European Multidisciplinary Conference in Thoracic Oncology (EMCTO 2011) 24—26th February 2011, Lugano

Role of predictive markers in personalised cancer treatment

Professor Ken O’Byrne (St James’s Hospital, Dublin, Ireland)

More now from ecancer.tv with Professor Ken O’Byrne from Dublin. Ken, you’re looking at molecular diagnosis and specifically the Lugano consensus; what would you say is the Lugano consensus on molecular diagnosis and its importance?

I think it’s important to go back a little bit in that process. The Lugano consensus re-emphasised the importance of the multidisciplinary team in optimising the care for the patient and the area that I looked at with my colleagues, including Fiona Blackhall and Rafael Rosell, is the specific role of pathology in determining what the best treatment might be for a patient and then, within that, molecular characteristics of the tumour that may allow us to personalise medicine.

It would be great if, in the future, you could use molecular diagnosis to really base the core of all personalised medicine; do you think that is going to be possible?

I do and I think there are some steps already that we’ve made to make that happen. So let’s start off again on the more simplistic basis. We, for many years, ignored the importance of pathology, the histology; we thought that all non-small cell lung cancers were the same. Now we realise, in fact, that the treatment we might give to an adenocarcinoma or non-squamous cancer, just simply chemotherapy, is somewhat different from the treatment we might favour giving to someone with a squamous cell cancer. One of the problems that we’ve had is that a lot of the biopsies that we get from patients, particularly with advanced disease, are very small so we have to try, insofar as we can, to subtype those tumours. Recent evidence has shown us that a marker called TTF1 and CK7 are good markers and immunostain are good markers for adenocarcinoma. Markers such as CK5/6 and p63 indicate that a tumour, which might be difficult to tell just looking under the microscope, would favour a squamous cell cancer. That’s quite useful, so that means that we can then, in more patients than we would have in the past, give a specific chemotherapy for adenose and a specific chemotherapy for squamous.

Now that’s chemotherapy, does the same thing apply to molecular therapy?

Exactly. Now the molecular therapies that we would use, some of them can be given in a broad way so, for example, avastin, bevacizumab which targets VEGF and that’s been shown to be of benefit for patients who have non-squamous cancers. Those agents potentially are dangerous in squamous cell in that we have significant side effects and we might use something else in squamous cell. So, although they’re not licensed yet, the drug cetuximab looks like it might be interesting in squamous cell cancer as an additional treatment, it targets EGFR. But within both squamous cell to a lesser extent but adenocarcinoma and the non-squamous tumours to a larger extent there is a sub-population of patients who have mutations of the epidermal growth factor receptor. Ideally we’d like to get that status at the beginning of the patient journey and there are two big reasons for that. The first reason is the trial data that shows patients who get EGFR targeted therapies who have got mutations have a much higher response rate compared to chemo, and have a better quality of life for the time that they’re on the targeted therapy than they are with the chemotherapy and they get prolonged survival, which I think is very important. That conglomerate would tend to help us to use the EGFR targeted therapy in the first line setting.

What sort of time sequence are we talking about? A patient presents, when do you start doing these molecular tests?

As soon as we establish the diagnosis, so it’s really important that the small pieces of tissue that you get from these patients are used well. What we would tend to do is look to the H&E stain to look down, and if it’s clearly an adenocarcinoma and we’re happy it’s a lung cancer then that’s potentially enough and then the rest of the tissue then can be used for your molecular testing. In the setting that you have uncertainty, then you just do a few small sections with the markers I told you, the TTF1, the CK5/6 etc, and then it’s after that point that we then tend to do the molecular tests. What’s interesting is that we’re not just looking at EGFR mutations any more, we’ve got EML4-ALK translocations and there’s good evidence that if you target those mutations in patients with a drug called crizotinib that you may get a higher response rate.

We also have evidence that HER2 mutations, which are a small proportion of patients, they may benefit from HER2 targeted therapies and perhaps Herceptin, perhaps lapatinib, perhaps one of the newer agents that’s coming through, such as BIBW2992, which is an irreversible binder to EGFR but also binds to HER2. So this is the way we’re moving forwards, so eventually we may be able to actually personalise targeted therapies to subsets of patients. And then, obviously, there will be a cohort of patients for whom there isn’t a target and they will be maybe preferentially started on chemotherapy.

We’re moving forwards but have we got there? Is the decision-making process clear-cut yet?

