A panel of European experts met during the first ESMO Signaling Pathways Symposium on “Targeting the HER/EGFR family in breast, lung and colorectal cancers” (1-2 March 2013, Sitges, Spain) to discuss on current and emerging issues in targeting EGFR/HER pathway, and highlighting possible shared approaches in a context of personalised cancer medicine. They also critically discuss why we see success in one disease type or setting but failures in others although they share the same underlying pathways. Furthermore, they try to understand why from the drugging perspective, one approach works in some, but not in all tumour types or patient settings. This text presents the key points from the discussion.
Dr Tim Maughan (TM), Gray Institute for Radiation Oncology and Biology, University of Oxford, Oxford, UK Panel experts:
Dr Luca Gianni (LG), Medical Oncology, San Raffaele Hospital, Scientific Institute, Milan, Italy
Dr Jean-Yves Douillard (JYD), ICO Institut de Cancerologie de l’Ouest Rene Gauducheau, St. Herblain CEDEX, France
Dr Dirk Arnold (DA), Tumour Biology Centre, Klinik for Internistische Onkologie, Freiburg, Germany
Tim Maughan: Welcome to the ESMO Signalling Pathways Round Table discussion. We’re talking about HER signalling pathway and how it’s working in the three important cancers: breast cancer, lung cancer and colorectal cancer and we have three international experts to discuss this. We’re going to first turn to Luca Gianni to talk to us about targeting the HER2 pathway in breast cancer.
Luca Gianni: We are lucky to start with a solid tumour, such as breast cancer, because we had a dissection of the molecular characteristics of the tumour very early and it became soon apparent that having an HER2 overexpression and amplification was associated with certain characteristics and there was the possibility to hit the HER2 and have therapeutic effects. And that has brought along a very successful story of very effective therapies, targeted therapies, for HER2-positive breast cancer that has basically changed the fate of these women effectively. In terms of targeted drugs and targeted therapy, precision therapy in a way, for breast cancer with HER2 overexpression was an outstanding example and continues to be.
TM: What difference has that made to cure for women with HER2-positive breast cancer?
LG: Well you know there has been a simulation that showed the application of HER2 directed therapies early on in the adjuvant phase to women carrying HER2-positive tumours cuts death by more than 50% and this is a stunning achievement because a single therapeutic intervention is capable of changing dramatically the course of the disease. Until 2005 having HER2positive breast cancer was associated with a worst prognosis than having cancer without HER2 overexpression. Nowadays it is paradoxically that doctors feel almost reassured when they have a case with HER2 overexpression because they have such fantastic drugs to hit the tumour and eventually eradicate it.
TM: There has been a great change in breast cancer; let’s talk about lung cancer.
Jean-Yves Douillard: The situation is a bit different in lung cancer, but to my understanding non-small cell lung cancer is the best example of personalised medicine application. In addition to a new pathological classification, essentially for adenocarcinoma, we now have a molecular classification and we have been able to identify multiple potential targets in terms of mutation or overexpression. Of course, our understanding of the EGFR mutations is more advanced in that they are found in roughly 15% of the patients with non-small cell lung cancer, but in addition to that we know there are ALK translocations for which we have a drug too. We also know that PI3K might be an important target and drugs are presently under evaluation. So the old concept of non-small cell, small-cell tumours has to be dissected into multiple diseases according to molecular profile. The more challenging situation for lung cancer as opposed to breast or colorectal cancer is the access to tissue; we don’t have that much tissue. The majority of patients diagnosed with non-small cell lung cancer are in the metastatic phase of the disease, so we have a small biopsy sample; this remains an issue on which we have to work to find some alternative techniques to identify those mutations in order to offer the patient the best treatment presently available. We have been making progress essentially by better patient selection based on molecular profile. That’s why we have certain groups of patients with longer survival, but in terms of a cure rate we are not as successful as our colleagues in breast cancer. We have not found any drug that works in the adjuvant situation so far.
TM: We’ll come back to discuss that in a moment. Dirk would you like to give us your opinion on situation in the filed of colorectal cancer?
