ASCO 2023: Highlights and analysis
Assoc Prof Bishal Gyawali - Queen's University, Kingston, Canada
Saira Ahmed – ecancer Medical Reporter
Hi, welcome to this highlights and analysis video from ASCO 2023. We have Associate Professor Bishal Gyawali here with us, thank you so much for joining us.
Thank you for having me. I’m pleased to be here, as always.
Thank you. So, what do you think are the important studies from ASCO 2023 that you’d like to highlight?
Thank you. This ASCO there have been a number of important trials that were presented. Not all of them are practice changing, some of them are, some of them are not. I was very interested that a common theme that I noticed among many important trials that were discussed was the appropriateness or the inappropriateness of the control arm treatment. So, in a randomised trial we are testing an experimental arm with a control arm and if the control arm is not the standard of care, if the patients in the control arm are not getting the best treatment, then any differences in efficacy that we see are false signals, they are not the true markers of efficacy. That is something that concerns me. So I’d like to highlight some of the adjuvant trials, for example, because there is a common theme that we see about the control arm patients not getting the standard of care post-progression when they relapse.
Let’s talk about the ADAURA trial in non-small cell lung cancer; it was one of the plenaries. I have previously written a number of papers about the ADAURA trial when it was first presented with disease free survival results. At this ASCO we got the results from the overall survival so that’s a good thing because three years ago when it was presented we were asking for overall survival results and we were saying that we cannot make any firm decisions based on DFS alone. This year we got OS results but the problem is, as people may know, ADAURA is the trial of osimertinib versus placebo for patients with early resected non-small cell lung cancer that are EGFR mutation positive. But we already know that osimertinib is a highly efficacious drug in the metastatic setting as well so any patients who relapse they should get osimertinib as the first line treatment. That was proven from the FLAURA study, the same drug, same industry, a similar phase III randomised trial.
Now ADAURA is the trial of the same drug in the adjuvant setting, in the early setting. So when the patients in the control arm, when they relapse, when the disease comes back, they should mandatorily get osimertinib because that was already proven. This was also published in The New England Journal of Medicine paper and when we look at the supplement we find that only 79 among the 205 patients who relapsed from the control arm actually had access to osimertinib, they got osimertinib. So that is less than 40% of the patients in the control arm who were supposed to get osimertinib did not get osimertinib. So if the patients in the control arm are not getting the best treatment when they relapse then any survival difference that we see might be an effect of patients not getting the right treatment rather than the experimental treatment being better. The survival difference that we see in ADAURA, it’s substantial, it’s a good survival difference that we see but we do not know what part of it is because the patients in the control arm did not get osimertinib upon relapse.
Now, there is an interesting theme here when people talk about these trials in social media or elsewhere. People think that it’s always either/or, it’s always black and white, that either you say that this is a really wonderful trial, this is a really wonderful drug and great news for our patients or you are supposed to say this is the worst thing ever and we are never going to use it. No, it’s not always black and white like that. There are problems with this trial which I have highlighted in two of my papers in the past, one in JCO, one in JAMA Oncology if people are interested in reading it. Having said that, what are we going to do to our patients tomorrow if they come to our clinic? Based on these results which are a very strong OS signal we need to discuss with the patient about this option but we should also remember this caveat. Using this drug or discussing this drug with the patient does not mean you should not criticise the flaws of the trial. We should be vocal about it, we should criticise these problems in the trial. These are unethical, almost inexcusable problems in this trial that the same company ran the trial of the same drug and proved that it is important and improves survival when given first line and the same company runs an adjuvant trial and does not mandate that their own drug should be given when the patients in the control arm relapse, that is unethical. So we need to criticise that. Even if we are going to use that drug tomorrow we need to criticise this trial so that such problems do not happen in future trials. We cannot let them just slide by; we need to be vocal about it, it is our duty to say that that was an unethical practice and our future trials should not be doing that.
