Latest in NSCLC from WCLC 2020

Bookmark and Share
Published: 24 Feb 2021
Views: 3477
Prof Solange Peters, Dr Melissa Johnson, Prof Heather Wakelee and Dr Jay Lee

Prof Solange Peters (Lausanne University Hospital, Lausanne, Switzerland), Dr Melissa Johnson (Sarah Cannon Cancer Institute, Tennessee, USA), Prof Heather Wakelee ( Stanford university, California, USA) and Dr Jay M Lee ( UCLA Santa Monica Medical Center, California, USA) discuss the latest in NSCLC from WCLC 2020.

Initially, they discuss the treatment and management of patients with NSCLC and the important types of NSCLC, their diagnostic features and testing.

They then move on to discuss the new treatment paradigms, novel drugs and verified treatment regimens and advances announced at the meeting for patients with NSCLC.

Dr Johnson and Wakelee discuss the IMpower110: updated OS analysis of atezolizumab vs platinum-based chemotherapy as first-Line treatment in PD-L1–selected NSCLC.

Dr Lee talks about the surgical and clinical outcomes with neoadjuvant atezolizumab in resectable stage IB–IIIB NSCLC.

Prof Peters also discusses the clinical/biomarker data for neoadjuvant atezolizumab in resectable Stage IB-IIIB NSCLC with the panel.

They further talk about data from other note worthy studies as well like Codebreak 100, KEYNOTE-799, EXCLAIM and ADAURA trial.

This programme has been supported by an unrestricted educational grant from Roche.

WCLC 2020

Latest in NSCLC from WCLC 2020

Prof Solange Peters - Lausanne University Hospital, Lausanne, Switzerland
Dr Melissa Johnson - Sarah Cannon Cancer Institute, Tennessee, USA
Prof Heather Wakelee - Stanford University, California, USA
Dr Jay M Lee - UCLA Santa Monica Medical Center, California, USA

SP: Dear colleagues, it’s my great pleasure to welcome you to this discussion round table which is focussing and mainly looking at the most recent data in the field of thoracic malignancies, particularly lung cancer, presented notably at the World Lung Cancer Conference 2020. It’s an interesting set of data; we’re discussing concepts mainly around immunotherapy but also targeted therapy. I will have the pleasure to be able to open the debate, if it’s not face to face it’s got to be virtual, with my colleagues in order to together share our opinions, point of view, and maybe our divergences in how we perceive this data. My name is Solange Peters, I’m the Head of Medical Oncology and Thoracic Unit in Lausanne in Switzerland in the University Hospital. I will let my colleagues introduce themselves too in order for you to have a fuller idea of who we are and why and how we discuss this data. Heather, do you want to start?

HW: Sure, so I’m Heather Wakelee, I am the Division Chief of Medical Oncology at Stanford University, the Deputy Director of our Stanford Cancer Institute and the President Elect of the IASLC.

SP: Jay, do you want to continue?

JL: Hello, my name is Jay Lee. I’m the Chief of Thoracic Surgery at UCLA and I’m also the Surgical Director for the thoracic oncology programme.

SP: And Melissa?

MJ: Hi everyone, I’m Melissa Johnson, I’m the Director of the lung cancer research programme at Sarah Cannon in Nashville.

SP: Thanks a lot. So we’ll immediately dive into this data. I must say this World Lung was more characterised by updates or diving into datasets that we knew about superficially. The first trial I’d like to discuss with you is this update or this additional level of granularity about the IMpower110 which was one of these trials trying to make the best use of monotherapy IO and basically, atezolizumab, the anti-PD-L1. We know this data because they have been presented last year showing that atezolizumab is better than chemotherapy but, as we were used to think, restricted to high PD-L1. But remember also that Roche is using a definition of high PD-L1 which is based on SP142 on the staining of not only tumour cells but also an additional subgroup of very strongly stained immune cells, so making the picture slightly different. Roy Herbst also presented that using 22C3 or using 263 still the more than 50% PD-L1 was benefitting from atezolizumab versus chemotherapy but it was a post hoc analysis. Of course, seeing the data of KEYNOTE-024 more than 50%, it’s convenient because the data are superimposable.
Seeing KEYNOTE 042 this is questionable because in some countries, and I know in some practices it’s still sometimes done to give pembrolizumab in all positive PD-L1 patients, based on KEYNOTE  042. So IMpower110 is now investigating in World Lung 2 and 3 TC or IC, meaning not only high but also intermediate staining and showing here that you lose the OS benefit. So is it something changing your practice, Heather, meaning that you will restrict the number of patients receiving monotherapy or anyway maybe you were not doing it in the practice?

