Let me start with a couple of plenary session abstracts relevant for global oncology. The first one is the OUTBACK clinical trial of locally advanced cervical cancer. This is a clinical trial that tested whether adding adjuvant chemotherapy to the standard of care cisplatin based combined chemoradiation improves outcomes for patients with locally advanced cervical cancer. The answer was that it did not – there was no improvement in overall survival or progression free survival by adding an extra four cycles of CarboTaxol as adjuvant chemotherapy beyond the standard of care which is chemoradiation. A couple of points to highlight about this trial – it is noteworthy that although technically this was a statistically negative trial in that it did not improve outcomes it was included as a plenary abstract which shows the importance of the statistically negative trials in informing or changing clinical practice. So a trial that did not improve outcomes does not necessarily mean that it’s a failed trial. More importantly, this is a globally relevant trial because cervical cancer is a huge disease burden, specifically in low and middle income countries around the world. It is still the fourth commonest cancer worldwide so this has huge global significance.

But one thing that I found a little concerning was that there were very few patients represented from low and middle income countries in this trial. This was a big trial and it must have taken a huge effort – running a clinical trial of cervical cancer is not easy – however, because this is such a big burden in low and middle income countries I would have loved to see more participation of patients from low and middle income countries which was not the case. That sometimes brings the relevance of the clinical trial findings into question because in many low and middle income countries even giving this standard of care, even the delivery of the control arm, may not always be ideal. Even the delivery of the control arm can be problematic – it’s not easy to give combined chemoradiation in the setting of low and middle income countries all the time.  So in the future when we are running clinical trials that are more important for low and middle income countries we should also make an effort to have more representation of participants from low and middle income countries.

But on a positive note this is the only plenary session abstract that was not funded by the industry, this was funded by Australia’s Medical Research Council. So this highlights the fact that there are many important clinical trials that need public funding that are being run by cooperative groups. So I’d like to take this opportunity to request more and more public funding of clinical trials that are helpful to answer globally relevant questions.

Not surprisingly this clinical trial had a primary endpoint of overall survival, unlike many other clinical trials we’ll be discussing now which did not have a primary endpoint of overall survival but rather had some surrogate markers.

The next clinical trial, again from the ASCO plenary, and globally relevant trial, is the JUPITER-02 trial. This is a trial in nasopharyngeal carcinoma, in metastatic nasopharyngeal carcinoma first line. This trial tested whether a regimen of toripalimab, a new checkpoint inhibitor, in addition to gemcitabine cisplatin chemotherapy for first line treatment of nasopharyngeal carcinoma improves outcomes. It did show that there was a significant benefit in progression free survival and there was a signal of overall survival benefit as well. The hazard ratio for progression free survival was 0.52.

A couple of points to highlight about this trial. Again, nasopharyngeal carcinoma is not a disease that is common in the Western world, it’s more common in Asia. This clinical trial provides us a lesson that countries should collaborate and work together to answer clinical questions that are relevant in their local setting. This clinical trial was a collaboration between China, Taiwan and Singapore. This can potentially be practice-changing for patients with nasopharyngeal carcinoma which is a very rare cancer elsewhere in the world so this is an example of how low and middle income countries or how different countries in different parts of the world can influence practice globally because it is next to impossible to run a trial of nasopharyngeal carcinoma in Europe or the USA or Canada, for example.

One more point I would like to highlight using the example of JUPITER-02 is about financial toxicity because toripalimab is a new checkpoint inhibitor and we already have had a lot of checkpoint inhibitors and there are a number of newer checkpoint inhibitors that are currently undergoing trials as well. One issue that we keep discussing is the issue of lack of price competition despite the availability of numerous me-too agents. But in this case this is an agent that is developed in China and from the example of lung cancer we have seen in China that when ecotinib, an EGFR TKI, was introduced in China, a locally developed me-too agent, there was price competition and it led to a reduction in prices, not only of ecotinib but also of erlotinib and gefitinib, other established EGFR TKIs. So that’s what we have seen happen in China and we have actually reported that with colleagues from China in Seminars in Oncology recently.

However, I would like to see the introduction of this new agent at a lower price, forcing price competition in the Western market as well. This is something that we have never seen happen in oncology – usually the new agents are priced similarly to the established agent. But in this case there has been some discussion about this agent might probably be priced at a cheaper rate than what other established checkpoint inhibitors are being priced at. If that leads to price competition and price reduction that would be the most influential outcome from this agent for the Western world.

Next I would like to move to the VISION trial; this is also a plenary abstract presented at ASCO 2021. This is for the patients with metastatic castration resistant prostate cancer. This trial tested whether a Lutetium-labelled PSMA agent improves outcomes. This is a newer technology so it sounds sexy, it sounds attractive because this delivers a high dose of beta radiation to prostate cancer cells and there was a lot of enthusiasm around this agent. The other thing is that there was highlighting of the fact that this improved not only radiological PFS but also improved overall survival. The hazard ratio for overall survival was 0.62 and usually we complain when it’s a surrogate marker but in this case it did improve overall survival.