One more thing I would like to highlight, and this has been a common theme about news coverage, several media have been reporting on these important trials that are coming out of ASCO. Several key people with billions of followers on social media are tweeting about this trial but they are not giving the right message, they are not providing these nuances in the discussion. For example, when people are talking about ADAURA people are saying that it reduces mortality by 50%. The hazard ratio is 0.5, 0.49, around 0.5, that means it reduces the risk of dying by almost 50% which is not the same as saying that it reduces death by 50%. Because people will be misled, we need to think about the absolute difference in survival. The absolute difference in survival if you look at the overall population is 10% if we look at stage 2, stage 3 it’s almost 12%, which is an important survival difference, but also the caveat that the control arm actually did not get the best treatment. But it’s not 50%. Even if we ignore that, it’s not 50%. When we say 50% people will think that, wow, this is really remarkable, this is something that is as good as imatinib. But that’s not true. The absolute difference is 10% to 12%, it’s not 50%. But you see that in important newspapers, important people are saying that it is reducing death by 50%, they are not even saying the risk of death or the hazard of death, and not saying that this is a relative reduction not an absolute reduction, which is misleading.
Now, going on to this issue of whether the patients in the control arm get the standard of care when they relapse or not, the other trial that was presented was the NATALEE trial. It was a trial that we very much looked forward to. It is a trial of adjuvant ribociclib in patients with hormone receptor positive breast cancer in the early setting. It’s similar to the monarchE trial that was presented at last year’s ASCO, maybe two years ago. We have recently published a paper about the monarchE trial and the flaws in the monarchE trial that limits its interpretation which is published in Lancet Oncology. Similar issues apply to the NATALEE trial as well.
The primary endpoint in this trial is three-year DFS rates and we see that the three-year disease free survival rates have improved by 3.3% in the NATALEE trial by using ribociclib. But we still don’t’ know the overall survival results; we do not know whether it’s going to improve overall survival or not. The follow-up time is too small, only 28 months. This trial has enrolled more than 5,000 patients; if you enrol so many patients, any sliver of benefit will look significant. But, again, the key question here is the patients in the control arm, what percentage of them got ribociclib or any CDK4/6 inhibitor when they relapsed? If patients in the control arm did not get CDK4/6 inhibitors when they relapsed then, again, OS data will be meaningless. You can show better results by having a better drug or by making the control arm worse. So, so far, the common theme that we are seeing is we are trying to make the control arm perform really badly instead of trying to show that the experimental arm is better. That is something that is concerning for all of us, or should be concerning for all of us.
In the same vein I want to talk about the IMbrave050 trial of atezolizumab plus bevacizumab in hepatocellular cancer. Again, this is an adjuvant trial, early setting, after resection atezolizumab plus bevacizumab versus active surveillance which is the right trial, an important trial, right design. Patients were randomised to get atezolizumab plus bevacizumab for one year versus active surveillance alone. The primary endpoint was recurrence free survival and they showed that the recurrence free survival significantly improved, with a hazard ratio of 0.72. But, again, this paper has not been published so I do not know the full details but whatever was presented it said crossover was permitted. Permitted means allowed if they wanted to.
Atezolizumab plus bevacizumab has proven overall survival in the metastatic setting. So these patients on the control arm, if they relapse they should get it. It’s the same combination – atezolizumab plus bevacizumab – the same company, same drug. But they say ‘permitted’, they do not say ‘mandated’. They should have been mandated to receive this drug when they relapse. I don’t have the full publication so I don’t know what the details are but if these patients were not mandated to get atezolizumab plus bevacizumab, despite being eligible for it, then that’s a big flaw in the trial design. Patients should get standard post-protocol therapy, even if it’s a different drug or a different treatment. But in this case it’s the same exact drug that they should have gotten so it’s the same industry running the trials so they should be providing those drugs for free and they should be mandating it. But this is something that did not happen, which is a problem.