The decision-making process is becoming more clear-cut and by that I mean compared to, for example, when I started off over twenty years ago looking after patients, we were still unsure about even the value of chemotherapy. Then we had a meta-analysis in 1995 which suggested that cisplatinum patients with advanced disease added six weeks survival but with a lot of toxicity. Then we got data showing that chemotherapy versus best supportive care actually not only had survival benefits but improved symptoms, quality of life. Then we began to look at second line chemotherapy, only in the last decade, showing an improvement in survival; then we had the EGFR tyrosine kinase inhibitor erlotinib coming in in second line, showing a benefit. Then we’ve moved on to the histologies, again showing small benefits, small gains, and then we’ve got maintenance therapies, both with pemetrexed and nilotinib in particular, targeted and non-targeted, but also things like gemcitabine etc and we have a benefit. So we’re moving forward slowly but surely and we’re getting incremental benefits in terms of survival for the patient. Those benefits in survival are linked on to quality of life and symptom control benefits.

Now which of the markers are strongly linked with prognosis and predicting the course of the disease so far? Which have already made it, so to speak?

I just want to distinguish between prognosis and prediction. So prognosis is a marker that tells you that the patients will do well or badly, irrespective of any treatment you give. The predictive marker tells you that that patient is likely to benefit from a treatment or likely to be resistant to a treatment. So it’s really the predictive marker, the nailed predictive marker is the EGFR mutation.

EGFR, what about the others waiting in the side-lines?

EML4-ALK is, I think, the next big one that’s going to come through. The early data showed that those patients have a very high response rate to crizotinib. The other targets that are coming through, I mentioned HER2 mutations, MetMAb. So the MetMAb is an antibody and is also a TKI that switches off cMET, so I think that’s going to be a big target that comes through. Then in another group of patients we’ve got PARP inhibition, I think that looks very exciting. The PARP inhibitors may work, again, in a subset of patients that have a characteristic called BRCAness, they have deficiencies in DNA repair; we think that’s going to be very, very important.

What about good old standard chemotherapy? Are there any of your markers which can help you choose which agent to use?

There are potentially markers but none of them have yet been proven in randomised controlled trials. So probably one of the markers that has been most talked about is ERCC1, this is a DNA repair enzyme that if it’s overexpressed appears to give resistance to cisplatinum based chemotherapies, cisplatin based chemotherapy in particular. But again we don’t have randomised controlled trial data, prospective data, to really nail that one. In chemotherapy it’s a little bit behind, even though the knowledge base was a bit ahead of the targeted therapies, it’s a little bit behind where the targeted therapies are in redefining who does and does not get a benefit from specific therapies. But there a lot of trials on-going in this area and we will get prospective data over the next few years.

And with the predictive markers, how much do you predict they will improve therapy if you get them all right?

Let’s take EGFR, a very simple one. The response rate to chemotherapy for patients with squamous or adenocarcinomas is between 20-30%, that’s the objective shrink rate, and by that I mean that the tumour shrinks by more than 50% in those patients. In about another 40-50% of patients the disease is stable and the next 30-40% of the patients progress with chemotherapy. With the EGFR targeted therapy the response rate goes up to between 70-80%. So that’s a big difference.

A massive difference.

We also know that that subset of patients, they probably respond a little bit better to chemotherapy, maybe up to about 40% with chemotherapy. So that’s a big, big step forward and with the crizotinib therapy I mentioned, with EML4-ALK, the response rate again looks like it’s about 70%. So these targeted therapies given to patients who have got these mutations work very, very well, they’re relatively free of side effects, free of the chemotherapy side effects like nausea, vomiting. So I think this is going to be the future, this is where we’re going with the personalised medicine approach.

So you’re predicting potentially a sea-change in the results of therapy for non-small cell lung cancer?

Absolutely. I think we’re going to see a… I’m very, very positive now about the future. It was interesting because I was involved in a lot of the early trials with targeted therapies, with chemotherapy, that were all negative and we were all saying “God, is a targeted agent actually going to work?” Part of the reason they were negative is because we weren’t quite sure that we were hitting the target, in other words we weren’t sure if the tumour actually had the target or not. Now with molecular diagnostics coming through, we know, we can tell in many cases that the target is present and that group of patients does very well.

Now you’ve been given the difficult job of saying what is the consensus from Lugano on pathology and molecular testing. Clearly we don’t yet have all the data, so what’s your version for the take home message for what the consensus is on the knowledge so far?

The two big points that we would make are that it’s really important to try and distinguish between squamous cell carcinomas and non-squamous, because that affects the chemotherapy we might give. The second consensus is that we’re saying that we really should try to test for EGFR mutation status in every patient before they go on to have either chemotherapy or the targeted EGFR TKI, particularly in the non-squamous which tends to be where the bulk of the EGFR mutated patients are. The third consensus point that we’re making now is that we believe EML4-ALK will be a target and that really labs should now be gearing themselves up, molecular diagnostic labs, to begin to optimise the testing for this on the basis that we thinks it’s likely that that’s going to become an established target for therapy really within the next few months.

It all sounds very exciting. Ken, thank you for being on ecancer.tv.

A pleasure, thank you.