HW: That’s a great question and, to me, it’s difficult with immune therapy to put all your faith in one biomarker. However, I do look at PD-L1 using the 22C3 assay connected with pembrolizumab as something that is meaningful to me. So when I see a patient who has a 22C3 PD-L1 greater than 50% I will consider single agent pembrolizumab. I have not used pembrolizumab in patients as a single agent in under 50% because when I look at the data from 042 I wasn’t convinced that we were really finding a meaningful benefit in the patients who were not the super-high levels in their tumours. So when I look at this update from the 110 with atezolizumab I see a similar story wherefore patients where there is very high PD-L1 we could consider using single agent atezolizumab but in those with the intermediate levels, and again it’s slightly different biomarkers but same idea, it confirms to me that it’s probably not the best choice.
Now, does that mean that we should never, ever consider single agent checkpoint inhibitors for patients who aren’t the super-high PD-L1 tumour patients? Well, there probably are some who would benefit from that approach but I’m not convinced that we have the right biomarkers to really distinguish them. The concern, of course, is that in some patients where you give a checkpoint inhibitor as opposed to chemotherapy, whether it’s hyper-progression or just complete failure of any benefit, they have rapid tumour progression and then become too ill to be able to get chemotherapy. So you miss the chance to treat them at all. So I’m still very cautious about that and I tend to give chemotherapy plus immune checkpoint inhibitor for anybody who doesn’t have a very high level of PD-L1 and who also doesn’t have a very significant disease burden such that they are symptomatic from their cancer. So those are the two criteria I need to feel comfortable giving single agent checkpoint inhibitors.
There are also patients who absolutely, no matter what you say, will refuse chemotherapy so I will think about it with them but I will have a more informed discussion. So this data didn’t really change anything for me, it just confirmed my already formed biases and gave a little bit more data for that. So I’d be interested to hear if Melissa has different thoughts about it.

MJ: Yes, I agree with you, Heather. To me this was less practice changing but more practice reinforcing. That it is those patients with high levels of PD-L1, whether you’re testing it using the Dako 22C3 assay or the Ventana 263 assay or the SP142 assay used in this trial. Those patients with high, high levels are the ones that will probably benefit enough from a monotherapy PD1 or PD-L1 in this case strategy. As I listened to this data it reminded me of the KEYNOTE-042 data and when it was reported at ASCO a couple of years ago that group that was 1-49 that also had a similar hazard ratio, also not very much benefit. This intermediate group that the presentation focussed on were patients where we certainly would add chemotherapy, at least in the United States, and we’re looking for other combination strategies that might be a better fit in the future.

SP: Dr Lee, do you want to complete these immunotherapy considerations?

JL: Yes, as a thoracic surgeon I’m not making this decision between atezolizumab and pembrolizumab but one question that I would pose to my colleagues here is in the high PD-L1 expressing patients how do you decide on IO monotherapy versus the combination? Is it purely driven by the metastatic tumour burden and could you expect these patients to have a good response to monotherapy? So I’m interested in hearing from my colleagues here.

HW: That’s a great question. When we look at what’s the likelihood of having a dramatic tumour shrinkage with single agent IO, even in the high PD-L1 patients it’s only about 50%. If you look at the combinations with chemotherapy it’s beyond that. So it’s really a matter of weighing what toxicity profile can the patient tolerate and how much do we need their tumour to shrink quickly? So it’s those two combinations. So if you have someone who is relatively young but has a very high tumour burden I’m going to go with combination therapy. If it’s an older patient or someone with comorbidities where their total tumour burden isn’t as high or it’s not in an immediately threatening position I’m going to be more willing to think about single agent IO.

MJ: I agree with that, Heather. Jay, I think even at this meeting we are starting to hear about other biomarkers that might predict for patients who will not benefit enough from monotherapy pembrolizumab or atezolizumab, STK11 and KEAP among them. While that is still an exploratory analysis in many trials, we don’t understand it fully. When I see biomarkers like that on a next generation profile I also am reluctant to prescribe monotherapy.