However, there is an important caveat in this trial that whenever we are talking about improving overall survival we should also be careful about what the control arm is. If the control arm is less than the ideal control arm then improving overall survival is not a big deal because the control arm patients did not get the proper standard of care. That concern holds true for the VISION trial as well because the VISION trial control arm said standard of care at the discretion of the investigator but specifically excluded chemotherapy, radium, immunotherapy. So that means that although they said that the control arm would get standard of care based on the discretion of the investigator, they specifically excluded the agents that could actually make a difference. So what patients in the control arm did get, we need to read the full paper once it is published, but based on the information that we have either they got nothing, that is best supportive care alone which is, again, not the control arm which is reflective of real world practice, we do not give nothing to these patients, or they got androgen receptor inhibitors although they had already progressed on androgen receptor inhibitors which is, again, not the standard of care that we would give to our patients in the real world.

So, yes, there was an OS advantage but this was an OS advantage against a control arm that is not reflective of the real world practice. So what this means, basically, is this new agent, Lutetium, is not better than the agents that we already have, what it says is it’s better than nothing. So if we were to use this agent the patient population that would be appropriate for this agent would be the very last line of treatment when nothing else is available or when we would offer nothing else for these patients. That means as long as we have other active agents for our patients with metastatic castration resistant prostate cancer they should be getting those agents and not this newer one.

So I guess the only reason it was in the plenary was because the mechanism of action sounded very science-y or very seductive rather than actually this trial being a big practice-changing trial. There were other trials that were presented at ASCO that I found were more important globally.

The other abstract I want to discuss, let me discuss three abstracts together because they fall under a common theme. The three abstracts related to adjuvant treatment in the curative setting and in all of these three abstracts we fall under a common theme of what is the benchmark needed for changing practice in the adjuvant setting and the debate between disease free survival versus overall survival as the most relevant endpoint in this setting. So first let’s talk about the KEYNOTE 564 trial. The KEYNOTE 564 trial was also presented as a plenary abstract, it’s a trial of adjuvant immunotherapy, pembrolizumab, in renal cell cancer. It showed that, yes, giving adjuvant pembrolizumab in renal cell cancer improved disease free survival but the overall survival data are not mature.

There was a lot of discussion about this trial on social media as well. It’s always easy to say, ‘Oh yes, this is a practice-changing trial,’ and cheerlead. But if we think critically we need to think about a couple of things in the adjuvant setting. One is that in the adjuvant setting whenever we are prescribing a treatment we are necessarily over-treating. That is a given because we are treating all patients in the hope of preventing some of them from relapse. The other thing is whenever we are prescribing a treatment in the adjuvant setting the only effect on quality of life is detrimental. There can be no beneficial effect because the alternative, the control arm, is placebo. The control arm is giving no treatment. So when you compare giving treatment to giving no treatment the only effect on quality of life that can happen is detrimental. The best case scenario is that it is detrimental but it is not clinically significant, it’s clinically insignificant detriment. But there can be no scenario where the quality of life will be better in the adjuvant setting.

Because a number of these patients are already cured by surgery and may not relapse, any side effects will be a big deal because we are necessarily exposing these patients to toxicities in the hope of preventing relapse when, in fact, a number of them might never relapse even without that adjuvant treatment. So we need to keep all these things in mind. Therefore, specifically in the curative setting, we want to see overall survival data before considering any treatment as practice-changing. In the absence of overall survival data, with disease free survival alone, that should not be a benchmark for changing our practice because the question is not whether delaying relapse for a given patient is important – yes it is, delaying relapse is important for any individual patient. But the question here is whether over-treating all the patients in the hope of delaying relapse for some of them is important when overall survival is going to be the same and there are chances of severe toxicities, including financial toxicity and sometimes even death due to treatment related adverse events. So that is the whole package that we need to consider when we are thinking about adjuvant treatment; we should not be looking at just the DFS graphs, we should be thinking about this whole package.

This is the debate that we had with the ADAURA trial, as well, last year in the case of adjuvant osimertinib and these are the points that we highlighted in our commentary in The Journal of Clinical Oncology where I wrote with Dr Jack West about whether the ADAURA trial should be practice changing or not. This is the same thing, we need to think again and again in the case of any adjuvant treatment.

Specifically in the case of KEYNOTE 564, yes, there is a DFS advantage but OS data is immature. So my take home is that we need to wait for OS data because if you look at side effects it’s not trivial – there is a 15% increase in severe side effects, grade 3 or higher side effects, with adjuvant pembrolizumab. It was 33% versus 18% with placebo so this extra 15% severe side effects grade 3 or higher is not a trivial thing when, in fact, many of these patients might not even relapse.

