The Value of Technology Interventions and Earlier Diagnosis of Symptomatic Lung Cancer
- Dr. Matt Evison, Manchester University Hospital NHS Trust
Dr. Matt Evison is a chest physician in Manchester and the Cancer Alliance’s lead clinician. He is responsible for the cancer pathway in Greater Manchester. The blog discusses the need for earlier diagnosis, which is the top priority for anyone trying to deliver lung cancer services. In addition, he covers the successful trial ongoing at his trust with Qure.ai’s AI-enabled Chest X-Ray solution — qXR. The presentation was originally done at UKIO 2022
We talk about the pandemic a lot, but we’ve been amid a lung cancer pandemic for a long time. It is the single most significant cause of preventable death and the single biggest cause of cancer deaths. Lung Cancer is responsible for nearly 2 million deaths worldwide every year. That is about 20% of all cancer deaths. So when the NHS long-term plan tells us that we have a responsibility to get to the point where we diagnose three-quarters of all cancers at stages one and two, the lung is the biggest challenge by far to try and achieve that challenge. And in Greater Manchester, it’s a significant problem. Greater Manchester has a population of around 3.2 million. And we have between two and a half to 3000 lung cancer cases a year. There are high areas of deprivation, which go linked with tobacco dependency. Unfortunately, both tobacco dependency and lung cancer disproportionately affect the poorest members of our communities.
Public Health England counts all deaths before the age of 70 as premature deaths. Lung Cancer is the most significant contributor to that. It is the single biggest cause of heart disease, lung disease, stroke, and often many of those combined. We get constantly challenged when we speak about addressing lung cancer mortality rates. We have to focus on promoting earlier diagnosis. There’s lots of interest in lung health checks, which is brilliant. It will change things, especially in screening high-risk asymptomatic patients. But we must realize that it will only diagnose around 20% of lung cancers. The remaining 80% will still be present via the symptomatic pathway. To achieve NHS England’s standards, we must put just as much critical focus on the symptomatic path.
We know a third of people have been to primary care three or more times in the months leading up to a lung cancer diagnosis. This is where we have the opportunity to intervene. Chest X-rays often get overlooked about how critical they are. They will always be the cornerstone of lung cancer diagnosis and the symptomatic pathway. People say that if X-Rays were invented now, they would never be brought into modern daycare because of their limitations. But I will say they are here to stay, and we need to think about how best we can optimize them.
Over half of all symptomatic lung cancer cases get diagnosed through a chest X-ray. But a good proportion of people with lung cancer will also have a normal X-ray. Any delay adversely affects outcomes in lung cancer like no other pathway. It’s the most biologically aggressive solid tumor that even if there is a day’s difference, it matters on the pathway. And this is why I think it’s essential to focus on how vital chest X-rays are. Let me show you some data basis a trial done in Leeds where they did everything they could to increase the number of chest X-rays done in people with persistent chest symptoms through public awareness campaigns through a self-referral chest X-ray pathway. So a patient could walk into the hospital and have a chest X-ray without a referral. This increased the chest X-rays by about 80%. Even though it increased the volume of work, it showed the biggest stage shift in lung cancer we’ve seen outside of a screening trial. We saw a nearly 10% increase in stage one cancer diagnosis with as much reduction in Stage Four cases.
We cannot underestimate the importance of getting a symptomatic patient to seek help. But, alas, there’s a huge barrier to this and how the public perceives symptoms. Often these are suppressed as a regular cough or rely on previous experiences where it just goes away or how they interpret their risk of developing lung cancer. Most tragically, the public is bombarded with news about how overwhelmed NHS is. And the cough they have had for a few weeks is not worthy of what is happening to the NHS. So we have got the public perception of their response to symptoms that could be symptomatic of lung cancer. We don’t do enough chest X-rays. If we did more, we’d pick up more lung cancers in symptomatic patients. Furthermore, we have significant capacity issues.
But we know what needs to happen. The ‘Getting it right the first time: The National Specialty Report for Lung Cancer and the ‘National Optimal Lung Cancer Pathways’ have told us that we have to report chest X-rays in 24 hours. And if you’ve got an abnormal chest X-ray that looks like lung cancer, you need a CT within 72 hours of that chest X-ray. So, the current metrics in the crumbling system are trying to keep A&E alive at the front door while every other cancer pathway tries to get its imaging through.
How do we get symptomatic patients to seek care?