SP: So basically the data will show you what you already were doing. It’s interesting because this is a European perspective, right, we never adopted in Europe the KEYNOTE-042 so we only were delivering monotherapy IO in more than 50%. So these data somehow comfort us in the fact that we were probably doing right, adding the component of chemotherapy in what we call the twos, the TC2, IC2, so the intermediate level. So it’s something which probably doesn’t change the practice but sometimes having data which are superimposed and confirmed you are doing rightly is a good way to go. So I think it refines this algorithm which is still confusing because there are still many options.
So, leaving the stage 4, it’s probably very important because we have a surgeon to speak about neoadjuvant strategies. Probably there is now really an emphasis, like in breast cancer, for neoadjuvant immunotherapy or immunochemotherapy because all of these data have shown major and unexpected what we call pathological response and also because we had a press release showing that complete response looks also to be better in the trial using nivolumab and chemotherapy. Also because we have been seeing this phase I/II expansion cohort using mainly atezolizumab, using also nivolumab and published in The New England using nivolumab/ipilimumab which without chemotherapy, even without chemotherapy, looks extremely promising, way better than expected.
In this meeting we were able to see some updates of this analysis of neoadjuvant atezolizumab only in the LCMC3 trial, a large phase II trial, trying to look first of all at two very important things: biomarkers, how can we potentially refine the way we select these patients for neoadjuvant IO only, and, very importantly, what is the surgical outcome. Because when you do that the question is are you losing patients from the curative intent context or strategies that you have planned, meaning are you increasing the number of non-resectable disease or are you increasing the level of complications. A very important paradigm in lung cancer.
So there were two abstracts: the surgical outcome which is a very important abstract, of course, and the biomarker research trying to find how we can learn better how to choose patients for neoadjuvant IO only. So maybe we’ll start with surgery and basically I’ll make a long story short. I’m not a surgeon but what I read from this abstract is neoadjuvant IO doesn’t look to impair our ability to perform optimal surgery, being fibrosis, being also the risk of losing patients from surgery. Dr Lee, do you agree with me that these data are more or less letting us the opportunity of doing that without too many risks or too high a level of risk? What did you think about this neoadjuvant IO and what’s your experience with these surgeries?

JL: Yes, the LCMC3 primary analysis was really the largest monotherapy with checkpoint inhibition in this space. The take home message there was that, one, we successfully met our primary endpoint of achieving major pathologic response in 21% of patients. Comparatively, surgery could be done safely with very low attrition rates. We didn’t see any major complications; patients went home in a timely fashion. Really no new safety signals. So overall it was really to address feasibility and addressing the path regression issue and safety surrounding surgery. The other measure of serious treatment related adverse events when we define as grade 3 and above, those were all very low, below 11% whether it was preoperative or postoperative. So I thought that it spoke very highly for checkpoint inhibition in this space.
There’s a lot of talk about increasing complexity of the operation from checkpoint inhibition and we really didn’t see that in LCMC3. I know that, based on Patrick Forde’s paper in The New England Journal and the companion surgical paper where there was high conversion rates from minimally invasive to open surgeries and alluding to the inherent difficulty of perilobular fibrosis and things like this. We just didn’t see that magnitude of scarring that changed the conduct of the operations. Overall it was done quite safely.

SP: How does it compare for you to the old days? Maybe in stage 3 we are giving neoadjuvant chemotherapy so does it change how the patient is at the time of surgery? Do you feel like the surgery is more or less difficult? Do you feel like the response quality is different? Because that’s what we had in mind in the past is neoadjuvant platinum something, etoposide or taxanes, so what’s your feeling about it?

JL: We did a post hoc analysis of this and it’s always up for criticism any time you pull data at the end. But we did see perilobular fibrosis in about a third of the patients. Comparatively, in my own opinion, I think it’s comparable to neoadjuvant chemotherapy patients; I didn’t think it was more difficult than that. If we look at more quantitative measures like conversion rates from minimally invasive to open surgery it was only 15%. So I think overall it is more difficult than comparable to what we would see with neoadjuvant chemotherapy but it didn’t really change the conduct of the surgery and I thought it was quite safe.