Again, why recurrence free survival instead of overall survival? That’s another question that we need to be asking. Especially in the adjuvant or early settings where we are going to treat all of the patients with the hope of saving some of the patients because, by definition, in the adjuvant setting we will be overtreating a huge number of patients who would have already been cured or who would have recurred despite the treatment. So overall survival data is a must there.
Which brings me to ATTRACTION-05, an adjuvant trial of nivolumab in patients with oesophageal gastric cancer. Actually, in this trial the trial was presented very well because the presenting author mentioned not only that the primary endpoint was recurrence free survival but also mentioned that they chose recurrence free survival because they looked at the validation of the surrogacy and they found that it was a strong surrogate for overall survival in this setting. That’s the reason why they chose recurrence free survival. I have not heard any other presentation where they provide such a logical argument for why they chose a given surrogate endpoint instead of choosing overall survival.
In any case, this was among pathologically stage 3 patients and they found that actually this was a negative trial in that the three-year recurrence free survival rates were 68% versus 65% but three-year overall survival rates were 81.5% versus 70%, again, not significant. Which is interesting because a similar trial, CheckMate-577, which tested nivolumab versus placebo after the patients had received neoadjuvant chemoradiation, was positive but ATTRACTION-05 is negative. So we need to think about what are the differences that explain these discrepancies.
Overall this was a negative trial but an important one. So, for patients after D2 resection pathological stage 3, adding nivolumab to regular chemotherapy does not improve outcomes and should not be the standard of care would be the message from this study. But, again, we need to think about why does it differ from the CheckMate trial.
Now let’s move on to neoadjuvant settings. Let’s talk about the KEYNOTE-671 trial which is the trial of pembrolizumab in non-small cell lung cancer. It had co-primary endpoints of event free survival plus overall survival. They gave pembrolizumab in addition to chemotherapy for four cycles before surgery, then the patient received surgery and then followed with pembrolizumab for one year. The two-year event free survival rates were significant – 62% versus 41% – but at this interim analysis the overall survival was not significant.
Again, a number of questions arise here. One is that event free survival is positive but overall survival, as of yet, is not positive. So in that case I don’t consider it to be a practice-changing trial. Also, given the fact that patients with PD-L1 less than 1% were 35% of the total population, probably the overall result that we are seeing is highly skewed by the PD-L1 more than 50% population who benefit more than the rest of the patient population. A little concerningly, EGFR ALK testing was not mandated in this trial, I don’t understand why. So if we can run a trial of pembrolizumab then we can easily do EGFR ALK testing for these patients. But it was not done so that is a huge limitation of this trial – we do not know the efficacy based on the mutation status.
Having said that, I also don’t see any reason to use this regimen instead of CheckMate-816. Because CheckMate-816 in a similar patient population is a trial of neoadjuvant nivolumab and this is just three cycles – you give three cycles of nivolumab plus chemotherapy – and it has proven EFS advantage, just like this KEYNOTE-671 showed. Overall survival results are still maturing. So if one wants to use immunotherapy in a neoadjuvant fashion why not just use three cycles of nivo – it’s much more less toxic and it will be cheaper compared to using it for one year. Again, with these trials, as I mentioned, we need to look at whether the control arm patients when they relapsed they got immunotherapy or not.
We are talking about neoadjuvant trials, the PROSPECT trial was presented at the plenary session. It was a very important trial. A couple of things I wanted to highlight about this trial is that, a, it was not an industry sponsored trial, it was sponsored by the Alliance Clinical Trial Group and that’s why it asked a meaningful question – whether we can safely omit radiotherapy for a selected group of patients with rectal cancer. It was a non-inferiority design trial. That’s why I want to keep highlighting that we need public funding of clinical trials, we need academia-sponsored clinical trials, because only publicly-funded, academia-sponsored clinical trials can answer important questions of clinical need.