We need a multifaceted approach to address this. It isn’t going to be one thing. I think we will make incremental gains at every point in the pathway if we’re truly going to do something transformative. We at Greater Manchester see it as our responsibility to develop that strategy. And we produced a document promoting earlier diagnosis and symptomatic patients, which tries to bring in all those aspects. So, for example, a patient in the community who has had a cough for three weeks and is at risk of lung cancer, how do we get them a chest X-ray? We also need to get the X-ray reported quickly and get a CT scan that is also reported quickly and brought in front of a clinician like myself.
We have to run public awareness campaigns that overcome some of those barriers. We need a primary care education package, so we’ve got gateway C, a primary care education platform that delivers modules on earlier lung cancer diagnosis, training AHPs to request chest X-rays. The primary care community is now a very varied workforce.
We’re doing direct access pathways so patients can turn to an x-ray department and have an x-ray without a GP referral. With the x-ray done, how are we thinking about getting that reported quickly as possible? Here is where the potential for AI comes in.
We still need to think about what happens with normal chest X-rays, but that person still has lung cancer. So how do we promote referrals? And how then, from the abnormal x-ray, we move through to a CT scan and fast reporting?
So that is the collaboration we’ve developed with Qure.ai, AstraZeneca, and Greater Manchester Cancer Alliance. We are deploying an AI platform, but deploying it on a regional scale to try and address the issues we’ve got. Currently, 97% of NHS Trusts have backlogs with chest X-ray reporting, most reporting up to three to four-week behind. This leads to outsourcing to remedy those delays. The HSI base has been a significant driver (Healthcare and safety Investigation Branch). This year, they released a report about missed lung cancer diagnoses on chest X-rays, which is quite powerful. It talks about the reasons for missed lung cancer, X-rays, and the errors that could potentially be removed by automated machine learning, among others. It also advises the cancer system on improving our patient safety netting and AI validation and implementation. The safety netting of patients is essential; there is good data out there that suggests that if we encourage GPS to refer to the cancer pathway, the risk of dying from lung cancer in that area decreases, but we have to push up referrals.
Chest X-Rays are the cornerstone of symptomatic care pathways for Lung Cancer.
Qure.ai’s qXR platform — AI interpretation of chest X-rays looks for around 30 Different chest X-ray abnormalities. It has got good data behind it regarding its diagnostic accuracy. Let me pick out a few things about how these could be beneficial for our region’s patients.
If we can immediately identify X-rays that are suspicious for lung cancer, it puts them at the top of the pile for reporting. And we can move that direct access to the 72-hour CT pathway. We can add safety netting and automated reporting features, and even an automated referral system. Sometimes about 20 to 25% of patients with lung cancer will have a normal X-ray. Do not be reassured by a normal X-ray. You should not hesitate to refer the case to the cancer pathway, highlighting the remaining suspicion.
Greater Manchester is in an excellent position to trial this out as we have a single PACs platform for the region, that’s 10 Different acute care trusts. We all use the same Sectra PACs and the same imaging. So we can deploy this technology on a regional scale. And that was our first objective. Is it feasible across 10 Different sister hospital trusts in the real world to deploy this platform, and does it improve efficiency? So the vast majority of chest X-rays will be normal. And if there is a pre-populated report that simply needs validating by the reporting radiologist and has a set safety netting already embedded within it. Does that improve efficiency? Are we starting to think differently, and work differently, which helps our system in crisis? Does it reduce chest X-ray capture time to report across the entire patch? Does it improve the time from the chest X-ray report when there is a suspicion of lung cancer? Does it get someone to a CT quicker? And if we implement that safety netting, what happens with our referrals?
We want to see referrals going up. Our responsibility is to deal with the increased caseload due to the referrals while keeping the system efficient. We know it also reduces the potential for missed diagnoses. So it does address many of the above-discussed barriers to that broader strategy for earlier diagnosis of symptomatic lung cancer.
It is exciting to hear about early lung health checks at this conference, but the symptomatic pathway will still diagnose 80% of lung cancers. Chest X-rays are the bread and butter. They will still be the cornerstone. So we have to now think about optimizing the chest X-ray and improving the diagnosis of the same. But we got to think about this at every level, public awareness, primary care, education, direct access, AI, and accelerated pathways. They all have to come together if we’re going to make a difference. But the chest X-ray is the cornerstone.
We’ve got to do more in symptomatic patients. We’ve got to improve how well their report is, the speed of reporting, and patient safety netting practices. So for us, this is quite an exciting venture. We’re doing something different with something we haven’t done before. We’re doing it as a cancer system together. It’s a pilot, and we look forward to seeing its results and reporting them back through it.