NG: Good afternoon. I’m Nicolas Girard, I’m a thoracic oncologist at Institut Curie in Paris and welcome to this ecancer discussion. This will be a round table focussing on real world data and diagnostics in patients with non-small cell lung cancer with EGFR exon 20 insertions. For this discussion I will have the pleasure to have several experts in the field: Dr Passaro, Dr Kerr and Dr Griesinger. Maybe you can present yourselves?
AP: Hi everybody, it’s a great pleasure for me to be here today. So I’m Antonio Passaro, I’m a medical oncologist in the Division of Thoracic Oncology in Milan.
FG: My name is Frank Griesinger, I’m a haematologist oncologist at the University of Oldenburg Pius-Hospital and I’m treating a lot of lung cancer patients. We’re doing quite a bit of molecular testing and also molecular studies.
KK: My name is Keith Kerr, I’m a pathologist with an interest in lung cancer working in Aberdeen in the UK.
NG: So today we want to discuss, based on the abstracts presented at the ELCC 2021 meeting, some of the key aspects in the epidemiology, diagnostics and the treatment of patients with EGFR exon 20 insertions. First, maybe, discussing those mutations and the differentiation of these mutations from the common EGFR mutations. Maybe Frank, you want to discuss the incidence and the heterogeneity of those exon 20 insertions?
FG: Thank you. So exon 20 insertions is one of the uncommon mutations. I think we all know the distribution that was proposed by Young [?] and this is the group 3 uncommon mutations. There’s the group 1, which can be treated most likely with first, second and also third generation TKIs; group 2 is normally a common mutation in combination with the T790M, sometimes also just a de novo T790M, which is pretty rare. Then the exon 20 insertions which is group 3 and these are about 6-8% of the EGFR mutations. We’ve actually looked in CRISP, in this large cohort, and it’s exactly 6.7% of all the EGFR mutated patients. There is a difference, and maybe Keith can dwell on that a little bit more, exactly where the mutations are; this is also reflected in one of the abstracts. So the most common ones are the insertions in the domain, the tyrosine kinase domain, and not in the helix and there seem to be differences with respect to the response to EGFR TKIs. There is also one aspect that came out at the WCLC showing that if one tests just with normal kits for EGFR mutations, let’s say like Cobas and other kits, one obviously loses up to 50% of the EGFR exon 20 insertions. We have now some targeted therapies: everybody heard about poziotinib and mobocertinib and there’s now also a study with amivantamab up and running. So these patients should be tested, they should be identified, they should be identified with sensitive methods and then also treated according to the availability of new drugs.
NG: Okay. Keith, maybe you want to discuss further the need for NGS to identify those mutations? Because there are many of these and then some complexity in identifying those mutations. This is what was reported in one of the abstracts presented during the meeting.
KK: Sure, thanks Nicolas. Just to echo what Frank said, the exon 20 insertion mutation family probably is the third commonest group of EGFR mutations after the common exon 19 deletions and then the L858R point mutations. So, as a group overall, they account potentially for quite a significant number of patients. The very wide range of numbers that are reported in the literature, anywhere from 1% of EGFR mutations up to some big studies have reported even 12% of EGFR mutations are of this type, is at least in part relating to the technology that’s used. So, again, as has been alluded to by yourselves, to some extent it’s a matter of you find what you look for. If you’re using allele specific single gene assays based on PCR you will only find the mutations that the test is designed to find. This has actually been a phenomenon that has caused problems with the reported mutation prevalences right from the start in the EGFR mutation story and, indeed, in some of the other mutations. So with an approach, however, like next generation sequencing or even Sanger sequencing, although virtually nobody is probably using Sanger sequencing these days in earnest in a clinical sense, you will find mutations in the exons that you look at. So you will find all of the mutations. So as we become interested in exon 20 insertion mutations, it’s really important to have a methodology that is going to find them. For all practical purposes now, next generation sequencing is the go-to methodology for that. We don’t properly understand whether there are differences, say, in different countries with different ethnicities around exon 20 insertions but from the data that have been reported from North America, now from Europe, from the ELCC study, and also from the Far East, there’s no obvious evidence that there is a marked difference in exon 20 insertions per se. They seem to fall into the same category as all the others.
NG: Antonio, what are the clinical characteristics of those patients, those patients with exon 20 insertions? Are these patients different from patients with common EGFR mutations? Do we have any clinical features that may be associated with exon 20?
AP: In the data that was reporting in ELCC by Christopoulos in the real world data, the clinical features were similar to the other exon mutations – exon 18, exon 21. So we have a prevalence of females and never smoker or light smoker. So we have a differentiation from the single alteration, the tyrosine kinase inhibitors, but maybe we have the same background, biologic background, for this kind of disease. This is very important to confirm the need to test all our patients, regardless of the first positivity or not of the exon 19, exon 21. So, as discussed before by Frank Griesinger, it’s important – we have a pretty big block of patients with EGFR mutations, so common, sensitive uncommon and the exon 21 insertion that’s a very different entity today. So, from my clinical point of view, moving from the diagnostic point of view of care, it’s important to understand that at the present time I think that the clinical is very, very slow compared to the diagnostic. For example, in Italy only 25-30% of non-academic centres are able to test our patients with an NGS analysis. So this is dramatic if we understand the therapeutic target, we have a diagnostic ready to identify these patients, we have a clinical feature to understand, maybe, these kinds of patients that we have not the [?? 10:05]. Please, Frank.