But the problem is this trial started in 2012 and ran through 2018. The way we treat rectal cancer has vastly changed from when this trial was started. So there is a problem of the trial results becoming outdated by the time the results are available. But I don’t want to discount the importance of this trial; this trial is very important. They looked at whether giving neoadjuvant chemotherapy, and this is a patient population with T2 node positive, T3 node positive or T3 node negative, they excluded high risk patients. Again, I was talking about how media can misinterpret or misinform the public, I think in important media outlets the headline was, ‘A new trial shows that patients can escape brutal radiotherapy,’ which is misleading for the public because, yes, you can escape radiotherapy, based on this trial you can escape radiotherapy for this specially selected group of patients, not for everyone. But that nuance will be lost if you look at the headline and you are a patient with rectal cancer then you say, ‘Okay, I don’t need radiation.’
What did this trial do? It randomised patients to get either the standard of care which is chemoradiation before surgery followed by surgery followed by adjuvant chemo if needed, or to get six cycles of FOLFOX before the surgery, skip the radiation and, based on the pathological status after the surgery, they can get chemoradiation or chemotherapy after the surgery. For the first time in this trial presentation, the presenter highlighted the reason for choosing the non-inferiority limit. In non-inferiority trials we have a limit based on how we define non-inferiority – if it is up to this limit it is non-inferior, otherwise it is inferior. Usually there are so many non-inferiority trials published in oncology they don’t give a rationale for why this particular definition was chosen. It’s just we chose a non-inferiority limit hazard ratio of 1.33, nobody knows why. But during this presentation it was clearly highlighted that it was chosen based on a consensus of the Alliance Steering Committee members that also included input from patient representatives, which is very important. We need to ask patients what is non-inferior for them, we cannot just say, ‘Okay, 33% increase in the risk of dying is non-inferior.’
What this trial found was that, yes, escaping radiotherapy was non-inferior. The five-year disease free survival rates were 81.8% versus 78.6%; five-year overall survival rates were 89.5% versus 90.2%. The five-year DFS non-inferiority limit should have been within 5%, based on the trial definition and that was met so non-inferiority could be claimed. But the rationale for a non-inferiority design trial was thinking that radiation will have too many side effects on the patients and we wanted to avoid radiation and therefore can we give just chemo and not do radiation. But when we look at the toxicity data, actually the radiation group did not show too much worse toxicity compared to the chemotherapy group. I think the quality of life data is also coming out and there was a brief slide about quality of life data so that will be very interesting to analyse in detail. But, on the first look, if you look at the toxicities, the toxicity in the radiation group did not look worse than the toxicity in the interventional arm.
But one more caveat is that, as I mentioned, the way we treat colorectal cancer has changed since the trial started. So nowadays we are trying to de-escalate treatment further. In this trial patients got six cycles of chemotherapy before the surgery and then they also got chemotherapy after the surgery. So many people, especially for this low-risk disease that are included in this trial, many of us would think that that is overtreatment, probably you do not even need that many cycles of chemo. If you are giving six cycles of chemo before surgery then probably you do not need to give any more chemotherapy after the surgery.
There are also some select groups of these low-risk patients who would not want to have surgery. There are different nuances to this but an important thing to highlight to the public who have been reading this in newspapers is that this is not every rectal cancer patient, this is highly selected low-risk patients who would have been eligible for up-front surgery, [? 22:02]-sparing, organ-preserving surgery, anyway. So this is not every rectal cancer patient.
Do you think the results from this study will have an important clinical impact now or in the future?
Yes, I think so because, as I said, although the way we are treating rectal cancer has changed, what this shows is that for a select group of patients radiation might be skipped and so it will help us in shared decision making. It won’t change practice for everyone, so when we have a patient who fits the criteria of this trial we can have these options about, ‘Okay, you have the option of getting radiation, you have the option of getting chemo and these are the side effects from radiation. With chemo these are the side effects of chemo only.' As I mentioned, we might even be able to skip the chemo after the surgery. So it will help us in having those discussions.
