ESMO 2024: Highlights and analysis

Assoc Prof Bishal Gyawali - Queen's University, Kingston, Canada

Hello everyone, this is Dr Bishal Gyawali from Queen’s University, Kingston, Canada; I’m very happy to be back with you all again with the ESMO round-up video. This is our annual tradition, we do this round-up after every ASCO and ESMO and thank you for joining us again.

At this year’s ESMO one of the themes that I wanted to highlight was the perpetual lowering of the bar, lowering of the standards for what we consider to be practice changing. To give you an example, there are several trials that were presented that had only progression free survival data that was significant. It was not that overall survival was not available, overall survival data were available, they were presented and they were negative. So you have trials where you have survival data, it’s not like we don’t know what survival is going to be – we know and it is negative. In that case you would expect the conclusion of the slide or the discussant or when it is published in the journal, the journal article would conclude that overall survival did not differ. But instead what we got was no discussion about overall survival or briefly in passing just mentioning that, ‘Oh, overall survival was not significant, by the way,’ sort of presentation and focussing entirely on progression free survival and highlighting how important that was.

I’ll give you some specific examples, like the PRIMA trial of niraparib in advanced ovarian cancer. This is a trial of maintenance therapy of niraparib versus placebo and for maintenance therapy and for combination therapies I have argued time and again for different reasons why PFS is a very weak and bad surrogate endpoint. For maintenance therapies, because of the increased therapeutic burden in order to justify that and in order to justify using it earlier compared to later for only those who need it,  you need to show survival advantage. This was the final OS results from the trial. We knew even before that PFS was positive, we were all waiting for OS and OS data were available and they were presented at this year’s ESMO and OS was not significant. The hazard ratio was 1.01, OS did not improve and, in fact, if you look at the numbers the overall survival with the drug was 46.6 months and with placebo was 48.8 months. So the drug had shorter overall survival than placebo. This was not statistically significant otherwise it would be significant harm and increased mortality but nevertheless these are important numbers that need to be communicated properly. But you know what the headlines were? The headline was niraparib confirms its PFS benefit in patients and should be the standard of care. Niraparib improves or delays time to progression. So all these misleading headlines without stating the fact that survival did not improve and, if anything, it was even shorter. This should not be the standard of care. It has already been the standard of care based on PFS and approval but after seeing these survival results we should rethink that decision and the guidelines should change, in my opinion. But, instead, we just ignored that survival result and we just kept celebrating that PFS was improved.

The MARIPOSA-2 trial of second line treatment after progression on osimertinib for EGFR mutant lung cancer, and this was a trial comparing amivantamab plus chemo versus chemo alone. Again, overall survival was not significant but everybody just focussed on progression free survival. This is not even a first-line treatment, this is after progression on osimertinib. You need to improve survival; if you can’t improve survival even in second, third, fourth lines, then that drug is not good enough. If you think that crossover negates the survival effect then the appropriate conclusion would be you can use the drug later on in the therapy and you don’t need to use it up front.

The same case with the LEAP-012 trial of lenvatinib plus pembrolizumab as an addition to TACE for patients with intermediate hepatocellular cancer. Again, there was a five months’ improvement in PFS but OS did not improve. We already knew that PFS had improved; OS did not improve. This is hepatocellular cancer. Again, the conclusion, the discussion, everything focussed on five months improvement in PFS, but when are we going to mention that overall survival did not improve. For a disease like hepatocellular cancer which has a poor prognosis, the minimum we expect is overall survival. If you look at toxicities, the toxicities are horrible with this combination. There was a 72% incidence of grade 3 or higher adverse effect.

The other problem is there was no crossover. The patients who got randomised to get placebo plus TACE when they progressed they did not necessarily get immunotherapy. We already know that immunotherapy is the standard of care. So you have no crossover when you were supposed to include crossover, so in a sense it’s a poor control, and this experimental agent has such a high incidence of serious and severe adverse events it’s not easy to take. Even then we failed to improve survival and still we are cheerleading on progression free survival and we think that it’s a good treatment.

