ESMO 2021 Roundup
Dr Bishal Gyawali - Queen's University, Kingston, Canada
Hello everyone, this is Dr Bishal Gyawali from Queen’s University, Kingston, Canada and it’s my pleasure to be back with you all today to discuss about some major highlights from the ESMO 2021 meeting. This 2021 ESMO meeting has been unique in that there are plenty of important trials that were presented at this meeting; it was in a hybrid format, both an in-person component and virtual. To be honest, it was a little difficult to keep track of all the major news that was coming out of this congress but here are some of my top picks.
First, let’s start on a very positive note, let’s talk about genitourinary cancers and especially some trials that were presented regarding prostate cancer. Let’s begin with STAMPEDE. STAMPEDE is one of those trials that keeps giving; I can’t imagine any other trial group having as much effect on changing the treatment landscape for a particular type of cancer as STAMPEDE has done for prostate cancer, this is really remarkable. At this year’s ESMO STAMPEDE did this again. As we all know, STAMPEDE is a multi-arm, multi-stage trial so there are multiple concomitant different arms going on against a common control. This time the STAMPEDE result that was presented was for non-metastatic high risk localised prostate cancer patients. Previously we have received a number of practice-changing results from the STAMPEDE trial for metastatic patients but this time it’s for non-metastatic localised prostate cancer, high risk which was defined as node positive disease or node negative but T3/T4 and a PSA of more than 40 and Gleason 8-10. So for this group of patients the standard of care so far has been androgen deprivation therapy plus local radiation. In this trial they compared this standard of care versus the standard of care plus abiraterone. There was a time when they also had a treatment arm that contained both abiraterone plus enzalutamide in addition to standard of care, so they pooled both those two arms, because we know that adding enzalutamide in addition to abiraterone does not add much to the efficacy.
So this is a comparison of standard of care versus standard of care plus abiraterone with or without enzalutamide. The primary endpoint was metastasis free survival and as we know that metastasis free survival is a validated surrogate endpoint for overall survival so it’s appropriate to use metastasis free survival as the primary endpoint here. They found that the metastasis free survival had significantly improved. In six years’ time metastasis free survival has improved from 69% to 82% with a hazard ratio of 0.53. As we said, metastasis free survival is a surrogate for overall survival so unsurprisingly overall survival was also significantly improved here, from 77% to 86% with a hazard ratio of 0.6.
So these are remarkable results and I think this is definitely practice changing. So patients with non-metastatic high risk localised prostate cancer, the standard of care has changed from ADT plus radiation to ADT plus radiation plus abiraterone.
The results from the PEACE-1 trial were also presented at this year’s ESMO and this is the trial for metastatic hormone sensitive prostate cancer. This trial randomised patients to actually four arms, one was the standard of care which is ADT plus docetaxel, since docetaxel was established here, or standard of care plus abiraterone and a standard of care plus radiation versus standard of care plus abiraterone plus radiation. The comparison was done first with standard of care plus or minus radiation versus standard of care plus or minus radiation plus abiraterone. So that means we are trying to see the benefit of adding abiraterone to the standard of care and this showed that there was a significant improvement in radiological PFS in both high volume and low volume disease. There was also a significant improvement in overall survival of nearly one year.
The second comparison was with standard of care with docetaxel versus standard of care with docetaxel plus abiraterone. So what we are basically trying to see is whether adding abiraterone in addition to docetaxel, whether combining both docetaxel and abiraterone, is of any benefit or not. Because we previously knew that adding docetaxel has independent benefit and adding abiraterone has an independent benefit but whether giving both of them has an extra benefit compared to giving docetaxel only in addition to standard of care. Again, here it seems that it does have a significant benefit, especially driven by high volume disease. The hazard ratio was 0.75.
So these two are very important trials for prostate cancer that were presented this year and I want to say a couple more things about the STAMPEDE trial. This is an academia-run trial, multi-arm, multi-stage. It gives some resemblance to the recovery time of COVID-19 as well, both led by the UK. I think this is a lesson on how trials should be done. It has multiple arms that we can compare simultaneously so it keeps delivering newer and newer results. As new results are available it’s quite flexible in changing the protocol and adding the newer standard of care to the original cohort of patients. So this is the type of trials that we need in oncology.