Dirk Arnold: Well, just to address the last sentence of Jean-Yves, we are even less successful, unfortunately, in targeting the EGFR receptor family in colorectal cancer. We have learned in colorectal cancer that the expressionper sedoes not mean that there might be activity; that’s the lesson we learned in refractory colorectal cancer for targeting the EGFR receptor because the efficacy of the antibodies is entirely independent from whether the gene is expressed or not, and even amplification does not seem to play a major role in this. In colorectal cancer what we have learned is that the downstream signalling is important or even more important for the efficacy of those drugs because we have learned that the existence of a KRAS mutation, meaning the permanent activation of the downstream pathway, really overrides the expression or the amplification of the receptor. So therefore, the focus is on the downstream pathway rather than receptor expression. We have also learned that even this single biomarker that we use in colorectal cancer today is not very predictive in terms of efficacy; it is a negative predictive marker, meaning it identifies the patients bearing KRAS mutations which should not be treated or whether they have to be treated because the benefit is not as large as in those in the patient group that exclude the mutation. So there has been less development and the picture is less clear in colorectal cancer compared to the other cancer types.
TM: So the question arises, why has the approach of targeting this pathway in lung cancer and colorectal cancer not fulfilled the promise that we have seen in targeting HER2 in breast cancer? Why do we think we are failing in these two, we raised a bit of the question there, would you like to give a precise answer to that?
JYD: In lung cancer as soon as we have identified a biomarker with a predictive effect, we have had nice results upfront.
TM: Nice results in terms of improved survival?
JYD: Nice results in terms of survival by patient selection, progression-free survival, better response rate, better quality of life if you think of, for example, the EGFR tyrosine kinase inhibitor (TKI); crizotinib in ALK translocated patients. But we won’t cure any of these patients, they will probably relapse; resistance develops and it is important to identify what is the reason for resistance. There are multiple possibilities and, as I said before, it would be nice to rebiopsy the patient to identify the resistance mechanisms and try to use the appropriate drug.
TM: So we have different types of molecular target here: Gene amplification, mutation, or what is the target in colorectal cancer for the EGFR. Is the type of the molecular target actually important in determining how successful we are?
LG: I think more important here is what type of molecular alteration role is seen, what the role of the molecular alteration is. In the case of breast cancer with a HER2 amplification there is a driving force associated with that and with the expression of HER2. But within the tumours that carry and express HER2 there are those that are not amplified and do not really benefit from exposure to a HER2-directed drug. So the point is: What is the alteration associated with the HER2 amplification that makes HER2 so relevant to the point where hitting HER2 gets a therapeutic benefit? And that is the element that I think is a common theme also for other diseases.
TM: Do we know the answer to that question?
LG: While we try to dissect and understand the answer it is not the problem in drugs; many of the drugs that we use have an affinity for the target that is way beyond what is needed to completely block or eliminate this target results. So we know that it is not insufficiency or inadequate exposure to the drug that plays a role; it must be something associated with the drivers of the tumours that we have to dissect.
DA: We can clearly see this in the colorectal cancer because we learnt that even if the tumour is expressing EGFR, even if the tumour has no KRAS mutation, we can see that the drug works in delaying of progression but in several weeks; meaning in refractory patients in about 6 weeks, more than 50% of those being treated are progressing at the next stage so we are delaying the tumour progression for about 6 to 8 weeks. Which may be important in the refractory situation, but this clearly tells us that this tumour is developing mechanisms against this pathway, this driving force. Therefore this pathway becomes not relevant for those patients. Very recently we have learned for example that KRAS mutations may occur in some of those patients initially wild type who are candidates for this treatment, but that again it is only the minority of those who become refractory. So this indicates that more than 50% of the patients develop other mechanisms and only the minority of them develop this one single mechanism we have identified to be responsible for resistance. So this shows the complexity of the situation.
TM: So Bert Vogelstein said in a paper, published in Nature last year; “Resistance isa fait accompli. The time for resistance to occur is simply a function of how long it takes the resistant clone to repopulate the tumour.” Do you agree, is it always the case, and if not, why not?
JYD: I would agree with that especially in lung cancer. The tumours are very heterogeneous and when we have a small biopsy we can only look at just what we have. We know that along the EGFR pathway for example there may be multiple mutations, this does not mean that a hundred percent of the cells are mutated, and those will be responsible for failure and later progression. We can act on the cells harbouring the mutation but the other clones within the tumour will progress at their own rate and this is to me the most probable reason for failure.
TM: In colorectal and lung cancers we have talked about stage IV disease so far with the use of these agents almost exclusively and we have seen the rapid emergence of resistance. In breast cancer you have used trastuzumab successfully in adjuvant therapy and that is really where the cures have come. So, is the success due to the nature of HER2 and breast cancer or is it a fact of using treatment earlier in the disease course, perhaps?