SP: And we keep in mind that it might even be better combining both. If we expect the time we can see the result of chemo-IO-neoadjuvant but for the time being we can still consider this is a promising and feasible strategy to even do chemo-IO. You would agree with that? You don’t see any reason why they shouldn’t be combined? Because it looks like the response rate and major pathological response is even better, do you think so?

JL: Absolutely. I think all of us here agree that what we have seen so far from NADIM and the SAKK study and then the Columbia University study from Dr Schuh, they all demonstrate that at least in the phase II setting the path regression rates, whether it’s MPR or pathCR or objective response rates are all much higher with the combination of chemo-IO. So we do expect those findings to be hopefully echoed in the phase III studies. But one hesitation I have is that a third of our patients, 35% of patients, are older, more than 75 years of age. They’re not all ECOG 0 and 1 and all the inherent grade 3 toxicities that, at least based on historical neoadjuvant trials, are somewhere around 50-60% that we expect grade 3 toxicity with chemotherapy. So what I’m getting at is there is still going to be a significant population where the chemo-IO may not be ideal in this population because of all the things that we are concerned with in an advanced stage when we’re giving chemo and IO together. So I think the field is still looking for an IO-IO combination, particularly with the newer checkpoint inhibitors coming out and other alternatives to chemotherapy combinations with IO.

SP: You’ve paved the way to personalised strategies. It’s interesting in the neoadjuvant, you will have these patients who can receive IO only, who can receive IO-chemotherapy, we have this press release of the CheckMate 816 showing that maybe even a benefit in complete response. It might be interesting to see that you can tailor what you do to the patient profile. But maybe your opinions, Heather and Melissa, about surgery, of course, but we are not surgeons. But personalising this treatment – what about the biomarkers? We have this second part about biomarkers, it’s 20% across all patients but how can you enrich in the ones who can receive only IO? We are not very surprised, we see some PD-L1 with some EGFR and ALK but not only. There were also some specific driver alterations or oncogene suppressors. So what’s your feeling about this biomarker assessment? Does it help you refining, customising, tailoring the treatment? Maybe Melissa to start.

MJ: I agree with you, Solange and Jay, that this is absolutely a flashlight into the future of where we’re going, how we’re going to be treating patients. The surgeons at Sarah Cannon love giving immunotherapy, they tell me that the tumours feel different than they have in the past after chemotherapy. I have had some bad experiences where patients have not tolerated chemotherapy and immunotherapy and so they haven’t gotten the opportunity to go to surgery when, of course, they were surgical candidates to begin with. So I do agree that there will be an opportunity to personalise.
My worry is that that takes time. LCMC4 is a neoadjuvant targeted therapy trial that is using biomarker analysis to assign different neoadjuvant targeted therapies for eight weeks prior to surgery. My surgeons have been lukewarm about this opportunity because it really does delay the surgery where two doses of atezolizumab did not. So I’m hopeful and looking down the line for neoadjuvant approaches but I think there’s a lot to know because we need to get the information fast in order to effect change for these patients.

SP: And Heather, what do you think about it, the biomarkers? Maybe a very high PD-L1 you can skip the chemo? And the question is do you really skip the chemo even if it was an N1 clinically documented, or even an N2 before the IO? So all these questions are open, how to personalise and what about STK11 or with EGFR? How can you handle that?
HW: My take home from the biomarker analysis for LCMC3 is it’s not straightforward and I don’t know that any of us expected it would be. So the PD-L1 analysis showed, and they were using the 22C3 assay, but in the high PD-L1, at least 50%, you still had a major pathological response of only a third, 33% or so. So this isn’t metastatic disease, this is curative where you are potentially doing something that could prevent someone going to surgery. So that’s not high enough where I would feel comfortable that just having a high PD-L1 would give me comfort in telling my surgeons, ‘Wait, I’ve got to give this person a couple of months of therapy.’ Again, this was only two doses but still, as Jay had presented, there were some patients that were upstaged while they were on IO. It was a much smaller number than were downstaged but it wasn’t zero so potentially you’re losing a chance to cure those patients so we have to be very cautious.
The STK11 data showed that patients who had mutations in STK11 in their tumour weren’t as likely to respond but was it enough to say, ‘Oh, we shouldn’t do it’? I wasn’t convinced. Then there were a lot of very complex looks at what were the immune infiltrating cells, things like CD68, a lot of different things. It’s very difficult to just look and know. So it was interesting they had some talk too about using artificial intelligence to sift through all of these different biomarkers. Because when we’re dealing with trying to predict an immune response it’s not like the era of the targeted therapies where the mutation is there or not and the targeted therapies work or not. In the neoadjuvant setting with immune therapy it’s very complex.
So that’s where chemotherapy is helpful, from my perspective, because with chemotherapy all those things don’t seem to matter as much. So you can give chemo plus immune therapy and perhaps that’s going to solve the complexities, that’s the simplistic way of looking at it. But my take home from the LCMC3 was it’s still very encouraging but I’m still very much more optimistic that the chemotherapy plus IO is where we’re going to be heading for neoadjuvant and maybe we can do IO alone for some patients but I don’t feel comfortable that we really know who those people are yet.