The other thing that we need to think about is we had this debate about DFS in the case of adjuvant RCC when the S-TRAC results were published for adjuvant sunitinib. Yes, adjuvant sunitinib also improved DFS, failed to improve OS. In that time as well there were two groups of oncologists, one said this should be practice-changing and adjuvant sunitinib should be the standard of care and the other group of oncologists said, no, adjuvant sunitinib should not be the standard of care. So if at that time we said that adjuvant sunitinib is the standard of care then KEYNOTE 564 is a problem because the control arm is placebo and not sunitinib if we thought sunitinib should have been the standard of care. But if we say sunitinib should not have been the standard of care, then again KEYNOTE 564 is similar to the sunitinib in that it has improved DFS but OS data are lacking so we need to be waiting for OS data again before considering it to be practice changing.

Finally, whenever we are thinking about adjuvant treatment we need to think about what percentage of the patients in the control arm did have access to this new drug when they relapsed. Because we know that immunotherapy is an important treatment for patients with advanced or metastatic kidney cancer. So what percentage of the patients who got placebo in the adjuvant setting did get a checkpoint inhibitor when they relapsed is an important data to consider. Because if a huge percentage of these patients never got any immunotherapy when they relapsed then this trial of adjuvant pembrolizumab simply becomes a trial of whether giving pembrolizumab is important versus not giving pembrolizumab. Of course it is important but the question that needs answering is whether giving pembrolizumab in the adjuvant setting is beneficial than giving pembrolizumab when the patients relapse. That’s the question that needs answering.

Based on the same theme, if we move to the other plenary, that is the OlympiA trial of one year of adjuvant olaparib versus placebo for patients with HER2 negative early breast cancer. Again, this trial showed that there was benefit in DFS. The three year invasive disease free survival was 85.9% with one year of adjuvant olaparib versus 77.1% with placebo with a hazard ratio of 0.58. But the overall survival at this point is not significant. Again, all those issues that are discussed with KEYNOTE 564 are relevant here: a) is DFS a good enough surrogate in the early disease setting? Again, no. This has been a study for breast cancer; in breast cancer only in the HER2 positive subgroup the DFS is a valid surrogate for overall survival, yes, in the HER2 positive subgroup. But for other subgroups of breast cancer DFS has not been shown to be a good surrogate for overall survival. Therefore in this case, a HER2 negative group of patients, it is not a good surrogate for overall survival. b) As we discussed in KEYNOTE 564 we need to see what percentage of these patients did get a PARP inhibitor when they relapsed when they became metastatic. If that percentage is low then, again, that simply means giving olaparib is important and does not necessarily mean giving olaparib as adjuvant treatment is important.

A couple more points to consider in the setting of the OlympiA trial is they also included patients with hormone receptor positive subgroup but they did not necessarily use Oncotype for separating who were at a higher risk among hormone receptor positive patients. 18% of the patients were hormone receptor positive so we need to be careful to extrapolate the data from the overall trial to the hormone receptor positive subgroup. Some of these patients may not necessarily be high risk hormone receptor positive subgroup.

The age range of patients in this trial seemed inclined towards a young age group which is understandable based on the eligibility criteria of the need for a germline BRCA mutation to be eligible for this trial. Again, this trial was stopped early and we need to be careful whenever a trial is stopped early that the benefit might be exaggerated compared to the true benefit.

As with all adjuvant trials the questions are the same – are we curing patients or are we just delaying relapse? Is it possible to gain the same overall survival by using the treatment at the time of relapse on a few patients versus exposing all patients to this treatment as an adjuvant therapy?

The other trial that also fits into this category of discussion is the IMpower010 trial of atezolizumab as adjuvant treatment for lung cancer. This trial also tested 16 cycles of adjuvant atezolizumab for patients with non-small cell lung cancer from stage 1b to 3. The analysis was done in a hierarchical design in that first they would look at only the PD-L1 positive subgroup of patients between stages 2-3. If that was positive they would look at stage 2-3 all patients and if that was also positive then they would look at stage 1b-3 all patients. So far the data shows that for the PD-L1 positive subgroup between stages 2-3 there is improvement in disease free survival with a hazard ratio of 0.66. That hazard ratio becomes less and less impressive if we include all patients irrespective of PD-L1 subgroup and if we include all patients from stage 1b to stage 3. Overall survival data are not mature.

