WCLC 2022 | LungIMPACT: impact of immediate AI enabled patient triage to chest CT on the lung cancer pathway

So the ePoster on LungIMPACT is describing a study that we’re doing on testing a already approved AI solution for reporting chest x-rays. I’m particularly interested in the whole field of AI as a way of reducing the burden on the service and also making the pathway much more rapid. I’m also quite concerned to make sure that these AI solutions do what they were intended to do. So this study is primarily aimed at seeing what the impact of an immediate AI-driven flag of an abnormal chest x-ray is on the lung cancer pathway. Now that may sound a bit sort of self-explanatory, but actually what we’ve seen in a similar study that we ran called the radioX study, where we looked at immediate radiographer reporting is we found that the time from chest x-ray to diagnosis was hard, from 60 days to 30 days due to logistic change in our NHS services. So we’re looking at the same thing this time, though, with an AI-enabled process.

So the study runs with the AI giving you an immediate bit of information and flagging it up to the radiographer who can then look at the chest x-ray and potentially book a CT scan and accelerate the pathway, versus the standard process where the chest x-ray is acquired, the radiographer may or may not look at it, and it’s then sent for reporting. Once the reporting process is being done, then the reporter will have access to the AI as well. So both arms of the study are getting access to AI. It’s just the timing of that AI. I have to say, I anticipate that this will have a similar effect to the radioX study. It’ll probably accelerate the pathway, and pathway speed in lung cancer patients is really, really important. People deteriorate really quickly, and they’re often not eligible for treatment if they wait too long. So that’s the purpose of the study. So it’s going to run over the next year, and hopefully, we’ll have some results in about 18 months.