But I now remember, I wanted to highlight one thing about this trial is that it is published in The New England Journal of Medicine. It is one of the best reported trials that I have read. There is no spin in the results. The publication has been written excellently. Enough information is provided about everything; the supplements are full of data, there is rich data in there. They are not hiding anything. I think that the quality of life data is getting published as well pretty soon. But when I read the trial… usually when I read several trial publications I find that, ‘Okay, why are they hiding this? Why are they not telling this? Why is there spin in the results?’ But this trial was written excellently.
Shall we move to supportive care?
Yes, before going there let’s talk about one more plenary, the INDIGO trial. It was among patients with glioma, grade 2 glioma. It was a trial of this new drug which is an IDH1/2 inhibitor called vorasidenib. The bottom line result is that it showed that it improved progression free survival and many people celebrated these results. But it’s a little premature to conclude that this is a really wonderful result, that this is practice changing. The reason why I say that is these are young patients, the median age is less than 40, 39-40. These are patients who had residual tumour or a relapsed tumour after surgery. If we look at the trial design, these patients are getting enrolled into this trial 1-5 years after the surgery, so this is not immediately after the surgery. The median time is 2½ years after the patient had surgery. So when these patients are being enrolled into the trial, the enrolling physician thought that, okay, they need to get some treatment. So they were not getting anything for 2½ years, for 3 years. Now if someone is thinking that, okay, we should start treatment of this patient then probably there is a reason why we want to be started. So the control arm cannot be just active surveillance, like the placebo; the control arm is placebo. So if we thought that the patient should be getting an active drug is placebo the right control arm is one concern.
The other thing is there is no granular data available to assess whether these patients in the placebo arm had a longer duration post-surgery at the time of initiation of the trial so that their progression free survival will be shorter. Because if they started relatively longer than the patients in the experimental arm then they will obviously progress sooner after the start of the trial. So, yes, that control arm is something that I cannot fully agree on.
The other thing is they mentioned that they have excluded patients with high-risk features based on the MRI findings and other findings. They specifically excluded patients with high-risk features. So if we are excluding patients with high-risk features and if we are including patients with low-risk anyway, that means, in a sense then, we are giving them placebo as a control arm. That means these are patients who do not need anything based on our normal clinical impression. That’s why we excluded patients with high risk and we gave them placebo. So if we are trying to treat people with an active drug which has side effects when the standard is to do nothing then we need to show that it improves survival otherwise why would we treat people whom we would have otherwise have done nothing? But the primary endpoint is not overall survival, the primary endpoint is just progression free survival. So we are going to put these patients on this drug, which is going to be pretty expensive, thousands of dollars every month, tens of thousands of dollars every month, I guess, but we are asking them to be on this drug for a long, long time, for years, for what? So that we can just delay starting the treatment? And that is, in fact, one of the endpoints – time to the initiation of the other treatment. We need to improve the survival, we need to improve their life, we need to improve their quality of life and we need to improve the length of life. We cannot ask these young patients to be forever on a pill…
To risk their lives in that way.
Exactly, when, in fact, it does not even improve their life. The other thing I want to highlight is our crossover. Two things happened in these examples. Crossover is when patients in the control arm get the experimental arm when they progress. So in the IMbrave trial, in the ADAURA trial, in those KEYNOTE trials, crossover should have been mandated because these drugs were already proven at the time of relapse. So if the patients in the control arm relapse they should mandatorily get this drug but it did not happen, they did not mandate it in these trials. They said crossover was allowed, some of the patients got access, less than 40% of them got access to osimertinib in ADAURA. It should have been mandated but they did not do it.
Here, in the INDIGO trial there should have been no crossover but they mandated it. Because this drug has never been proven in glioma. So if a drug has never been proven there should be no crossover because it will dilute your survival results. What I mean by that is, for example, when these patients in the control arm progressed then they would have gotten the standard of care which would have involved chemoradiation or second surgery. But instead of that now they will mandatorily get this new drug. If this new drug, that’s what we saw with sipuleucel-T in prostate cancer as well, if this new drug has never been proven to improve survival then the patients in the control arm will do worse because they did not get access to this surgery and radiation and other treatment that would actually have helped them. They spent their time getting this new drug.