LG: It is about both, I believe. I have no doubt that emergence of resistance is a matter of time in any tumour that we know of. And it is now obvious and clear, that heterogeneity is a classic feature of any tumour that arises in the body. So the problem is to devise strategies that exploit different mechanisms to hit the tumour and that, for instance in the case of HER2 overexpressing tumours, is the reason why it is not so easy to give up on chemotherapy because heterogeneity means also that you have overexpression and amplification of the HER2 in a fraction of the cells and if you are too selective with your targeted drugs you basically loose other cells that do not carry the target. So you need to address and attack the tumour on several levels and several aspects.
The additional point you made is also very important; in breast cancer we apply very early targeted drugs against HER2 in the adjuvant and neo-adjuvant setting and this puts the drugs in a context where you also include surgery, radiation therapy and other drugs so that overall you have a multi-dimensional attack on the tumour and you may have a chance of eradicating it.
TM: So let me put that to our two experts here; what about using targeted therapy in earlier stage disease in lung and colorectal cancer. Where have we got to on that?
JYD: Well in lung cancer regarding the use of EGFR TKI: Some studies have been performed and have been terminated early because of failure, so obviously it is not working in the adjuvant setting. We don’t have the results yet of other agents namely bevacizumab or cetuximab in the adjuvant setting in non-small cell lung cancer but we may face what we have seen in colorectal cancer as well, I’m afraid.
DA: As I have already mentioned, in colorectal cancer we have a clear story of failure for monoclonal antibodies targeting the EGFR receptor; two trials have been negative. We can now speculate whether this is because of the interaction with the maybe not best chemotherapy as combination partner for this, but the standard approach of adding those drugs to standard chemotherapy clearly failed. And to be honest, we don’t exactly know why. I think there were many bets on this - that these two trials would have been positive due to the anticipated mechanism of action and I think many oncologists were surprised but the message was very clear, they were entirely negative.
TM: So that brings us on to the issue of what can we do about this? And many people will say combination therapy - combination with chemotherapy we’ve heard in breast cancer. In colorectal cancer and in lung we’ve seen combination with conventional therapy being a failure. So how do we move forward on this?
LG: In my opinion a very good way to move forward is that of attacking a valid target by using different drugs to different mechanisms and this is what we are doing for HER2 overexpressing tumours in breast cancer. We have the good luck to have many drugs against HER2 and they attack the tumour though different mechanisms and if you use, for instance, tumour attacking antibodies such as trastuzumab plus pertuzumab you have a tremendous bust of activity and an additional benefit when you put the two drugs with chemotherapy. And this is something that is changing radically; I mean we can double the rate of complete pathologically assessed complete responses in HER2-positive tumours by adding a second HER2 targeting drug. The same can be accomplished with lapatinib and trastuzumab and now there are clear indications that a drug such as TDM-1 (which is an antibody-drug conjugate consisting of the monoclonal antibody trastuzumab linked to the cytotoxic agent mertansine), or pertuzumab plus trastuzumab plus docetaxel, can achieve outstanding results as well. So that approach at least in breast cancer works I think because of the unique role of HER2 in that subset of tumours in breast cancer.
TM: OK, so those early results are translating into improved long-term survival?
LG: They are translating into improved progression-free survival and survival in metastatic breast cancer and the expectation is that the ongoing adjuvant trials with those combinations will provide evidence of an additional benefit also in terms of survival in women with early breast cancer.
TM: The evidence points very strongly to the fact that HER2 amplification in breast cancer is a driver mutation for at least a proportion of these patients and that targeting it results in improved overall survival and cures. We don’t have clear outcomes in lung or colorectal cancers. Now is this success perhaps related to the use of an antibody based treatment approach vis a vis the tyrosine kinase inhibitor, why aren’t we seeing successes in EGFR mutated non-small cell lung cancer?