SP: I would agree and for this some patients that you might consider only IO the biomarker assessments were not surprising, with the KEAP1, the STK11, the PD-L1, the never smoker. So something probably we knew from advanced disease. So we shouldn’t apply so many different characteristics than we apply usually in terms of the risk or what my French colleague would call hyper-progressive disease, that I would call just the failure of the strategy. But we already anticipate what might be the failure of the monotherapy strategy.
In terms of strategy there’s one I like, and I’m biased here because I’m working a lot on that trial, is the new checkpoint, the new immune checkpoint TIGIT which basically an anti-TIGIT is to be found almost across all pharma companies for the time being. But the first data we have been seeing through the phase II CITYSCAPE data, frontline patients with positive PD-L1 combining atezolizumab and tiragolumab, an anti-TIGIT, versus atezolizumab had shown already this year at ASCO that really an amazing hazard ratio and benefit was found for tiragolumab atezolizumab versus atezolizumab but all driven by high PD-L1. Still the same very high PD-L1, more than 50% range, and the benefit was driven by this subgroup in terms of response and PFS.
At this World Lung meeting we had some additional data because everybody wanted to see TIGIT as a biomarker. Why would you look at PD-L1 or why would you only look at PD-L1 when your drug is targeting TIGIT? I remember the first answer was probably they move in parallel but still there are discrepancies here, it’s not so sure that TIGIT varies accordingly according to PD-L1 and paralleling PD-L1. I have seen everything to the contrary on that side. So is TIGIT a reliable biomarker in addition or instead of PD-L1 to guide anti-TIGIT and anti-PD-L1 treatment? Heather?

HW: It didn’t really seem to be so it was interesting. They showed some nice data where they had the correlation with TIGIT and PD-L1. When we went to ASCO and heard the data they didn’t show us the TIGIT data and the assumption was that those were linked and that’s why. But now that we see the data, if you look at where TIGIT was found there was some correlation but it was not exact. There were definitely patients with TIGIT expression who did not have high PD-L1 but the responses tracked with the PD-L1 and not the TIGIT. So it was a bit of a head scratching experience for me; I don’t think, at least from this data, at least the way they were measuring TIGIT didn’t seem to be the answer. It’s still PD-L1 and so perhaps it’s still just a marker of a more inflamed tumour environment where this drug is adding something to the PD-L1 checkpoint inhibitor.

SP: Then if we somewhat just think about it, and what about this threshold because finding a threshold for TIGIT looks like a difficult exercise. And they used this threshold which was qualifying the vast majority of patients. It was not like more than 50% PD-L1, one out of four, they used the one having 105 patients out of 135. So what do you think about these data? Is there any additional look to have or maybe an additional way to measure TIGIT or do you think TIGIT is simply not a biomarker?

MJ: I think it is an interesting combination strategy and we are hopeful that perhaps just in a PD-L1 high group of patients that this may be another chemotherapy-free option. For me, Solange, it reinforced that it is a redundant biomarker to PD-L1 and perhaps we are beginning to understand. For me, one of the take home lessons from World Lung is not so fast, PD-L1 is still important. We must still continue to check that, even as more and more options are available across a PD-L1 spectrum. We will, I think, in the future separate options into PD-L1 high that will use and recapitulate the PD-L1 pathway and then the PD-L1 low where there will be other pathways that are relevant. So, for me, that reinforced the divide.
I’ll just say that the other interesting aspect of this biomarker analysis was that it confirmed not just 22C3 was predictive but also 263. Of course, in the CITYSCAPE data PD-L1 was confirmed locally by 22C3 so it was important and, I think, showed us that, once again, those two assays, the Ventana and the Deko, run pretty close to one another.