FG: [?? 10:08] back that up from Germany in terms of NGS and in CRISP, this registry, we do have basically specialised lung cancer centres, hospitals, comprehensive cancer centres, cancer centres and also practices in town because we have a specialty in Germany that also with oncologists that are practising in town, they are also treating patients with non-small cell lung cancer. We have a rate of now 45%, we started with 30% of patients that are tested with NGS and we are now at 45%. We know that in Austria they are close to 100% and Switzerland I think they’re at 90%. We have in our guidelines it’s stipulated please do NGS but there are still hurdles to do this in the broad spectrum and I’m not talking about small hospitals, or not specialised hospitals, there the rate is probably even lower.
NG: Still, this testing, this NGS testing, has also maybe to be viewed in the general context of molecular characterisation of non-small cell lung cancer. We have more and more alterations to look at because we have drugs even for very rare alterations – I’m thinking about KRAS G12C – but still highly frequent. But we have also the ERBB2 mutations, we have the RET fusions and so on and so on. So I guess availability of new treatment dedicated to exon 20 will actually push everyone to do a systematic NGS. But even in France where we have those academic platforms and so on, maybe less than 50% of patients actually are tested in the front line setting because of financing issues and so on. I agree with the kind of differences between small centres and larger centres. Still, this molecular characterisation may have some impact when we classify the exon 20 insertions between the near loop and the far loop with regard to the efficacy of available agents, including mobocertinib, poziotinib or amivantamab. But maybe before discussing those new treatments, Antonio, maybe you want to comment about this abstract looking at the standard treatment for those patients and especially reporting on the efficacy of immune checkpoint inhibitors which are standard of care in non-small cell lung cancer overall and without such alterations?
AP: Yes, so this is a very key point. In this particular setting we have that chemotherapy performed in the same time of fifteen years. So we have a global overall survival of less than 24 months. In this particular subgroup of molecular alterations, exon 20 insertion, we have the failure of immune checkpoint inhibitors. So we have a particular target without any very active agents. This is important to understand and to confirm when we have new data that’s coming out and also to understand how to improve the sequence of these kinds of patients in clinical practice. If we have that those patients are not responding to immunotherapy and we have the availability of clinical trials after, we have to improve the enrolment of patients with exon 20 insertion in clinical trials if we have no new drugs in the clinics. On the other hand, it’s very important to confirm and to stress that these kinds of patients should be not treated with the standard tyrosine kinase inhibitors; first, second and third generation tyrosine kinase inhibitors are not effective in this kind of subpopulation. So we have very small data that confirms limited improved activity in favour of osimertinib with a double dose but it was based on a very, very small phase II and very small overall patient population. So it’s important to confirm that the chemo, in the absence of any clinical trials or any specific data, is the standard of care for patients with exon 20 insertion. From the abstract from Christopoulos and colleagues it’s not clear if the combination of chemotherapy plus immune checkpoint inhibitor is able or not to improve the standard effectiveness of the chemotherapy alone. So data based in a very, very small group of patients. But considering the hypothetical sequence of chemo followed by tyrosine kinase inhibitors direct for an exon 20 insertion or bi-antibody, for example amivantamab, I think that we should prefer the chemotherapy alone instead of chemo plus immunotherapy but this is only in my opinion.
NG: No, I agree with you. We have some reports in the literature about the limited efficacy of immune checkpoint inhibitors in this population. These patients are still EGFR mutant non-small cell lung cancer patients so probably with limited tumour mutation burden, with an immunosuppressive microenvironment and with the limited amount of neoantigen overall. So, to me, the rationale to use immune checkpoint inhibitors in those patients remains limited and from the abstract presented during the meeting we see that immune checkpoint inhibitors alone are associated with a very low response rate. I know that this is a very limited subset of patients but this seems to confirm that immune checkpoint inhibitors should not be used in those patients. There were also some reports of hyperprogressive disease in the literature. So I agree with you that chemotherapy alone is probably the standard of care. Frank, what do you think?
FG: We also have in the common mutations the IMpower150 data so at least there it seems that immune checkpoint inhibitors have some effect. I agree completely that a single agent immune checkpoint inhibitor probably should not be used but, as far as I know, the studies that are ongoing also have as a control arm either chemo or chemo plus immune checkpoint inhibitor. So I think the call is out – we don’t know yet, we probably have to see in registries or also in the trials how these patients fare with the combination. But, just to make this point clear, immune checkpoint inhibitors as single agent probably does not work.