After prostate cancer let’s talk about urothelial cancer for the neoadjuvant treatment. In the neoadjuvant and adjuvant setting there was the VESPER trial that was discussed at this ESMO meeting. We know that the standard of care in the peri-operative setting for genitourinary cancers is either gemcitabine/cisplatin or MVAC, dose-dense MVAC. Both of these are equally used across the world with institutional preferences. This VESPER trial actually tried to compare which one of them is better, GC is better or dose-dense MVAC is better. Although this trial allowed both neoadjuvant and adjuvant setting patients to be enrolled, most of the patients, 68% of the patients, were from the neoadjuvant setting. In fact, the results are not statistically significant here although they look in favour of dose-dense MVAC. If we look at only the neoadjuvant subgroup of patients then that result looks statistically significant. So what this probably means is because both of these treatments are standard of care otherwise and if we have to make a choice between these two standards of care then probably we should go with dose-dense MVAC as long as we can give the 4-6 cycles of this treatment. Because if we look at the neoadjuvant setting only then dose-dense MVAC has a significant benefit here but also the toxicity should be kept in mind.
So now moving on from the prostate cancer trials, another tumour site that had a very good time at ESMO this year is breast cancer. There are multiple important trials that were presented. First in the curative setting the BrighTNess trial looked at patients with triple negative breast cancer whether adding veliparib, a PARP inhibitor, in addition to platinum plus paclitaxel would improve event free survival and pathological complete response rate. We know that pathological complete response is not a good surrogate anyway but this trial was important because the results of the I-SPY 2 consortium had gratiated veliparib in this setting, saying that there is a high likelihood of success. However, the addition of veliparib did not seem to help in the BrighTNess trial and this is another lesson that we need to take away. We should take these pathological complete responses or the results of these platform-based trials with a grain of salt because actually veliparib having graduated from I-SPY with a very high probability of success, more than 85% chance of success, and still it did not improve outcomes here.
But having said that, this trial also reinforced the need for adding platinum to this group of patients because adding platinum to paclitaxel did improve event free survival as well. So we have always known that the platinum group of drugs have a special role in patients with triple negative breast cancer and this validates that hypothesis. I would argue that platinum should be given to this group of patients.
The other interesting finding from this ESMO was about the duration of adjuvant hormone treatment. This has been a major, big debate for a long time. But the results from the GIM4 trial show that giving five years of letrozole after 2-3 years of tamoxifen, so that’s 7-8 years of adjuvant hormone therapy in total, improved both disease free survival and overall survival. The previous duration of treatment trials usually showed an improved DFS only without showing significant improvement in overall survival but this trial showed improvement in both DFS as well as OS. It also showed that actually the benefit came later, several years after treatment, so this highlights the need to continue on adjuvant hormone therapy for a long time and to think about the ways we can improve compliance, especially when the patients might get complacent many years down the line after taking the same medication every day.
Metastatic Breast Cancer
Moving on to the metastatic setting of breast cancer, one of the remarkable results that were published was from the DESTINY Breast 03 trial which compared the new antibody-drug conjugate trastuzumab deruxtecan to TDM1 in the second line treatment of HER2 positive metastatic breast cancer. TDM1 is the appropriate control arm so there is no issue with the control arm here. This trial showed that PFS was significantly improved but OS was still immature and has not improved at this time point.
There are a couple of things about this trial. One, the magnitude of benefit is super-impressive – the hazard ratio is 0.2 with a quite significant p-value. The median PFS reached here is probably longer than the median PFS we had seen with bortezomib plus trastuzumab plus taxane in the first line trial. So in this sense this is quite impressive results and I am in no way discounting the results from this trial. I think this is a practice-changing trial although we don’t have OS results yet but I think the OS will also be out very soon.
