyde and Chief Investigator of the study, said: “Every week in ourEmergency Department we see patients with serious but previously undiagnosed heart and lung conditions. Earlier identification can speed access to specialist care, enable timely treatment, and ultimately improve patient outcomes.
“While there is growing international interest in AI, most evidence to date has focused on disease-specific use cases, rather than its role in supporting frontline decision-making in busy Emergency Departments. SEARCH-ED is designed to address this gap, generating the robust, real-world evidence the NHS and health systems internationally need to make confident decisions about adoption.”
Dr Mark Phillips, Chief Clinical Officer at Harrison.ai, said: "Heart failure, COPD, and lung cancer are the kinds of chronic conditions that often get missed in a busy emergency department not because clinicians aren't good at their jobs, but because the presentation is subtle, the environment is pressured, and the diagnostic window is small.
SEARCH-ED is designed to rigorously test whether earlier AI-assisted identification changes what actually matters - time to diagnosis, readmission rates, mortality, and quality of life - in a real Scottish emergency department, with real patients. Our collaboration with GGC and AstraZeneca is built on a strong shared research track record and a genuine belief that rigorous evidence is what drives meaningful change in the NHS. We are genuinely excited about what this exceptional team can achieve together.”
Tom Keith-Roach, President of AstraZeneca UK, said: “Innovation in healthcare is most powerful when it helps us identify and treat disease earlier, before it progresses and outcomes worsen for patients. That requires partnerships like SEARCH-ED, which bring together NHS, academic, industry and technology expertise to test new approaches in real-world care settings.
“By exploring how AI can support earlier recognition of serious conditions such as lung cancer, heart failure and COPD in the emergency department, this collaboration has the potential to enable urgently needed, risk-stratified case-finding approaches that deliver better outcomes for both patients and the wider health system.”
Anna Arent, Head of Oncology, AstraZeneca UK said: “Earlier identification of lung cancer is critical, particularly in the context of the NHS Scotland’s ambitions to reduce late-stage diagnoses, but what makes SEARCH-ED particularly compelling is its potential to help patients sooner across multiple serious conditions from a single chest X-ray, reflecting how people often present in real-world emergency settings. By generating evidence in lung cancer, heart failure and COPD, this approach could support earlier diagnosis, faster treatment, and ultimately better outcomes and quality of life for patients.”
While AI applications in radiology have advanced rapidly, there has been comparatively little prospective evaluation within live UK emergency care settings assessing impact on patient outcomes. SEARCH-ED has been designed to help address this gap, generating robust real-world evidence on whether AI-supported interpretation can enable earlier diagnosis, faster treatment, and more timely intervention for patients within routine NHS emergency care.
---
Patients and members of the public have helped shape the SEARCH-ED study from the outset, including reviewing study materials and advising on how patients should be contacted if the AI identifies a finding on their X-ray.
The study has been approved by an NHS Research Ethics Committee (reference 25/SW/0124) and is sponsored by NHS Greater Glasgow & Clyde, with results expected in 2027.
First published: 4 June 2026