NG: So new treatments, new options for those patients – TKIs? So poziotinib, mobocertinib and one antibody, amivantamab, targeting MET and EGFR. Antonio, maybe you want to discuss poziotinib, mobocertinib data? These data presented at the last ASCO WCLC meeting and reporting on, actually, some very interesting response rates.
AP: Yes. So this kind of tyrosine kinase inhibitor moved from the data that we have in the standard common EGFR mutation and during the past [?? 18:42] it was limited exclusively, in my opinion, for the toxicity profile. Both the new drugs, including the tyrosine kinase inhibitors poziotinib and mobocertinib, showed very interesting data about the overall response rate surely then over chemotherapy. But both the drugs were limited for the toxicity profile, in particular poziotinib, has a very difficult [?? 19:09] with a very different step of a different reduction dose for patients that received it in the clinical trial. I think for poziotinib the very limiting factor for the clinical practice should be the toxicity. On the other hand, mobocertinib, maybe in my opinion is the best tyrosine kinase inhibitor compared to the other drugs and has a very, very interesting activity for this kind of patient. But I’m not sure that comparing them, balancing very well efficacy and toxicity, there will be room for mobocertinib in clinical practice. So the data that are coming out from the evaluation of the different kinds of region in which we have the exon 20 insertion maybe are the key to understand that using tyrosine kinase inhibitors in this particular molecular subset is not a great way for our targeting, compared to the other drugs that are a monoclonal antibody.
NG: Keith, maybe have you seen any data regarding the co-mutations associated with exon 20 insertions? Do you think that these may also impact the efficacy of any kind of targeted agent?
KK: There are some published data with respect to co-mutations occurring with exon 20 insertions. In summary, the pattern of co-mutations is more or less the same as is found with the common EGFR mutations. So in that sense the picture with exon 20 insertions appears to be no different. The question, though, that you posed is will the co-mutational picture influence outcomes and I have a feeling that probably it will. When we think about the data that we have seen with TKI treatment of common mutations and the presence or absence of a p53 mutation, there seems to be quite compelling evidence. This, I think, is a consistent biological phenomenon because it’s actually been reported with ALK and I’ve seen reports also with ROS that the co-existence of the p53 mutation makes the therapy less effective. Therefore there may be an argument in those patients of adding chemotherapy with the TKI. So why would exon 20 insertions be any different in that sense? At the moment we know no reason why they should be.
NG: I agree with you – actually those TKIs probably we have some difficulties to interpret the results in terms of efficacy because of the heterogeneity of exon 20 insertions, because of the co-mutations. So it’s not the same story as common EGFR mutations and probably here we have more variables to integrate in the model to predict the efficacy of those drugs. The second strategy to target exon 20 insertions is amivantamab, Frank, maybe you want to discuss some of the key data with amivantamab?
FG: Well we saw some data at the WCLC and actually amivantamab had the highest response rates and also the highest and longest PFS and there mobocertinib and poziotinib were much, much inferior. Also the tolerability of amivantamab seemed to be much, much better – so GI toxicity and also skin toxicity were much, much less pronounced. So I think we see now a shift from TKIs to antibodies and maybe also antibody-drug conjugates because we’ve seen that in for HER2. This is not the topic today but with this trastuzumab plus deruxtecan that has really fantastic response rates in HER2 mutated, not in over-expressed but mutated, patients. So one wonders whether amivantamab plus a payload might be even more efficacious and as far as I know there are also drugs that are being developed that have these characteristics. I think this is very interesting data and there is also a study that is testing now chemotherapy plus/minus amivantamab in the first line setting of exon 20 insertions. So this is a very interesting study and it’s a competing study with the mobocertinib study which, as far as I know, is on hold because the futility analysis has to be done and it’s going to be continued. So we will see how the TKIs in the exon 20 insertions will fare. I would like to make one point, just for the understanding of poziotinib, mobocertinib, what I understood from the exon 20 insertion is that it’s a very small pocket, it’s very difficult to fit drugs in this very small pocket and that the poziotinib has the problem that the IC50 is lower for wildtype than for the mutation. So by definition the toxicity is very high than with mobocertinib, this is the only drug that has a better affinity to the insertion mutation than to the wildtype but nevertheless it still has pretty high affinity to the wildtype so we see a lot of toxicity. So we have to see which drug approach will come out as the leader; my prediction would be that in reality it’s probably going to be antibodies.
NG: Thank you. Thank you for this very interesting discussion on this very specific subset of patients – exon 20 insertions. We have these challenges for the identification of patients with the need for using NGS to fully characterise exon 20 and not miss some of those patients. We understand that the standard of care, especially the use of immune checkpoint inhibitors, is associated with limited efficacy and that we have new options coming through clinical trials right now with very promising data with TKIs and probably even more with antibodies. So we will obviously have to follow this story in the next meetings. So thank you for your contribution to this round table and, again, thank you for following this discussion.