Because you don’t know what the answer is until the end.
Exactly. So in this trial there should have been no crossover but they mandated it. So we are seeing flaws everywhere.
Let’s talk about the DUO-O trial in ovarian cancer. This was a trial that tested chemo plus bevacizumab plus durvalumab plus olaparib against chemo plus bevacizumab followed by bevacizumab maintenance and showed that, compared to chemo plus bev followed by bev, if you do chemo plus bev plus durvalumab plus olaparaib followed by bev plus durvalumab plus olaparib then it improves PFS. I find it hilarious. We are giving four drugs…
I’d like to know the tolerability report of this study.
Yes. In fact it would be five drugs because chemo is a combination of two different chemos, plus bev, plus durvalumab plus olaparib. Then even in the maintenance phase we are giving three drugs, bevacizumab plus durvalumab plus olaparib. Then we are just looking at PFS. So if we give five drugs, three drugs, you will improve PFS. You need to improve survival to show that it is beneficial. Improving PFS is not a big deal if you combine… you can combine three more drugs, why stop there? Let’s combine another three. It will improve PFS, it’s no wonder that if you keep combining different agents it will improve PFS. You need to show improved OS to prove that combining these all agents has a net therapeutic benefit for the patient otherwise they can get it sequentially.
Again, in one of the news outlets the headline was that this improved survival, which is misleading. It does not improve survival, it’s only the PFS results. Interestingly, the control arm here is chemo plus bev when, in fact, the PAOLA-1 trial has previously shown that if you combine olaparib then it improves PFS. But there was no control arm that said chemo plus bev plus olaparib followed by bev plus olaparib. They had only chemo plus bev as the control arm and then in the experimental arm they combined everything else. So it’s no wonder that it improves PFS, why would it not because just bev plus olaparib itself improves PFS versus bev alone.
So usually I argue that if it’s a combination treatment or if it’s a maintenance treatment you must have OS. In other settings you might have some arguments but if it is a combination treatment or it is a maintenance treatment you are increasing the therapeutic burden on the patient and you are using drugs that the patient could have used in the future up front. So you need to show overall survival benefit. But in this case it’s both, it’s combination plus maintenance both and it still does not have any OS and still people think that it’s a practice-changing result.
Would you like to highlight the tolerability or toxicity report of this paper?
That’s an important point. I don’t remember the data clearly but when you combine drugs you obviously, by definition, increase toxicities. Here we are combining an immunotherapy plus olaparib so it is going to increase toxicity substantially, I’m pretty confident. Having said that, I remember that in ovarian cancer there was one more trial, the MIRASOL trial. One of my colleagues, Enrique Soto, he highlighted it in one of his tweets. I don’t remember the exact numbers but I think this drug improved survival in the MIRASOL trial, however, if you look at the toxicity profile there is an 8% chance of grade 3 or higher blurring of vision. What does grade 3 or higher blurring of vision mean? If we look at the CTCAE criteria what does grade 3 and grade 4 mean? It’s impairing your daily life, it’s almost being blind. So it increases the risk of being blind and thereby having a poor quality of life. So 8% of the people becoming almost blind, blind enough to have impairment of their daily life, is a pretty significant number but people say that toxicities are acceptable, manageable. By definition it’s not manageable and we did not ask patients whether it’s acceptable to them or not.
There is a new way of spinning the side effects or the toxicity results. People say that there was no new safety signal observed. Are we expecting to observe new safety signals in every trial? If we start seeing such new safety signals in all the trials then it’s a problem. We do not expect to see new safety signals all the time but we have to report our toxicity results in a clear, non-biased and honest way.
