RecruitingObservational

Evaluating artificial intelligence for detecting diseases on medical images and diagnostic tests

This study is checking how well artificial intelligence (AI) can spot diseases and problems on medical scans like X-rays and MRI, and other tests like ECGs. Although many AI tools are approved for use, we don't know much about how well they work with real NHS patient data. This research will use anonymised (identity-protected) medical images and test results from NHS hospitals. Experts will first review these to confirm diagnoses, then the AI will be tested against these expert opinions. This helps understand if AI is accurate and useful for doctors, potentially making diagnoses faster and improving NHS services. Patients are not directly involved, as it uses old, anonymous data.

At a glance

Status

Recruiting

Sponsor

Oxford University Hospitals NHS Trust

Enrolment target

10,000

Start

01 Mar 2026

Estimated completion

31 Dec 2028

Evaluation of artificial intelligence algorithms for detecting various pathologies across multiple diagnostic modalities including medical imaging (X-ray, CT, MRI, ultrasound) and other diagnostic tools (e.g., electrocardiograms). Specific conditions will be defined at the sub-study level.

What is this study about?

This research project is looking at how Artificial Intelligence (AI) can help doctors find diseases or problems when looking at medical scans and other diagnostic tests. You might have heard about AI being used more and more in different areas, and it's starting to be used in hospitals too. While many of these AI tools have been given the green light to be used safely, we need to know more about how well they actually work when faced with the kind of data we see every day in the NHS.

The main goal of this particular study is to evaluate and compare different AI programmes designed to spot abnormalities on various medical images, such as X-rays, CT scans, and MRI scans, as well as other tests like electrocardiograms (ECGs). By testing these AI tools against real patient data from NHS hospitals, we can figure out if they correctly identify problems, correctly say when there are no problems, and ultimately if they could be genuinely helpful for doctors in their daily work. The NHS needs really strong proof before it starts using new AI tools widely, and this study aims to provide that proof, especially for how these tools perform with UK patient data.

It's important to know that this study doesn't involve patients directly. Instead, it's a 'data-only' study, meaning researchers will look at medical information, scans, and test results that have already been collected as part of your routine care at NHS hospitals. All this information is fully anonymised – meaning all details that could identify you are removed – to protect your privacy. This research will help the NHS make informed decisions about whether to adopt these new AI technologies, potentially leading to faster and more accurate diagnoses for everyone in the future.

Key takeaways

This study evaluates AI's ability to detect diseases on medical images and tests.
It uses only anonymised, existing patient data from NHS hospitals.
No patients are directly involved; no extra appointments or tests are needed.
The goal is to provide evidence for safe and effective use of AI in the NHS.
Potential benefits include faster and more accurate diagnoses and reduced waiting times.
Data privacy is strictly protected under UK laws.

Who may be eligible?

This study doesn't involve people directly, so there are no rules about who can or cannot take part in terms of age or health conditions.

Instead, the study looks at medical scans and test results that have already been collected from NHS hospitals as part of routine care. Data can be included if it can be made fully anonymous (so no one can tell who it belongs to) and if it meets certain quality standards for the AI to analyse.

Your data would only be excluded if it couldn't be made anonymous without losing important information, or if you have previously opted out of sharing your health data for research through the NHS National Data Opt-Out programme.

Could this study suit you?

Answer these quick questions to see if you may be eligible. This is a guide only — the research team makes the final call.

I have had medical scans or tests at an NHS hospital.
I have not opted out of sharing my NHS data for research (via National Data Opt-Out).
I understand my personal details will be fully anonymised.
I am comfortable with my past, anonymised medical data being used for research purposes.

Answer every question to see your result.

What does participation involve?

Taking part in this study means that anonymised medical scans and test results from your past routine NHS care might be used. You won't have to do anything extra, attend any appointments, or undergo any new procedures or tests.

Here's how it works: First, researchers will collect old medical images or test results from NHS hospitals. These will all be made anonymous, so all your personal details are removed. Then, expert doctors will independently review these anonymised results to decide the correct diagnosis for each. This 'expert opinion' then becomes the standard check. After that, the AI software will look at the same anonymised images and tests. Finally, the researchers will compare what the AI found with what the expert doctors said, to see how accurate the AI is. No patient contact, extra scans, or clinical procedures are involved at any stage. This is purely about looking at existing, anonymous data.

Potential risks and benefits

The potential benefits of this research are significant: if the AI tools prove accurate and reliable, they could help doctors make diagnoses faster and more precisely. This might lead to shorter waiting times for patients and could ease the workload on NHS staff. If successful, this work could also help other healthcare systems worldwide use AI safely. There are no direct risks to patients involved in this study because it uses only anonymised, historical data. No one participates directly or undergoes any procedures. All data is handled extremely carefully and is protected by strict UK data protection laws (GDPR). The study is designed to identify any weaknesses in the AI before it's ever used in patient care.

Locations (1)

Some site locations are approximate. We're improving this — please verify with the trial team before travelling.

Oxford University Hospitals
Unverified
Oxford, England

Common questions

Will my personal data be used in this study?

No, all data used in this study is completely anonymised. This means all information that could identify you is removed before it's used.

Do I need to do anything to participate?

No, you don't need to do anything. This study uses old, anonymised records and does not involve any direct patient participation.

What if I don't want my data used for research?

If you have opted out of sharing your NHS data for research through the National Data Opt-Out programme, your data will not be included.

What kind of tests or scans are being looked at?

The study looks at various medical images like X-rays, CT scans, MRI scans, and other diagnostic tests such as electrocardiograms (ECGs).

Who is funding this research?

The study is funded by Oxford Clinical Artificial Intelligence Research (OxCAIR) in the UK.

How to find out more

Abdala T Espinosa Morgado

abdala.espinosa@ouh.nhs.uk +44 (0)1865 221499 Official trial record

Always speak to your GP or specialist before deciding to take part in a study.