JYD: That’s probably true. As often as TKIs have been combined with chemotherapy, they have not improved the outcome. We also know that these patients with EGFR mutation are good responders to chemotherapy, but it is important to use both agents not in combination but sequentially - first line, second line. It is our responsibility in academic research to design those trials that still should be done. To my understanding, the situation is a bit different with an anti-EGFR antibody; if I refer to the FLEX study comparing vinorelbine, cisplatin with or without cetuximab, this study in a rather unselected patient population, except for the EGFR expression which was even minimal, was significantly improving overall survival in stage IV disease. Additional studies looking at the intensity of the expression of the receptor using immunohistochemistry scores really showed a huge improvement in hazard ratio. Unfortunately, the drug was not registered but the data exist and confirm that, again and maybe similarly to pertuzumab and trastuzumab, monoclonal antibody added to chemotherapy is providing a further benefit which we have never seen with TKI.
TM: Well we don’t see that in colorectal cancer.
DA: Combination with other multimodal treatments is really difficult because we have learned the story that likely we have been awaiting, that the interaction of anti-EGFR treatment with chemotherapy plays an important role as some of the regimen work synergistically and others don’t for chemotherapy. So for example, we would consider the combination of oral fluoropyridines on the basis of two randomised phase III trials, to be not the best combination partner, whereas infusional regimens of 5-FU/LU with oxaliplatin/irinotecan work nicely with these drugs. So therefore the interaction with chemotherapy seems to play a role, not only for increased toxicity but also for the synergy of both. And this becomes more complicated when we add radiotherapy; we have a handful of radiotherapy, chemo plus EGFR trials in rectal cancer and we can see that the effects are even detrimental or are likely to be detrimental in terms of pathohistological response. So therefore I think we have much more caution on the interaction of chemotherapy- cell cycle, and EGFR - cell cycle, whichever works as the mechanism of action.
TM: So may I ask you, we’ve made progress in, it’s moving forward - we’ve made some progress in lung cancer but we are quite stuck in colorectal cancer. How do you see us as academic researchers and clinicians helping to move this field forward? What must we do next? Can I come around the table, would you like to start us off?
DA: My personal interest would rather be to see which patients really hang on the inhibition of this mechanism of action and this really drives the tumour for a long time. The best way to identify such patients is to molecularly characterise upfront treatment but to use this drug for example in the situation where you only have minimal residual disease, meaning after combination with chemotherapy, and you have the remaining tumours in stage IV. I think you may have even test the larger subgroup of patients who are highly responsive to that drug. This is a field more precisely identifying the molecular characteristics of the patients who have high chance of being long-time controlled by those drugs.
JYD: I fully agree with Dirk. This is what has been done in the field of lung cancer with the concept of maintenance therapy, a reduction of the tumour burden with induction chemotherapy followed by a maintenance treatment with TKI for example; trials have shown that there is a benefit in progression-free survival and on overall survival. The patients were not selected based on EGFR mutation, however.
TM: We are taking such a concept further in colorectal cancer. So, in breast cancer, what is your next step as an academic leading researcher?
LG: Our next step is that of focusing as much as possible on the very early stages of the disease and exploiting to the best the approved tools of neoadjuvant chemotherapy and adjuvant therapy in breast cancer. That is a unique opportunity. We know that we can safely perform the administration of drugs before surgery for a month without affecting the end results of the overall treatment in women with breast cancer and with the application of drugs before surgery we have a tremendous opportunity of investigating the characteristic of the tumour, not only on its own but along the course of the administration of the drugs. And that is basically giving not only thea prioricharacteristics of the tumour that are associated with the highest probability of eradication but also the opportunity of providing the possibility of investigating what residual aspects after exposure to the drugs are associated with long-term failure and lack of success. And this approach is completely changing the field. We see years in advance in the neoadjuvant setting what will become apparent in the adjuvant setting after 5 or 6 or 7 years of observation. And I think that this is an outstanding opportunity which maybe can be tested in a trial in other disease types although it is not so easy to transfer this concept from one disease to another disease.
TM: Thank you. So perhaps we’ve learned this morning that one of the most important things is that we must have a target which is actually driving the cancer: HER2 does that in breast cancer but perhaps EGFR does not do that in colorectal cancer and in some patients the mutation does act as a driver in non-small cell lung cancer. We’ve seen the importance of designing trials that can really identify which agents are effective in which patients and the neoadjuvant window studies or the molecularly selected studies in maintenance following first line therapy are good opportunity for that. We’ve had comments about the importance of the translational research with analysis of the circulating tumour DNA and sequential biopsies are an important part of this. And so it’s a challenge. Combination therapy has proven to be a mixed bag, often taking us backwards rather than taking us forwards and is a tremendous challenge to the application of these concepts to more forward, particularly towards a cure for our patients.