We have really, really lowered our expectations, lowered our bars. The way to claim victory is by actually improving outcomes; the way to claim victory is not by lowering expectations and lowering the bar.

One more example: the DESTINY-Breast06 trial of trastuzumab deruxtecan. It was published in The New England Journal of Medicine simultaneously as well, the overall survival results. Again, these are patients who have been pretreated, hormone receptor positive, HER2 low or ultra-low patients who have already received at least one line of treatment before but the median is two and there were patients who had received up to five lines of therapy. PFS improved by almost four months which is fine but you would want to know what overall survival results are. First I read only the abstract and I thought overall survival results were still not available because there is nowhere in the abstract that discussed about overall survival. Then I accessed the full paper and in the full paper I don’t see any overall survival graphs. Then I go to the supplement and I find overall survival mentioned in page 23 of the supplement and overall survival was not significant. So I had to go to page 23 of the supplement to find the fact that survival did not improve and to see the graphs of the overall survival. If overall survival data is available that should be in the abstract, that is what patients care. The patients care whether they are going to live longer or not. If we don’t have OS data, that’s a different story. When we have OS data, putting the graphs in the supplement, it serves no purpose. We already have data, we already have the graphs, the data on overall survival should be mentioned in the abstract and the graphs should be available with the full text of the paper. Not everyone is going to read up the supplement.

One interesting trial this time was from supportive care, that is a drug, a new drug, for cancer cachexia. We usually don’t have any drugs to treat cancer cachexia so it qualifies as an unmet need. We did have one trial back in 2015, I think, of a drug called anamorelin which was supposed to be for cancer cachexia. It did not get FDA approval but has been approved in Japan. That drug improved muscle mass but did not improve muscle function. At that time I had written and I had thought that what’s the point in having 1kg extra muscle mass if you can’t button up your shirt, if you can’t open the door? So muscle function is important. But this was a trial and it was published in New England Journal simultaneously of a new drug called ponsegromab versus placebo for cancer cachexia. It was a phase II trial, not a phase III, and randomised patients to four cohorts: 1:1:1:1 ratio. Three of those cohorts are for the drug at different doses and one was placebo.

I saw a lot of cheerleading for this drug, admittedly people are excited that they have a new drug for cancer cachexia, an area where we do not have any drugs before. But this is a phase II and the primary endpoint is body weight, it’s not even muscle mass, it’s body weight. So that body weight may not even be muscle, it could be something… body weight is composed of muscle mass, adipose tissue and so on. So body weight is not a good endpoint in my opinion, it should be muscle mass like the anamorelin trial, the ROMANA trial, plus muscle function. In this trial they did try to measure some ability to do physical activity based on some questionnaires from the FACT2 but we should specifically measure muscle mass and muscle function. One of the good ways to measure muscle function is actually hand grip strength. We need to know muscle function properly because, again, the increase in body weight they report as somewhere between 1.2 – 2.8kg between different doses as compared to the placebo arm but what’s the point in having a little bit more body weight of which what percentage is muscle mass, I don’t know? What’s the point of having 1.5kg or 1.8kg or 2kg extra body weight but not being able to open the door, not being able to button your shirt, not being able to stand up from the toilet? So these are things that affect quality of life and I’m happy that they did try to measure quality of life and physical activity but I also want to see specifically muscle function. But the caveat here is it’s a phase II trial so the appropriate conclusion is let’s do a phase III trial.

I also want to make a point about de-escalation of treatment and there were some nice trials that tried to look into this. But also a couple of trials in the perioperative setting where I think is an opportunity to do de-escalation trials. We need to fund more and more de-escalation trials through public funding. For example, the NIAGARA trial where they looked at the perioperative durvalumab in metastatic bladder cancer both in neoadjuvant followed by surgery and followed by the adjuvant phase. We need to ask the question is the adjuvant phase even necessary. If the adjuvant phase is not necessary and the benefit is driven mostly by the neoadjuvant phase then we are just subjecting our patients to unnecessary therapeutic burden and cost and time toxicity and all the problems that arise from treatment without any evidence of benefit.