There are a couple of things to consider here. One is that many people were talking about how significant the p-value was and they were equating the p-value with the magnitude of benefit and this is a common misunderstanding. The p-value was something into 10 to the power of 22, 23, so there was an impression that such a significant p-value is reflective of the efficacy of the treatment. I would like to take this opportunity to highlight that the p-value does not tell us anything about the magnitude of benefit. The p-value only tells us that if the null hypothesis is indeed true what is the likelihood that we are seeing these results? So what it probably tells us is that the difference in efficacy results that we are seeing are probably not the result of randomness but it does not tell us how good the treatment is. I have seen this often and often, even when discussing the results of STAMPEDE, it also had a number of zeros after the decimal point in its p-value and, again, people were discussing that prostate cancer also had so many zeros in its p-value but breast cancer has had more zeros in the DESTINY-03 trial, one day we’ll reach breast cancer. That sort of conversation makes me realise that many of us probably don’t understand what we are talking about. Many of us probably don’t understand how to interpret the trials and we are supposed to use this data for the benefit of our patients. So I think it is very important for us to understand what a p-value means, it does not mean what we think it means. It does not tell us anything about the magnitude of benefit. The hazard ratio does tell us about the magnitude of benefit and the differentiation median survival times also do tell us about the magnitude of benefit. And the hazard ratio in DESTINY-03 is pretty impressive, it’s in the range of 0.2. But one caveat is that this trial was stopped early and I have mentioned this again and again, even when discussing the AURORA trial last year, that if a trial is stopped early then its hazard ratio will necessarily look far more impressive than it actually would have been if the trial had been continued to maturity. So what I’m saying is that the actual hazard ratio for this drug is probably not 0.2, it’s probably higher than 0.2, but because it was stopped early necessarily it had to cross the stopping boundary so it does look far more impressive. This drug would still be beneficial but the hazard ratio would be less impressive than 0.2.
Again, progression free survival is not a valid surrogate endpoint here but given the totality of the evidence, looking at the median PFS with this drug in the second line compared to the median PFS we see in the first line, and the fact that the overall survival curves are also starting to separate, I anticipate that the overall survival difference will be reached and it will be significant. We have seen from so many other trials that if the drug is really this good then it should have no problem with reaching statistical significance for overall survival in a very short span of time despite crossover. We have seen that with multiple other trials, even those trials that were presented at this year’s ESMO, that if a drug is really good then it does achieve statistical significance with overall survival despite crossover. If it fails to achieve overall survival significance then we have to ask or we will have to question about the role of the drug in earlier line versus later line. But we should not get too carried away with the DESTINY-03, we should not forget the patients we still have. We know that this is only for second line, this should not be replacing our first line practice because we don’t have any information there yet. We need to also see at the publication, I haven’t seen any published results of this, in order to know what the post-progression treatments were, whether they were balanced in both arms and, most importantly, whether TDM1 would still work after trastuzumab deruxtecan. Probably it won’t because they are both antibody-drug conjugates with trastuzumab as the key component.
In fact, there was one more trial published called the TULIP trial with another antibody-drug conjugate called trastuzumab duocarmazine. This trial also compared against TDM1 in the third line setting. This trial compared against physician’s choice treatment post-TDM1 in the third line setting. 87% of the patients in this trial had already received TDM1 before, however, only 60% of the patients had received bortezomib before so that’s not an adequate treatment in my opinion. OS was not significant here although there was a significant benefit in PFS. So I don’t think TULIP-1 is practice-changing yet and it also has important ocular toxicities. Even with DESTINY-03 we saw important lung toxicities there so those should be kept in mind when interpreting the results of this trial.
Sorry, one more trial in breast cancer – the OS results from the KEYNOTE-355 were presented. For patients with PD-L1 positivity there was a substantial improvement in median overall survival of nearly 7 months which I find quite meaningful. I find it surprising that when atezolizumab for triple negative breast cancer was withdrawn from approval in the United States some of us were complaining that atezolizumab had been withdrawn, an important treatment option has been withdrawn. But I can’t reconcile the fact that atezolizumab failed in a concomitant trial to improve PFS or OS and you have pembrolizumab in the exact same setting that has improved OS by 7 months then why not just put in effort to have access to pembrolizumab in this setting rather than arguing for having atezolizumab that has clearly failed.
Moving on to cervical cancer, the KEYNOTE-826 trial showed that the addition of pembrolizumab for patients with advanced cervical cancer significantly improved outcomes. But, interestingly, the treatment landscape of cervical cancer has changed many years ago from the use of chemotherapy alone to using chemotherapy plus Avastin which improved median survival by around 3 months. But in this patient population of KEYNOTE-826 if you look at the subgroups of patients who received Avastin versus those who did not then there was no significant effect for patients who did not get Avastin. So that’s an interesting finding. But, more importantly, cervical cancer is a disease of low and middle income countries – almost 90% of cervical cancer deaths in the world are currently in low and middle income countries. These are the countries that can’t even afford Avastin so for patient populations that can’t even afford Avastin, how practice changing would these results from KEYNOTE-826 be I wonder? If they can’t afford generic biosimilar bevacizumab probably they will never be able to afford pembrolizumab. So, yes, this is a practice-changing trial provided that these patients can have access to this drug. So something needs to be done to ensure access to these drugs.
