This text presents the key points from the colorectal cancer session held during the ESMO Signalling Pathways Symposium on targeting the HER/EGFR family of receptors in breast, lung, colorectal cancer (1-2 March 2013, Sitges, Spain). Prof Eric Van Cutsem, session moderator, summarises the key topics of the interactive discussion. Herewith, we present his talk.
Prof Van Cutsem:The session we had this morning was excellent. I must confess it was one of the best, one of the most dense sessions on EGFR blockades in colorectal cancer that I have attended over the last couple of years.
What we have seen is that we are making progress, that we are learning; but also an important message is that the more we learn the less we know and the more open questions there are.
Clinical evidence of EGFR inhibition in metastatic colorectal cancer
Prof Van Cutsem:During the session, I reviewed the clinical evidence of EGFR inhibition in metastatic colorectal cancer. We started from the clinic, looking at where we saw activity in the clinical studies of EGFR antibodies.
Both adjuvant studies with anti-EGFR antibodies failed. In metastatic disease, antibody tests started from later lines, moved up to second and later to first-line. We reviewed the antibody studies data and also the data on KRAS. The initial studies were in the non-rich population, but later they were carried out in a KRAS enriched population. It is now clear that KRAS mutation is a good predictive marker to predict resistance; not all KRAS wild type patients will benefit from anti-EGFR therapy. I will return to some of the important messages in this discussion...
Looking at the clinical data, one of the initial pivotal studies was the NCI Canadian-Australian study with cetuximab, as single agent, in chemo-refractory patients. The non-rich patient population showed a survival benefit, but later on, when KRAS analysis was added to that, the hazard ratio of survival improved from 0.77 to 0.55, with a doubling of median survival with the antibody alone. These are well-known data published a couple of years ago, but are still pivotal data in the field.
Antibody testing was later moved to earlier lines in combination with chemo. Second-line studies showed both antibodies, cetuximab and panitumumab, plus an irinotecan backbone clearly improved outcome of patients. Both studies showed differences that were not spectacular but that are accepted as being clinically relevant and statistically significant.
When these antibodies were later moved to first-line testing, there were some positive and some negative studies, and that makes it difficult. It is clear that we have to go in depth and we really have to select our patients better. We have some explanations in some of the negative studies, but we don’t understand everything, like in the COIN and the NORDIC studies. But the antibodies are now being tested in earlier lines and they are being integrated into some guidelines, for example, the ESMO consensus guidelines, because as demonstrated, when you combine chemo with EGFR antibody, you induce responses, meaning that we can convert more non-resectable patients to resectable patients --which is important for individual patients.
We’ve seen some disappointment in this setting. Initially, we had some preclinical data that anti-VEGF and anti-EGFR antibodies may have been synergistic or were synergistic in preclinical models and then there was the initial clinical study in mouse lines. But for reasons that we don’t understand completely --and it goes clearly beyond toxicity, probably there is an interaction between anti-VEGF and anti-EGFR antibodies. We’ve seen that when we combine oral chemo plus an anti-VEGF plus an anti-EGFR antibody, there was a deleterious effect from the combination of antibodies compared to just anti-VEGF antibody. All this clinical data, as I mentioned, is integrated in to complex treatment algorithms; without going in the details of it, there are probably in early lines some subgroups of patients, of course KRAS wild type, but probably some subgroups of further patients that will benefit. Especially, as we wrote in the guidelines, patients that we want to convert from potentially non-resectable to resectable and also patients who are very symptomatic, where we really want to achieve the response.
But it is clear from this clinical knowledge that we really have to go further, which was discussed in several talks.
Signatures of EGFR dependence: Are we there yet?
Prof Van Cutsem:Dr Gabriel Capellà from Barcelona looked into the data on signatures for EGFR dependency as potential indicators of response. He began discussing KRAS testing with some validated tests. In fact, data from the European consortium, which had a large number of patients, illustrate the impact of testing for a single gene in a non-rich population of chemo refractive patients: 24% of objective response was seen with cetuximab plus irinotecan. Enriching this population for KRAS, 36% of objective response was seen; enriching further for BRAF and PI3K, we see that response did increase slightly but there was no further increase. Today the recommendation to test for KRAS is seen in clinical guidelines and in practice, but the other gene testing is still debatable and not yet in the label of the different antibodies.
Dr Capellà furthered the story of gene signatures based on work that he did with others. He mentioned that one of the future aspects will be to determine secondary resistance and, in this regard, also liquid biopsies. We have seen two nice papers in Nature last year, wherein liquid biopsies were used to look at circulating DNA; this was also discussed by Dr Alberto Bardelli during the meeting, that this may help us to make progress and may facilitate our understanding of some of the important aspects of secondary resistance.
Dr Capellà went further into details of the EGFR mutation signatures and mentioned that there are EGFR mutations that may play a role in resistance, for instance, to cetuximab –which was also shown in the paper in Nature. But it is unlikely, because of the complexity, that there will be specific signal genes, and we must develop gene signatures. Some gene signatures are under development and Dr Capellà tried to convince us that there is some progress being made. We had an intense discussion on that –we are not there yet, but there are some attempts that may be realized in a couple of years that may provide some gene signatures that may help us, on top of KRAS testing, to see which patients really do benefit from the EGFR antibodies.
Dr Capellà concluded that we are making progress, but we are not there yet. Maybe this last conclusion that we are getting closer was a bit optimistic. I do believe it will take some time before we can come up with a validated signature that you and I will use in clinical practice, but there is really progress being made in this setting.
Predictive biomarkers of EGFR inhibition: What is there beyond KRAS?