We had OS results from the KEYNOTE-522 trial in triple negative breast cancer as well. Again, it was significant but does the adjuvant component add anything or is neoadjuvant alone enough? These are important questions to ask because we saw results from the NICHE-2 trial – one dose of ipilimumab plus two doses of nivolumab – there was 100% three-year disease free survival rate in MSI high colon cancer. We have trials for melanoma which show that neoadjuvant is enough. Even in non-small cell lung cancer we have neoadjuvant only trials. So future trials should be comparing neoadjuvant only versus a neoadjuvant plus adjuvant strategy because we need to find out whether an added adjuvant component of the treatment adds any benefit. There were some post-hoc analyses presented at the World Lung Cancer Conference about this and they concluded that adjuvant therapy probably adds something but this is not a question that can be delegated to post-hoc analyses, there should be proper trials addressing this question.

With regards to KEYNOTE-811, the trial of a regime of pembrolizumab to Herceptin plus chemotherapy in patients with metastatic HER2 positive gastric cancer, we got the overall survival results and overall survival improved by three months: 17 months without pembrolizumab versus 20 months with pembrolizumab. This was statistically significant and three months improvement. So I don’t have much complaint but I wanted to highlight how far low our expectations have gone in oncology because this three-month overall survival advantage is pretty good compared to all other trials that I have discussed so far where there was no OS benefit and there was marginal PFS benefit, toxicities, and still everybody was cheerleading. Compared to that, this is pretty good but on Twitter we saw some of the comments from non-oncologists saying that, ‘Why are you guys cheering for three months survival benefit in advanced cancer?’ which is the very right question to ask. But the reality is even that three months survival benefit in overall survival has become a very rare thing. We don’t even see any survival benefit, we see a couple of weeks of PFS benefit and drugs getting approved based on that. So three months overall survival benefit for us oncologists, we felt like it was a big deal, but it was a reality check that non-oncologists were questioning us, ‘Why are you guys even celebrating just a three month improvement in survival?’

Anyway, continuing with the gastric cancer trial, there were the results of the TOPGEAR trial presented as well as published simultaneously. It is a trial of a regime of chemoradiation pre-op to the perioperative regimen of chemo for patients with gastric cancer and gastro-oesophageal cancer. A very important question to ask, there has always been a debate of whether radiation adds extra benefit in addition to the perioperative chemotherapy strategy. Not surprisingly, this was a publicly funded trial. Both PFS and overall survival did not improve so we have an answer to that question here.

Speaking about publicly funded trials, I also want to talk about the AMBASSADOR trial in bladder cancer. This was a positive trial in terms of disease free survival which improved, or more than doubled, from 14 months to 30 months, but did not have overall survival. Overall survival was negative with a hazard ratio of 0.98. But I want to highlight this trial as a trial that properly reported the results. I read its publication in New England Journal of Medicine and I was happy with the way the trial was reported. They did a pretty good job because in this trial, unlike in many other trials, I saw that they had provided exact numbers and percentages for patients who were censored, the reasons for being censored. They also accounted for DFS results based on informative censoring. When they assumed worst-case scenario and treated all censored patients as having had disease progression, the disease free survival significantly still held. So that’s a very proper way of presenting the results; I wish all trials presented results that way, giving us exact numbers for patients who were censored and then doing a subgroup analysis controlling for those informative censoring.

In addition, they also provided very nice data on subsequent treatment. They gave exact numbers of patients who were eligible for subsequent treatment and at subsequent treatment what they got. So kudos to the authors, they did a pretty good job and unsurprisingly this was a publicly funded trial. But a couple of caveats: again we don’t have overall survival results yet. Sorry, we have overall survival results but it is negative so in the absence of overall survival, based on only DFS, how do we incorporate this into practice? We need to look at crossover and what patients got in the placebo arm after they progressed. As I said, they give this information very well, but we see that only 52% of the patients who were eligible to receive subsequent therapy actually got immunotherapy. I thought this was a trial where it would have made sense to mandate crossover. If crossover was mandated then that percentage would be much, much higher, 80-90%. Overall survival is negative anyway so one can also conclude that you can reserve immunotherapy for the future at the time of progression and give it only to those patients who actually need it.