But this should not also make us forget from the key point in cervical cancer that we need to target elimination of this disease through vaccination rather than concentrating our efforts on… We should not forget that goal while we are trying to fight for access to pembrolizumab.
Adjuvant Immunotherapy Trials
I want to discuss a couple of adjuvant trials of immunotherapy, very interesting. KEYNOTE-716 in melanoma, stage 2 melanoma, disease free survival hazard ratio of 0.6 but OS data is not available yet. So this is a trend of using adjuvant treatment especially with therapy in earlier and earlier stages of the disease. We already know that this is effective in stage 3 but here this is the data for the stage 2 patient population. When we try to give adjuvant treatment to an earlier and earlier stage disease patient we are necessarily over-treating many patients because some of these patients may not relapse and most of these patients would benefit from immunotherapy even at the time of relapse, achieving long-term cure rates. So in that situation, in the absence of overall survival data, we should not just drop into changing practice based on disease free survival alone. Because if we are not improving overall survival rates and we are just delaying progression, is it worthwhile treating all the patients in that cohort in the anticipation of delaying progression for only a small fraction of patients who may be able to achieve the same outcomes anyway if they receive this treatment at the time of relapse. So in the absence of overall survival data I don’t think that it’s practice changing.
That is exactly true with the atezolizumab adjuvant trial in non-small cell lung cancer as well. The IMpower-010 trial, this was a trial of adjuvant atezolizumab for one year versus best supportive care. This is not a placebo controlled trial even. The most worrying factor about this trial is that only 32% of the patients who were in the control arm had access to immunotherapy at the time of crossover. So we know that immunotherapy for this group of patients is the standard of care when they relapse so if only one-third of the patients are getting immunotherapy at the time of relapse and even then we don’t have overall survival results then this should definitely not be practice changing on the basis of disease free survival alone. Here the disease free survival is not even as impressive as we saw in the AURORA trial with osimertinib.
GI cancers – there were a couple of really interesting results published with regards to GI cancers. One is the Atezo TRIBE trial which showed that adding atezolizumab to FOLFIRINOX in the first line treatment of patients with metastatic colorectal cancer did improve progression free survival. But we need to remember that this is a randomised phase II not a phase III, a small trial, and these results are interesting and worthy of being tested in a phase III but not practice changing yet.
The other interesting trial for me was the NEONAX trial of peri-operative neoadjuvant plus adjuvant gemcitabine Abraxane versus adjuvant gemcitabine Abraxane alone. There are a number of things to discuss about this trial, especially the ITT population versus the modified ITT population. It also shows how we lose patients at every step of treatment. More importantly, one take home message would be that some chemotherapy is probably better than no chemotherapy. With the neoadjuvant setting we are able to offer patients at least two cycles of chemotherapy before taking them to surgery compared to losing many patients with the opportunity to offer them chemotherapy after resection. However, these are just intriguing results.
The other thing I want to highlight is the GASTRIPEC-I trial of adding HIPEC to chemotherapy for patients with peritoneal metastatic gastric cancer. In short, it did not improve outcomes so HIPEC should be considered experimental for patients with metastatic gastric cancer. There was one interesting result being discussed about adagrasib, the KRAS G12C inhibitor. I was intrigued why there was so much hype around that drug but I saw that the response rate was actually only 22% in a cohort of only 45 patients. Adding cetuximab to adagrasib improved the response rate to 43% but, again, it’s a cohort of 28 patients. I don’t think a response rate of 20% should make us that excited.
I think what this means is we should be testing this drug in a randomised trial but, for me, this does not definitely mean that we should be changing practice immediately. Sometimes people argue about the difficulties of conducting randomised trials and so on but there was a nice lesson for us from the FIRSTMAPPP trial which is a trial for metastatic pheochromocytoma paraganglioma which has the incidence of metastatic disease an incidence of less than 1 per million per year. Even then we were able to do a randomised academic trial there so doing a randomised trial is absolutely possible in most cancer populations that we have.