SP: Yes, it’s always good. It’s also a way to help implementing anything in the future because you can show equivalence. Just a question to Dr Lee: we are piling up our immunotherapies – CTLA-4, anti-TIGIT, anti-PD-1, even with chemo – all hoping for this five year survival. We’ve been seeing 30% five year survival if you receive pembrolizumab and you have a high PD-L1. Do you think at some point the surgeon will be lost in the middle of immunotherapy and will lose his role in terms of long-term control? Do you really believe immunotherapy will make the difference or do you think that surgery, on the contrary, is gaining in importance because you can be more multimodal and combine strategies, explore the tumour? What’s your feeling about this amazing development of IO across agents?

JL: Yes, I personally am not threatened by the fact that immunotherapy is using into the perioperative space. I think it’s exciting and I’d like to see targeted therapies move in as we see trials opening up. So I think overall it’s great. So I’m not at all threatened by that. I think it’s going to be… So far when we look for response and check RECIST criteria for the pre-imposed imaging studies, for the most part we’re not seeing dramatic changes based on RECIST, mainly because the timeframe from the pre- and post-imaging scans are quite short and we’re in a hurry to get these patients into surgery and not lose them. So we’re in this predicament. Along those same lines we’re seeing some patients that look radiographically progressing, particularly in lymph nodes and in NEOSTAR they demonstrated this nodal immune flair issue. But when you look back historically with other immunotherapies the concept of this sarcoidal-like reaction or immune mediated lymphadenopathy exists. So the RECIST criteria to assess response prior to treatment, I’m not sure that’s all that accurate or reflective of what’s happening pathologically, particularly as we see immune infiltrates and fibrosis pathologically as a component of assessing path regression. That is not reflected accurately radiographically. So what I’m getting at is you’re left with a situation similar to like subdural cancer where you get chemoradiation and you really can’t accurately assess who are the complete path responders or the MPR patients and a significant proportion of those patients still have microscopic residual disease. So I still see surgery as a key component to this. Personally I’m excited, I think this is great to see immunotherapy and targeted therapies move into the perioperative space and overall I think it’s a very good thing. Surgery, I don’t see it going away.

SP: I agree. It probably makes, again, the surgeon closer to the medical oncologist. It just recreates these platforms and bridges that we had probably lost a little bit in the past. So I think it’s very positive for what we call the multidisciplinary oncology. I have many abstracts, I will just maybe question one of you, it’s mainly medical oncology, but on the other abstracts just making an introduction and maybe a one minute answer. There was this disappointing early stop of the trial of pembrolizumab ipilimumab versus pembrolizumab front line in metastatic non-small cell lung cancer, the KEYNOTE-598. Prematurely stopped after the reading of PFS and OS but an OS which was close to the median OS so lots of censoring down the road. I found it a little premature because we know from 227 and 9LA that if any benefit of CTLA-4 is to be seen it’s probably later on. Now it’s stopped so we probably will never know more. So to me this trial doesn’t help me understand the role of CTLA-4. In two minutes, Heather, does it discourage you to use it or does it change your opinion?

HW: It is a bit discouraging. I think we have to remember that this study was only in patients with high PD-L1 versus the 227 which, of course, had a span. If you look there was some benefit in that, even in those with higher PD-L1 but it was perhaps even more so in those with lower. So it perhaps is additive importance, that’s the simple way of thinking about it. But also I would echo what you said, that perhaps we were just looking too early and that we really needed to see how this has a longer term benefit. Because you’ve got the initial response rate for what it’s meaning but where the real promise is with the checkpoint inhibitors, especially the combinations, is with that tail. We don’t know what the tail of the curve is going to look like. So I haven’t been a big adopter of the combination, even though we do have options in the US to do that because of toxicity and because I just don’t know if we can figure out who are those people where that toxicity is worthwhile to give them that opportunity for the tail. So that’s how I look at that.

