Tracking lung cancer

Lung cancer is one of the world’s most common malignancies. Regrettably, most of the patients that doctors see tend to present late, and with relatively advanced disease. So we actually know very little about how this disease begins and therefore what signs to look out for or how to stop it. On the other hand, if we can catch it early, there’s every chance we can prevent it. And with this in mind, what Sam Janes has managed to do, at University College London, is to follow a group of patients with changes in their airways that can be considered precursors to cancer. Some of these patients did later develop tumours, but others didn’t; and by looking back at biopsies taken right at the start of the study and comparing the DNA and the chemistry of the cells - the team have begun to uncover the molecular fingerprints that signal when a cancer could be starting. Chris Smith spoke to Sam Janes...

Sam - As many people will know lung cancer is a pretty fearsome disease. It's the biggest cancer killer of men and women in the UK, and sadly 10 year survival for people that get lung cancer is only about 5 percent. My ambition is, we try and detect lung cancer at a much earlier stage when it's potentially curable, and secondly understand much more about the biology of lung cancer at the earliest stages. We think if we know more about that then we'll be able to perhaps design ways of understanding whether people need urgent treatments and indeed, perhaps we can even develop treatments in the future which stop lung cancer forming at all.

Chris - When you say whether or not people need treatment under what circumstances might someone not need treatment.

Sam - Yes. So this is quite interesting. So one of the types of lung cancer develops in the airways and that's called a squamous cell carcinoma. Squamous cell carcinoma appears to develop from what we call a precancerous lesion. What that means is that when we look at the lining of the lung the cells look abnormal but in fact they're still behaving in a fairly normal manner. Some do progress to form invasive cancer whereas some actually regress or disappear over time and we think if we understand the difference between these lesions then we can start to understand how we can stop cancer forming.

Chris - And equally, presumably there will be people whom you could diagnose with these precancerous lesions that are never gonna turn into a cancer and therefore they don't actually need aggressive treatment. They may just be safely left and watched.

Sam - That's exactly right. So around half of these lesions will actually never progress and in many cases just disappear. So if we were to give people treatments such as surgery or something like that for these lesions then that's really what we call an overtreatment. Whilst the other half, they do progress to invasive cancer. So if we knew which ones they were, then we can treat their cancer or even these precancerous lesions really early.

Chris - So how can you sort this out, this is a sort of wheat and chaff sorting exercise, how can you do that?

Sam - Yes. So we've spent 10 years or more now mapping out, what were really previously uncharted waters of the biology of these cells that we term precancerous. And what we've looked at is the genetic code. We've also looked at how much the genes are actually expressed. And then finally the methylation on the DNA. So these are little proteins on DNA which basically switch on or off whether the DNA is being read and the proteins are being produced. And what we found is really pretty stark, and that is we can really accurately predict whether this lesion will become a cancer in the future.

Chris - How did you actually do this because this is proper patient data isn't it?

Sam - Yeah. This is an incredible study and I owe a huge debt of thanks. Many of these people have actually been coming up to the hospital now for 10 years or more. And what we do is a telescope test where we look down into the lung with a special fluorescent light and that enables us to see these really early, what we call these precancerous lesions we can see them because they fluoresce slightly abnormally. We take a biopsy and from those cells we then look at their DNA, RNA, and proteins.

Chris - And then because you know the outcomes in these people having watched them for the best part of a decade, you can go back to those samples collected really early and ask; “Are there any chemical messages or changes in here that would have told me if I'd known this at Day Zero that the outcome for that patient was going to be what we've seen”?

Sam - That's exactly right. So we very much worked in reverse, where we followed our patients and now we know which of the lesions developed into a cancer and which disappeared. And then we've gone back and got those biopsies out of the freezer and then we look at the DNA and RNA and proteins to decipher which of those are abnormal in the people that eventually formed cancer.

Chris - And with what degree of accuracy can you make those predictions?

Sam - Really pretty good. So 90 percent accurate

Chris - And what have you learned about how these cancers do, or don't progress, and why they happen in the first place?

Sam - Yeah. So that's going to be the real challenge because the amount of data that we've produced and now published in our paper this week in Nature Medicine is enormous. And actually is going to take us and many groups around the world now, to dissect out which pathways and genes we think are the most important so that we can in the future perhaps develop a therapy or a treatment to target these abnormal pathways.