Speaking about honest reporting and clear results, I very much like this clinical trial called the SONIA trial that was presented today morning in the breast session, metastatic breast session. CDK4/6 inhibitors palbociclib, abemaciclib, ribociclib, they have been proven in second-line treatment, they have also been proven in first-line treatment. I have been talking about the control arm not getting good treatment when they relapse. The first-line trials, in most of these trials this has been a problem in the first line – patients in the control arm don’t get CDK4/6 inhibitors when they relapse when, in fact, we have already shown that they are effective in second line. In fact, in one of the MONALEESA trials, the first-line trial, patients in the control arm, 25% of the patients did not get any treatment after they relapsed which is unthinkable for patients with hormone receptor positive breast cancer. There are several lines of treatment – there is hormone therapy, there is chemotherapy, but they did not get any treatment. They should be getting CDK4/6 inhibitors when they relapse.
So there has been a question about should we give CDK4/6 inhibitors to everyone up front or we can give it as second line. Because if we give it as second line we will be sparing toxicities for many of these patients and if we can achieve the same outcomes by giving it in second line then we’ll be giving it for a shorter duration, we’ll be giving it for a lesser number of patients and still achieve the same outcomes, which is pretty important for the patients, which is important for the health system.
SONIA is a very pragmatic trial that tried to answer exactly the same question. So it’s a trial of first-line CDK4/6 versus second-line CDK4/6 in patients with hormone receptor positive, HER2 negative metastatic breast cancer. Not surprisingly, this is an important question but not in industry’s interest so not surprisingly this was funded by the government, funded by the Dutch government, academia-sponsored trial. It ran through all but one cancer centre in the country so everyone was aligned in this trial, which is a very nice thing to see. We want to see such trials happening more and more because if the public funds are not used to answer these important trials then who is going to?
What this trial showed is that no matter whether you give a CDK4/6 inhibitor first line or second line, overall survival is the same. The PFS2 results were the primary endpoint and, unsurprisingly, in PFS1, after the first-line therapy, there is benefit for first-line CDK4/6 inhibitors. But if the control arm patients get it second line then that benefit vanishes, meaning that you can give it second line and have the same benefit. So you don’t need to give it to everyone first line.
The tragedy is, if those first-line CDK4/6 inhibitor trials had mandated crossover, then we even did not need this trial; we would not have even needed this trial. We would have known that the overall survival would be the same and we can give it second line. But because they did not mandate crossover we needed a different trial. Interestingly, the presenter of this trial made an excellent point that the trial was self-funded because the patients in the control arm got it at second line, they got it for a shorter duration. The differences between the duration of treatment for patients who got it first line versus second line was almost 16 months. So we are sparing patients 16 months of extra therapeutic burden and toxicities and achieving the same outcomes. That also leads to cost savings. So the presenter showed that that cost saving, in itself, funded the trial. The cost of the trial was recouped by just those cost savings. Now we can imagine how much cost saving we can have in the future when we are implementing the results of this trial to everyone.
So this is the philosophy of the dose escalation or the dose optimisation clinical trials. With immunotherapy we know that obviously we are overtreating patients, we are giving too high a dose at too frequent intervals for too many patients for too long a time. So we have so many opportunities to run these types of non-inferiority trials testing whether a lower dose for a lower frequency, for a lower duration in selected patients will be enough. So how does the funding for these trials come, because it is not in industry’s best interest? But it will be in the best interests of the payers because, similar to this trial, the experimental arm patients who will be getting the less intense regimen will lead to some cost saving and that will fund the trial. But in the broad scheme, if these trial results are positive then that will lead to a changing treatment plan as a less intensive treatment for everyone within the healthcare system so that will lead to a huge amount of cost saving. So it is in the best interest of payers to fund these types of trials.
I wanted to give a shout out to a supportive care trial because we have been talking about new drugs all the time which may or may not be relevant to patients or physicians living in low and middle income countries. But this supportive care trial actually comes from India. This is a trial of topical diclofenac gel versus placebo to prevent the incidence of hand foot skin reactions in patients getting capecitabine. Capecitabine is an important drug, we use it for GI cancers, for breast cancer. They showed that actually using topical diclofenac gel reduces the incidence of all grade, as well as severe grade, grade 3 and higher, hand foot skin reactions. So this is an immediately practice-changing trial. We can start recommending topical diclofenac gel to our patients on capecitabine from tomorrow.