Going on to the ADRIATIC trial of durvalumab in limited stage small-cell lung cancer after receiving definitive chemoradiation. Compared to placebo the durvalumab arm performed very well, in fact. It improved median survival by 22 months which is not something that we see every day. So this is a very good result and this should be practice changing. However, again, this was also a trial where mandating crossover would have been very beneficial because there were several patients in the placebo arm who could not get durvalumab at the time of relapse. But knowing the lethal prognosis of small-cell lung cancer I think these are meaningful results.

Now let’s go on to some of the GU trials, first the TiNivo-2 trial. This trial has a story to tell because if you remember the TIVO trials here, the TIVO-1 trial actually showed worse overall survival. The hazard ratio was 1.25 for overall survival for tivozanib versus sorafenib. In the TIVO-3 trial, which actually looked at only those patients who were receiving the drug as third line and beyond, there were only two months of PFS benefit and no OS. So TIVO-1 harmful OS, TIVO-3 no difference in OS, only two months PFS benefit. The renal cell cancer space has so many cancer treatments now, it’s not like we are running out of options in a couple of lines. But still this drug got approval and because it got approval in the TiNivo-2 trial they used this as a control arm and they tried to compare a regime of nivolumab, tivozanib plus nivolumab, versus tivozanib alone. So that’s the trial. So what I’m trying to say here is a drug that should not even have been approved, that is potentially even harmful, because it got approval based on two months of PFS and despite negative and even harmful OS results, because of that approval status it can now be used as a control arm in a trial and a new combination can be tested against that bad control. Even then the combination failed, so this was a negative trial. But this is how, with time, if we continue to approve drugs that do not work we harm present as well as future patients because the future patients are going to get this treatment as the standard of care in the control arm.

Then there was the LITESPARK-005 trial of belzutifan versus everolimus after patients have received 1-3 lines of treatment. It’s laughable, it’s almost comic, because PFS was statistically significant and the graphs do not follow [proportional hazards] but if you look at the median it’s 5.6 months in both arms. It’s 5.6 months in the control arm, 5.6 months in the experimental arm. So it’s zero months, no difference, in median PFS but it’s statistically significant. And overall survival is not improved. So the median PFS no difference but statistically significant, that’s because of the shape of the curve, and overall survival not significant at all but people are considering this is practice changing and this is the new standard of care. Go figure.

Then in prostate cancer the CONTACT-02 trial, cabozantinib plus atezolizumab versus the control arm that we have read a lot about, that we have criticised a lot about. It’s, again, patients getting androgen receptor blocker after having progressed on the same class of drug. Androgen receptor antagonist, they progressed on this and they got androgen receptor antagonist again. It could be abiraterone followed by enzalutamide or progression on enzalutamide are getting abiraterone. That’s not the right control. Again, OS was negative – even against a substandard or inferior control overall survival was negative but guess what was the headline? The drug significantly improves PFS and should be the standard of care, that’s the headline. So that has been a consistent theme in this ESMO.

So that’s the data that I have for you from this year’s ESMO. I hope it has been useful but I want to make it clear that the bottom line is we want success for our patients. We want our patients’ outcomes to improve and the way to achieve that is not by lowering the bar so much and not by lowering the expectations, it’s by actually improving outcomes. Our patients deserve that.

I would say it’s unethical when you have survival data to not provide that survival data information and to highlight only progression free survival. A patient would not know whether their tumour size is 3cm or 4cm, we tell them it was 3cm before and now you have 4cm, otherwise they would never know that. No-one knows what their tumour size is unless it is affecting them in terms of symptoms and causing quality of life issues, in which case we can show that our drug improves quality of life. Or they worry about it because we tell them that the fact that it grew from 3cm to 4cm means that you are going to die soon. But that’s not true, progression free survival does not correlate with overall survival in most cases.

So when you don’t have OS data that’s a different story but when you have OS data, that OS data should be the headline, that’s what patients should hear, that’s what the media should hear, that’s what we should be conveying, that survival data was this, it did not improve or it improved, whatever it is. We can’t hide survival data.