SP: Melissa, KEYNOTE-799. We all used durvalumab consolidation after radio-chemotherapy in non-resectable patients and the question is how can you optimise the synergy between radiation and IO. What should you give, of course, and can you position IO earlier in the course of radio-chemotherapy? It’s, of course, a trial which is helping us understanding more or less how we could move in this field of radiotherapy, non-resectable radio-chemotherapy and non-resectable locally advanced non-small cell lung cancer. What’s your feeling about this data? To be adopted in the clinic already in non-resectable patients or are you enthusiastic to this toxicity, efficacy?

MJ: This was an interesting trial. There were two cohorts that were not compared to each other. One cohort was carboplatin and paclitaxel with pembrolizumab and the other was platinum pemetrexed and pembrolizumab. So there were a couple of different questions that the trial asked in addition to can you add the pembrolizumab up front when you’re doing the chemoradiation. To me, it’s a small study, non-randomised. Interesting, and I think it shows that, and there have been other trials to show this, that adding the immunotherapy to the chemoradiation portion does not add toxicity per se. I was interested in the data; it will not change what I do in the short term. I’m not convinced that we fully understand the toxicity of all of these modalities given together and I have many patients who are smokers, who have difficulty with tolerating the chemoradiation and then the addition of the immunotherapy out back is enough. To add it all together would cause more inflammation that would make it difficult to know whether to treat or stop and give steroids. So, I thought it was interesting, it doesn’t change what I do and I’ll look for a larger experience going forward.

SP: Which will study the same concept in a randomised fashion, having the concurrent concomitant radiotherapy, chemotherapy IO. Some trials even consolidating stronger with the IO later on. So we’ll look at all these data, I hope, in the next two years. Two other abstracts I’d like to cover. One is important, to me a little disappointing abstract, is the difficult mutation exon 20 EGFR mutations where these insertions are difficult to target. We know they are drivers of tumours, they are driving a phenotype of a never smoker non-small cell lung cancer. However, it’s not like exon 19 or point mutation in exon 21, it’s difficult to address. We had these big hopes around various compounds, I will not quote all of them but the last one was the TAK-788, the mobocertinib, which was showing initially impressive results, leading this compound into a comparison in front line versus chemotherapy, giving some hopes. We had an update at World Lung. Still the phase III is ongoing against chemotherapy but what about these results, Melissa? Do you still believe it’s better than chemotherapy front line?

MJ: It is a really interesting, difficult group of patients to treat. So I thought the interesting thing about this abstract of mobocertinib, previously known as TAK-788, was that it was in a patient group who had received prior platinum and prior TKI. Still we saw a response rate that had come down quite a bit to 23% overall. The most notable thing about the abstract that I will take away was the diarrhoea incidence of 90% in that group of patients. Despite that, progression free survival 7 months, I think, so a durable response for those patients who could continue to take it. So I think it’s important and I’ll continue to follow this agent because obviously the chemotherapy and the TKIs don’t work for these patients.
There was another abstract, though, just because you asked, that I was also interested to see, the amivantamab bispecific targeting EGFR and MET in the same group of patients, EGFR exon 20 insertions, with a response rate of 40%. A newer set of data, so that’s another one that I will follow. I think we’re making progress but not there yet.

SP: And as a reminder, this compound is now also in phase III but combined with chemotherapy versus chemotherapy front line while the mobocertinib is tested at the time as monotherapy. It might be a little difficult to combine with chemotherapy because of the GI. So different concepts but a real difficulty in having new front line strategies, exon 20. It’s a difficult disease to treat.
So the last abstract I would like to cover, and I will go to Dr Lee because it questions the paradigm of working together, it’s EGFR mutated patients and, you would agree with me, amazing data for adjuvant osimertinib in the ADAURA trial. Something as hazard ratio never seen, 0.2 something. This update of the data was first looking at how adjuvant chemotherapy was delivered when you offer such a strategy – do you change the habitude of delivering chemotherapy mainly in stage 2 and 3, not in stage 1. So this was the first question and the second question is do you still really have to give chemotherapy? It’s probably a little abusive question according to the trial design but that’s something which basically the data presented might question in the spirit of the community. So what do you think? Is it time to give surgery followed by osimertinib, skipping the chemotherapy not useful? Or, Dr Lee, would you tell your oncologist that you like the traditional way of chemotherapy is to be given for N1 disease and so on and so forth. So what is your tumour board opinion about it?