These are the types of locally important, locally developed and planned clinical trials that are meaningful globally but these do not become the centre of attraction. I was actually disappointed about the SONIA trial not going into plenary. I think the SONIA trial was pretty important in several respects. It showed the importance of public funding and academically sponsored trials in answering important clinical questions. It was an important clinical question for patients with breast cancer. It was a proof of concept for the feasibility of running such a trial in the sense of having a self-funded trial and it is practice changing. I was disappointed when I heard that people were saying, ‘We still do not have good biomarkers to select who can safely be given second-line treatment.’ That’s the other way around – the trial shows that you can give it as second line to everyone. You will need a biomarker to prove someone gets it first line, not the other way around.
So it balances it out. We will move on next to what are the issues with drug approvals that we can see?
Yes, there have been a number of discussions and debates recently about the drug approval and reimbursement processes, especially with regards to accelerated approval. The FDA has published a new guidance as well and we are writing something about it as well. But there is always this tension between two groups of stakeholders – one group wants early access to possibly all treatments that may be beneficial; the other group wants to make sure that these treatments are actually safe and effective. There is the policy angle where the system decides even if it may be beneficial for one or two patients whether it is in the system’s best interest to fund it for everyone, as opposed to funding something else.
So at the heart of this controversy is the issue of surrogate endpoints and the lack of validation of the surrogate endpoints and the issue of the quality of the trials – we talked about so many flawed trial designs today – and the issue of the high cost of cancer drugs. So how can the system sustain if the cancer drugs are this expensive and are very marginally effective? That’s actually the theme of our paper that we published in Lancet Oncology just one day before ASCO.
In that paper we look at a cohort of cancer drugs that have the least magnitude of benefit, based on the MCPC score. All of these drugs have been approved by the US, of course, and, by definition, Medicare has to pay for them. But we found that several of these high-income countries they don’t reimburse these drugs, they don’t pay for these drugs. For example, New Zealand pays for 0% of these drugs. If you look at the mortality rates of these countries then they are not too far off from what the cancer mortality rates are in the US. That means that in the US we are spending a huge amount of money for nothing, in a sense, because we are spending a huge amount of money on these marginal cancer drugs. So if people are interested they can read our paper in Lancet Oncology that was published on May 31st or June 1st.
So we are talking about spending on cancer drugs, these expensive cancer drugs. For example, in this ASCO we had a trial of talazoparib plus enzalutamide versus enzalutamide plus placebo, the TALAPRO-2 trial. We had a similar trial of olaparib and niraparib in the past as well. Nowadays we are trying to combine more drugs up front instead of giving them sequentially. As mentioned earlier, this leads to increased therapeutic burden, it leads to increased cost but the primary endpoint is not survival. So if we want to give more drugs and increase the expense then we need to see survival benefit.
There was also an important trial, the CONTACT-03 trial published in Renal Cell Cancer. As second line treatment atezolizumab plus cabozantinib versus cabozantinib among patients who have already progressed on first line immunotherapy. It was a negative trial but it was not surprising that it was a negative trial because these patients had already progressed on immunotherapy. There was no reason that they would have benefitted from continuous [?? 46:13] therapy. But it’s a good proof of concept that if a patient has progressed on immunotherapy then probably giving immunotherapy again is not a good idea.
I hope that’s what everyone was doing because if they have progressed already on one drug there is no point in giving a similar drug again. But this trial solidifies that once you have progressed on a similar drug you should not be giving the similar drug again even if it’s an immunotherapy. That’s also why we think that giving abiraterone after progressing on enzalutamide, or vice versa, is also a bad control arm. Because these patients have already progressed on a similar agent in prostate cancer.
So that’s all that I have from this ASCO.
Thank you very much for your valuable insights and we hope to see you next year at ASCO.
Thank you very much, it’s always a pleasure.