Prof Van Cutsem:Dr Andrea Sartore-Bianchi, on behalf of Salvatore Siena, gave another talk on what’s beyond KRAS. He started off with the important message that we don’t understand completely some of the recent published data on KRAS mutant patients. There are some data from preclinical models done by Dr Alberto Bardelli and different retrospective analysis, but from large databases, regarding cetuximab but not panitumumab, indicating that patients with the G13D mutation in colorectal cancer may have similar benefit as KRAS wild type patients. It’s not in the label, it’s not in the general recommendation to treat patients like that. Some of the differences between the two antibodies were discussed but we don’t have a clear picture why we saw this data with cetuximab and not panitumumab. Different hypotheses have been brought forward regarding this important fact and perhaps the further development of gene signatures may help us in this understanding.
Dr Sartore-Bianchi then went further in details regarding a recent important paper, that examines testing for different other mutations, BRAF, NRAS, PIK3CA, PTEN and so on, in addition to KRAS; the data are consistent, it’s not new information but it’s important that in this Pyrosequencing-based study, patients with the BRAF tumour had no benefit from panitumumab, patients with NRAS tumour had no benefit from panitumumab. However, if you look for instance in the PIK3CA mutations versus wild type patients, you can see that these results were not consistent, that benefit from panitumumab was not consistent into one of the subgroups.
Dr Sartore-Bianchi detailed some of the mechanisms of EGFR resistance, including, as Dr Bardelli discussed during the meeting, what is called “the myriad of EGFR resistance”. He outlined some of the mechanisms where we may have druggable options in the future, such as HER2 amplification. There are indeed more and more data that HER2 may provide a target in patients with resistance to the EGFR antibodies. And our Italian colleagues and friends have already started a study in HER2 amplified colorectal cancer patients to see whether some of these preclinical data can be applied to make a druggable treatment target for these patients; a relatively small group of patients with colorectal cancer have a HER2 amplification, but for these patients it could be important; and this also provides an important message that may enrich our knowledge.
From the presentations of Dr Salvatore Siena and Dr Andrea Sartore-Bianchi some of the important messages, as I already stressed, involve the absence of KRAS mutation in codons 12 and 13 as a prerequisite for clinical benefit and also that genetic alterations of resistance and secondary mechanism of resistance may lead to benefits for small groups of patients where we may look into further treatment options.
On-treatment biomarkers of EGFR inhibition
Prof Van Cutsem:Dr Dirk Arnold pursued another topic in a talk on treatment biomarkers; a difficult topic with a negative attitude, we may say, that’s not so relative because patients are first treated and only after that do we find the biomarkers. But I do believe --and it’s the belief also of the community, that we have to explore in detail some of these markers because they may become increasingly important as indicators of treatment outcome.
Functional imaging with PET, early response prediction by CT scan or eventually MRI, were discussed in detail. We have seen data not yet validated, but some nice data in small groups of patients, that early response on PET as well as objective tumour response on CT scan, correlate with late clinical outcome. Of course, we don’t see this beforehand, as I mentioned, but after 4 to 8 weeks of treatment. If validated in some studies, these techniques may, in the future, help us change strategies in patients whose treatment potentially warrants change. But again, these concepts have to be validated, as does the concept of early biomarkers for treatment. Dr Arnold explained the different options of these early biomarkers, summarising what the potential clinical consequences on treatment could be: treatment may be de-escalated to avoid toxicity, treatment may be escalated or modified if early resistance correlates with poor outcome and the treatment strategy may be re-adjusted or tailored for specific drug use. So this may have profound implications, but again, prospective validation is needed and that’s going to be difficult. But we always have to be optimistic --like an oncologist always has to be, although it is going to be difficult in this setting to validate early biomarkers for treatment outcome in a prospective way.
Mechanisms of secondary resistance to EGFR inhibition: Role for EGFR inhibition re-challenge
Prof Van Cutsem:Dr Andrea Satore-Bianchi discussed in his talk on mechanisms of secondary resistance to EGFR inhibition. You have also heard during this meeting about some data on mechanisms of secondary resistance and that some of the conclusions regarding molecular selection according to KRAS data and other downstream effectors provide a step forward. I won’t say it is a big step: it is a first step, and we don’t have to be over optimistic, but it is a step forward. Nevertheless, secondary resistance eventually arises in most of the patients and we have not done very well in evaluating this mechanism of secondary resistance because of the fact that it is extremely difficult to re-biopsy patients. I’m sure we will address this later on in the debate. We have to re-biopsy patients to understand some of the mechanisms of secondary resistance. EGFR secondary mutations do occur, we have some early data on that: Nature papers were cited in this regard. Rechallenge is very controversial: some early data have been mentioned that help us to endorse the concept of only studying rechallenge. I would not advocate in clinical practice the thesis that we should rechallenge patients, but at least there is some evidence at the molecular level, preclinical data, that endorse the concept of looking into the value of rechallenging with anti-EGFR antibodies.
New opportunities for EGFR-pathway inhibition
Prof Van Cutsem:Finally, Dr Josep Tabernero detailed the different opportunities of looking into new EGFR pathways; He mentioned the different ways of looking into better drugs: possibly more efficient anti-EGFR monoclonal antibodies, monoclonal antibodies directed to other members of the EGFR family, monoclonal antibodies directed to other receptors and then combinations directed to downstream effectors and inhibitors. He also gave the clear message that we must try to develop rationally based treatment combinations for the patients; there are potentials and we now have on-going phase 1 and 2 studies using these strategies. We have to develop rational treatments by integrating molecular marker studies with translational research in order to understand better how to develop new drugs that show activity, despite the toxicity shown by some new drugs, such as skin toxicity, which is not trivial. Also we have to understand better which patients may benefit most from these new drugs.