JL: Yes, I understand why trial design, you have to incorporate chemotherapy in the adjuvant setting because it’s standard of care. So if you omit that then you’re going to have sites that are going to say that’s just not ethically acceptable. So I see how the adjuvant chemotherapy still should fit in as standard of care. But when you look at the ADAURA data in a stage dependent fashion you see the attrition of patients that actually didn’t get chemotherapy. It’s somewhere 30-40%, depending on which stage you look at. If you look at meta-analyses, there’s one from the Ontario Clinical Trials Registry, there’s one in JAMA Oncology for the North American US data, of when do patients receive adjuvant chemotherapy a significant number, 30%, 40% are getting it outwards of 8-10 weeks out from surgery and these are based on clinical trial data. So I think there’s a natural delay post-surgery, even though we’re in the era of minimally invasive surgery with VATS and robot, there’s still an inherent recovery period for patients to go on to adjuvant chemotherapy.
What I’m hoping is that this data from ADAURA may accelerate that timeline, that patients… that overall the medical oncology community and surgeons would be more proponent to starting the adjuvant treatment earlier when you have a targeted agent which has a much better toxicity profile. So I’m hoping that that timeline is earlier starting adjuvant therapy. Whether that’s more impactful on overall survival at the end is unclear. What is the cut-off? It’s not at all clear. But I think it’s just a better tolerable drug and it’s really ideal. So I think that’s great.
The other thing surrounding ADAURA is really PFS sufficient? For me, when you look at all the perioperative chemotherapy data, the OS benefit is only 5% whether you give it before or after surgery. So this is such an impressive PFS improvement that, to me, it’s good enough and, to me, it’s practice changing. I also think there’s value added that if you delay recurrences, and we didn’t talk about the CNS recurrence being significantly reduced with adjuvant OC, I think there’s definitely value added if the only thing we did was to delay the recurrences by preventing CNS metastases, bony metastases. These things all impact quality of life and I think that with a well tolerable drug the PFS data is very impressive and the recurrence profile is consistent with that, that we are making a difference.

SP: Thanks a lot. I agree with the difficulty in mitigating or being conservative facing such data. I mean, it’s completely changing the fate of these patients, including the CNS. The component of chemotherapy will remain a question mark but I agree with you, it has its own limitations and we know it’s been a while. I like your very practical point of view, being also working with these patients before and seeing them after what we do.
Maybe the last abstract, which was probably paradigm changing is the idea that KRAS mutations found in many diseases and particularly in lung cancer is not druggable. So maybe not all KRAS mutations are druggable but the G12C looks like being druggable, thanks to more than one compound being in development. But at World Lung we saw the new dataset, so updated data, about sotorasib, the compound from Amgen targeting KRASG12C. Reporting a response rate which was above 35% in this patient population which I wouldn’t call non-druggable. So I’d like to have the comment of our colleague and surgeon because we’re speaking neoadjuvant, adjuvant. Do you think this compound and this strategy is promising and maybe you would like to see it in early diseae?

JL: Thanks for bringing that up, Solange. I thought the Codebreak 100 study was really impressive. It’s a space where about 13% of patients are going to have a KRAS mutation and it’s really an area of unmet need. I thought that the sotorasib looked really promising and the data was impressive. I personally would like to see it move into the perioperative space, drugs like this that show a positive signal would be ideal in the neoadjuvant adjuvant space as well. So I would very much be in favour of that, along with other targeted therapies that are opening up in the perioperative space as well. So I think this is really exciting news.

SP: Thanks a lot. I see we’ve covered a lot of abstracts, there were many others, including KRAS including other drugs, but for the sake of time we will let you go through this wonderful virtual meeting that now you can see streaming these slides. I think you, colleagues, for your opinions about these abstracts and I hope it will be useful for everyone listening to us tonight. So thanks a lot and just go to the IASLC World Lung Congress website, you can get access to the presentations as well as the slides. It’s very easy to use so if you missed the data just go back to that. Thanks to all of you and have a nice day.