Continuing with the GI cancers, the CheckMate 649 trial was quite interesting. We already know that nivolumab plus chemotherapy improves outcomes compared to chemo alone in patients with advanced gastric cancer first line. But this is primarily driven by patients with a CPS score of more than 5 although the intention to treat population is also positive but this is primarily driven by a CPS score of more than 5. If the CPS score is less than 5 then the benefit is probably not as good. But there was also a nivolumab plus ipilimumab arm in this trial and the results of that arm were not presented up until now but it was presented at this ESMO meeting. Nivolumab plus ipilimumab, surprisingly, did not significantly improve outcomes here. So nivolumab plus ipilimumab should not be considered for this group of patients. This is a little surprising because it goes against the CheckMate 648 trial of oesophageal squamous cell carcinoma where nivolumab plus ipilimumab also had a significant benefit. However, nivolumab plus ipilimumab should not be the standard of care in patients with gastric cancer with this CheckMate 649 population. There was also a fatal adverse event of 2% which is something that you cannot ignore. That means two out of a hundred patients would die from side effects alone.
Head and Neck Cancers (Negative Trials)
A couple of negative phase III trials that I want to highlight, especially in the head and neck cancer population. The GORTEC-REACH trial showed that the addition of avelumab maintenance did not meaningfully improve outcomes for patients with locally advanced head and neck squamous cell cancer. For patients with metastatic disease the CheckMate 651 tested nivolumab plus ipilimumab, again this did not improve outcomes. This is also a reason that immunotherapy is probably not always useful in every tumour type. Some people are wondering why CheckMate 651 was negative in advanced head and neck cancer while KEYNOTE-040 it was positive. But I would like to remind you that even KEYNOTE-040 it was not positive at first, there were just a couple of patients where data were unavailable at the first analysis. During the first two analyses of the trial it stayed negative and only at the final publication it became positive. That means it was a very fragile result, just a couple of patients missing turned the trial to negative or positive.
Again KEYNOTE-122, pembrolizumab versus chemotherapy in nasopharyngeal carcinoma. This was also a negative trial, it did not improve PFS, it did not improve OS. So in the setting of head and neck cancer immunotherapy has very low benefit overall. This was in contrast to the results of the JUPITER trial presented at ASCO for toripalimab. I think toripalimab remains the standard of care for nasopharyngeal carcinoma here.
A number of trials, single arm trials, that were celebrated, even DESTINY-Lung01. I think DESTINY-Lung01 also showed one finding that we always thought was important to consider. In fact, I had published a paper in The Journal of NCCN last year showing that the response rate and duration of response for the same drug in the same setting false when you do a randomised trial versus a non-randomised trial. In DESTINY-Lung01 last year when they presented at ASCO they had shown a median PFS of 14 months and this year when they presented the median PFS had shrunk to 8.2 months. So I think it’s just a matter of follow-up time and increasing the sample size because with very few patients you can have those outliers in there that will change the numbers. So any trial result with a small sample size reporting a huge response rate and huge duration of response, we should always consider those results with a grain of salt and we should not consider them practice changing immediately.
These were some of the important findings that I thought were interesting at ESMO 2021. But a special shout-out to one of the sessions in which I was also a part of. It’s the WHO ESMO combined session on access to relevant cancer drugs in low and middle income countries. In that session Professor Chris Booth from Queen’s University, Kingston, Canada, presented about what we call the Desert Island Project, this is simultaneously published in Lancet Oncology as well, in which we ask oncologists practising in several countries in the world about what medications they consider to be most impactful for them, most useful for them, and what medications they actually have access to. I think the results are quite eye-opening. Most of the oncologists in the world still consider the drugs that are listed in the WHO EML, Essential Medicines List, to be most useful, most impactful for them but they are having issues with access to even these old, supposedly cheap cancer drugs. So there is a long way to go in terms of equity in access to cancer care globally. I would encourage everyone to read that paper.
Moving further, the ESMO-MCBS session also had some lectures on how to read a clinical trial well, how to operate a clinical trial well. As you can know from other discussions that we have had so far, many clinical trials are presented or are published with several flaws in them, several deficiencies in them, which we should all be vigilant to in order not to be persuaded straight away by the hyping, to see the clinical trial results properly for the benefit of our patients.
One special thanks to ESMO for inviting physicians and investigators from several countries in the world, including low and middle income countries, to serve as discussants at some of the important sessions. This shows that ESMO is getting more and more inclusive globally and is not an association of only high income countries inside Europe but several investigators from LMICs can feel ESMO to be their home. So this trend continues and other organisations also follow this. We have seen several times that even when some practice-changing data has been presented from LMICs the discussions are usually from high income countries. But here for the first time at this year’s ESMO I saw that some of the practice-changing trials that were presented from high income countries’ investigators had a discussant from low and middle income countries. That was a very pleasant thing to notice.
That’s all. Thank you very much