Again, all the points related to ADAURA or KEYNOTE 564 will be valid here. In fact, in this case the disease free survival is far less impressive than what we saw with ADAURA. In ADAURA the hazard ratio was 0.2 and here the hazard ratio is 0.66. So we need to see what percentage of these patients got checkpoint inhibitor when they relapsed. We need to see whether this translates to OS or not. In the absence of OS data, simply on the basis of disease free survival, this cannot be considered practice changing. It is important to note that there was a doubling of major side effects, grade 3 or grade 4 side effects – 22% of the patients who got atezolizumab versus 11% of the patients who did not, who were in the control arm. Grade 5 treatment related side effects were seen in 0.8% of the patients – we need to keep this in mind. So some of these patients might never have relapsed or might have had the same survival if they got this drug at the time of relapse but some patients are dying or having these serious side effects – 0.8% of the patients had grade 5 adverse events – when they got this treatment simply as an adjuvant treatment, especially in the hope of delaying relapse without having it. So we need to keep these things in mind when we are considering whether or not to use this trial as a practice changing trial in our practice. For me it is not a practice changing trial.

Next I would like to discuss the CheckMate 648 trial of nivolumab in oesophageal squamous cell cancer. This trial tested nivolumab plus chemotherapy and nivolumab plus ipilimumab versus chemotherapy alone for first line treatment of patients with squamous cell cancer of the oesophagus. Again this was a neat trial in that it showed OS difference, there was an OS advantage of nivolumab plus chemotherapy versus chemotherapy alone and there was also an OS advantage seen with nivolumab plus ipilimumab versus chemotherapy alone although the advantage seemed more impressive with nivolumab plus chemotherapy rather than nivolumab plus ipilimumab. There was also a PFS advantage seen with nivolumab plus chemotherapy which was not seen with nivolumab plus ipilimumab.

There was some discussion about how this is a wonderful thing because it provides us with a chemo-free option of using nivolumab plus ipilimumab in patients. Yes, it is chemo-free but I just want to highlight that chemo-free does not necessarily mean side-effect free. Immunotherapy has its own bunch of side effects so I would rather recommend nivolumab plus chemotherapy rather than nivolumab plus ipilimumab based on the data that we have.

The other thing is we keep talking about the PD-L1 levels and treatment recommendations based on those cut-offs but these cut-offs, in my opinion, are arbitrary: 1%, 2%, 5%. This should be looked at as a continuum rather than a strict below and above a certain arbitrary cut-off limit.

The other thing to consider in the case of CheckMate 648 is does ipilimumab actually add anything. I’m assuming that one relevant question to ask is whether nivolumab alone would be a standard of care and we need a trial to have an answer to that question. In the second line setting we have seen nivolumab alone is a treatment option but in the first line setting does adding ipilimumab add any meaningful activity to nivolumab alone. So that might be a relevant question to answer in the future because that could be a safer chemo-free option than nivolumab plus ipilimumab.

Finally I would like to discuss about the RELATIVITY-047 trial. This is a trial of relatlimab plus nivolumab versus nivolumab alone as first line treatment of advanced melanoma. I would like to discuss this trial because this is also a teaching opportunity on how to think about these combination treatments. Whenever we are adding treatment to a standard of care, whenever we are adding a new treatment to a standard of care, then we are necessarily increasing the therapeutic burden, we are necessarily increasing the toxicities. So in the case of combination treatment versus single agent trial, A+B versus A type of trial, the primary endpoint should necessarily be overall survival because PFS alone is not meaningful if we are adding a new agent, if we are comparing doublet versus single or triplet versus doublet or triplet versus single. Whenever we are adding an extra agent then delaying progression is understandable, it’s very easy to achieve but what patients need is improvement in survival because they are tolerating higher side effects, higher therapeutic burden by getting a combination of drugs. In fact, in first line melanoma nivolumab has been such a game changer, it has dramatically improved outcomes for our patients. So if we were to add on that we definitely need to show that it is improving survival rather than just delaying progression. So that’s an important concept to understand whenever we are thinking about combination treatments, A+B versus A or A+B versus B sort of clinical trials.

So, in the case of RELATIVITY-047, yes, there was an improvement in progression free survival but that alone should not lead to any change in practice. We need to wait for overall survival data. One final thing about all these trials that we have discussed today and all the trials that we’ll be discussing in future, we’ll be reading in papers, is that we need to think about a number of issues whenever we are dealing with these trials. I along with my colleagues at the ESMO-Magnitude of Clinical Benefit Scale Working Group, we have listed ten points to think about when doing a clinical trial. That is published in ESMO Open as an open access paper but to briefly summarise beyond some of the points that we have already discussed today we should also be thinking about crossover – whether crossover design was in the trial and whether it was appropriate or not. We should also think about censoring because informative censoring can heavily skew the clinical benefit appraisal, especially when we are using surrogate endpoints and we should also be thinking in terms of quality of life, post-progression treatment – was the post-progression subsequent treatment ideal and it is reflective of the standard of care – and so on.

I hope this roundup of ASCO 2021 was helpful for you and I look forward to listening to your comments and criticisms. Thank you.