WHAT IS THE NHS DOING WITH AI AND ROBOTS TO DETECT LUNG CANCER?

Context: This piece reflects how AI tools and public debate looked at the time it was written.

Introduction

NHS England has just announced a pilot programme combining AI software with robotic equipment to help diagnose lung cancer more quickly. The trial is happening at Guy's and St Thomas' Hospital in London, with plans to expand to other hospitals if it works.

This is specifically a UK National Health Service initiative. If you're reading this from outside the UK your healthcare system may be exploring similar AI-assisted cancer detection programmes, as this technology isn't unique to Britain, though the NHS trial is, as far as I can discover, one of the most advanced currently running.

This isn't AI replacing doctors but rather AI handling the pattern-spotting grunt work while doctors use better tools to reach tiny nodules deep in the lungs that would otherwise require repeat scans or risky surgery. The combination could reduce weeks of anxious waiting to a single procedure.

For people aged 55 to 74 who have ever smoked, this matters because the NHS is massively expanding lung cancer screening over the next few years. More screening means more tiny suspicious spots being found, which means more people needing quick, accurate answers about whether those spots are cancer, and that's the problem this technology is trying to solve.

What's actually happening in the trial

The pilot combines existing robotic biopsy technology with new AI scan analysis in a systematic workflow. The trial uses a two-step process where AI software first analyses lung scans to identify nodules (small lumps or spots) that look suspicious, flagging the ones most likely to be cancerous based on patterns it's been trained to recognise. This happens much faster than a human could review the same scans.

Second, doctors use a robotic catheter (a thin, flexible tube) inserted through the throat to take a biopsy directly from the flagged nodule. The robot can reach nodules as small as 6mm, about the size of a grain of rice, hidden deep in the lung where they're too risky or difficult to access using standard techniques.

The tissue sample goes to a laboratory where specialists analyse it to confirm or rule out cancer. For many patients, this could replace weeks of repeat scans and procedures with a single 30-minute biopsy, reducing both the waiting time and the need for more invasive surgery.

So far, 300 patients have been through this robotic biopsy procedure, and of those, 215 went on to receive cancer treatment, which suggests the technology is successfully identifying genuine cases rather than creating false alarms.

How the AI part actually works

The AI has been trained on trained on extensive datasets of lung scans showing what cancerous nodules typically look like, and it spots patterns in size, shape, density, and location that correlate with cancer. When it analyses your scan, it's comparing what it sees against those learned patterns and flagging anything that matches the profile of a likely cancer.

This is pattern recognition, not diagnosis. The AI isn't making medical decisions but rather doing the same job a radiologist would do when reviewing scans, just faster and more consistently. Radiologists get tired, have bad days, might miss something subtle, whereas AI doesn't, though it also doesn't understand context the way a human does.

The AI reduces the workload for doctors by filtering out obvious non-issues and highlighting the spots that genuinely need attention, which means radiologists can focus their expertise on the cases that actually matter rather than reviewing hundreds of scans where nothing's wrong.

The robotic bit explained

The robot is essentially a very precise guidance system for taking biopsies. Traditional methods involve inserting instruments through the airways and trying to navigate to the right spot, which gets harder the deeper into the lung you need to go, and small nodules can be impossible to reach, leaving patients with a choice between risky surgery or waiting to see if the nodule grows enough to become accessible.

The robotic catheter solves this by using far more precise control than human hands can manage, allowing it to bend, twist, and reach deep into the lung's branching airways to get to nodules that would otherwise be out of reach. This means more nodules can be biopsied safely, and patients get answers sooner without needing major surgery.

The doctor is still controlling the procedure, as the robot doesn't operate autonomously. Think of it like using one of those grabber tools to retrieve something that's rolled under the sofa - your hand can't fit in the gap, but the tool lets you reach in, grip precisely, and pull it out. You're controlling every movement, the tool just extends your reach into spaces your hand won't go.

Can you trust it with something this important

The technology has already been used on 300 patients with results that seem solid, given that more than two thirds of those patients received cancer treatment, which means the system is identifying genuine cases rather than sending people for unnecessary treatment or missing cancers that needed catching.

But this is still a trial, and the NHS is being cautious about rolling it out nationally precisely because they want to make sure it works reliably before using it on a larger scale. The pilot at Guy's and St Thomas' will be expanded to two more London hospitals, and if results remain good, wider adoption will follow by around 2030.

The AI isn't making diagnostic decisions but rather flagging suspicious areas, which doctors then investigate using the robotic biopsy tool. The final diagnosis still comes from specialist cancer teams reviewing the tissue samples in a lab, with multiple layers of human oversight sitting between the AI's initial scan analysis and any treatment decision.

That's appropriate because AI is good at pattern recognition but terrible at understanding context, spotting unusual cases, or knowing when standard procedures don't apply. Doctors handle those parts, and the technology supports their decision-making rather than replacing it.

Why the NHS is doing this now

Lung cancer screening was rolled out nationally in 2023, targeting people aged 55 to 74 who have ever smoked. Since 2019, more than 1.5 million people have been screened, and the programme is expanding to invite another 1.4 million people next year alone.

That screening is catching cancers earlier, which is brilliant for survival rates, but it also means finding far more tiny nodules that need investigating. Some will be cancer, many won't be, and the problem is working out which is which without putting every patient through weeks of repeat scans or invasive procedures.

The AI and robotic combination addresses that bottleneck by speeding up the initial analysis, improving access to hard-to-reach nodules, and reducing the time patients spend in anxious uncertainty waiting for answers. Officials estimate the expanded screening could diagnose up to 50,000 cancers by 2035, with at least 23,000 caught at earlier, more treatable stages.

Lung cancer is one of the biggest killers in the UK, with around 49,000 diagnoses and 33,000 deaths each year, and it contributes significantly to health inequalities between richer and poorer areas. Earlier diagnosis makes an enormous difference to survival rates, so technology that makes early diagnosis faster and more accessible matters.

The honest assessment

This is a sensible use of AI and robotics in healthcare, where the technology handles specific tasks that benefit from speed and precision while doctors maintain control over diagnosis and treatment decisions. It's not AI replacing human judgment but AI helping humans do their jobs better.

The trial results so far look promising, as the technology has been used on real patients with good outcomes, and the NHS is being appropriately cautious about wider rollout. Expanding the pilot to more hospitals will provide better data about whether it works consistently across different settings and patient groups.

The timeframe is realistic since the NHS isn't rushing to deploy this everywhere immediately but rather testing it properly, expanding gradually, and aiming for wider availability around 2030 if results continue to look solid. That's the right approach for technology that affects cancer diagnosis.

For patients, the benefits are clear: faster answers, fewer repeat procedures, less invasive testing, and better access to hard-to-reach nodules that might otherwise go undiagnosed until they're more advanced. The weeks of anxious waiting could be reduced to a single procedure with results coming back quickly.

The technology doesn't remove doctors from the process but gives them better tools and faster information, which lets them focus their expertise where it's most needed. That's exactly how medical technology and AI should work together, with the AI spotting patterns, the robot reaching difficult areas, and doctors making the final call based on proper analysis of actual tissue samples.

If you're in the age group being invited for lung cancer screening, this trial is good news because it means the NHS is investing in technology that should make diagnosis faster, more accurate, and less stressful. Whether you'll personally benefit from it depends on where you live and whether any suspicious nodules get found on your scans, but the principle is sound and the early results are encouraging.

Browse all topics → Index