ESMO is the leading professional organisation for medical oncology, and the annual ESMO Congress attracts thousands of oncologists.At the ESMO Congress of 2022, although Chinese oncologists could not arrive at the European Conference, we had the honor to interview professor Solange Peters, President of ESMO.
Oncology Frontier: Chinese doctors were unable to attend the ESMO on site due to COVID-19, could you introduce the highlights of this ESMO conference to them?
Dr Peters: Thanks a lot for the question. It is quite important for me, first of all, to recognize how difficult it is sometimes to not be able to be present at a Congress; not that the data can’t be seen, because we will share the data very soon, but for the Chinese doctors, it is important for them to know that after three months, ESMO members can see everything, and after six months, it will be completely accessible for all. But in three months, the data will be available on the website. But the highlights are mainly what we tend to select and present for the Presidential Session. We have very strong selection criteria and scrutiny to select the Presidential presentations. For us, Presidential means it has to be practice changing. When you go back home, you change the way you work, or maybe the way you behave in your environment. So the highlights are in the Presidential Session. You can look and read about the Presidential Session. There will be a lot of articles from journalists because they highlight what will make you change your practice. That is basically the definition. There are many highlights, and I would like to outline three important ones that were presented yesterday. First of all, the role of pollution, which can be interesting and important for some European cities, as well as some Asian cities. The fact that pollution can be the source of cancers in smokers, particularly lung cancer with EGFR or KRAS mutations. Yesterday, Charles Swanton showed the relationship between pollution and lung cancer, but also the mechanism behind that, and could explain how to potentially take action - of course, by changing and protecting the environment, but also in terms of biology. To me, this is the phenomenal abstract of the meeting, because it speaks for sustainability initiatives, but also it speaks for high level biological explanations. Sometimes what you see in China more than we do in Europe are people with lung cancer who have never smoked, which is a most terrible situation. The abstract is there for you to go through. It is the beginning of a new explanation for lung cancer. The second abstract yesterday was on the use of T-cell therapy - tumor infiltrating lymphocytes (TIL). You take a tumor. You take the lymphocytes out of the tumor, and you culture them. You reinfuse the cells after chemotherapy. It is a little different than CAR T-cells in leukemia or lymphoma, but this is in melanoma and derived from the lymphocytes found in the tumor. John Haanen did a European trial comparing this TIL therapy to the usual ipilimumab in second-line melanoma. He could show a major improvement in disease-free survival. So this is very good for melanoma patients, but also a completely new way of treating patients. It might be expandable to other disease types and to other strategies. It is like a new class of drugs. Last, but not least, yesterday we heard about the gamma secretase inhibitor, another new class of drugs. It was used in a rare desmoid tumor, which had no previous standard-of-care. We are showing the community that we need to address clinical trials for those patients who have often rare diseases, and this is feasible thanks to international collaborations. These are very important landmarks. Today and tomorrow in the Presidential Sessions, we will see melanoma data on neoadjuvant or adjuvant therapy. This is very important for immunotherapy. We will have the answer. We will see renal data - neoadjuvant or adjuvant: how to make things better in the adjuvant setting? Maybe we can be combining more drugs - immuno-oncology agents, antiangiogenic, TKIs - but we will have the answers in the next Presidential Session. Last, but not least, for the first time we have the randomized data of the KRAS inhibitors. When I started in oncology, I was told KRAS was not actionable. Now we have KRAS G12C inhibitors compared in second-line with docetaxel and showing superiority. And lastly, a trial that I really like addressing those patients that are always excluded from clinical trials - PF2 and PF3 are the patients with PF1 but elderly and with comorbidities. They are deemed platinum-ineligible. In this trial, we test how atezolizumab can be used as monotherapy compared to gemcitabine or vinorelbine with a single arm and a single agent strategy that we usually use. It shows a benefit in everything - OS, PFS, response rate, quality of life. So, diversity is also about patients who are frail and fragile.
Oncology Frontier: You presented the KEYNOTE-091 study about the PD-L1 expression and clinical outcome. The benefit of anti-PD-1 adjuvant therapy in the patients with high expression of PD-L1 is not as ourstanding as expected.How should this be explained?
Dr Peters: The first trial, the KEYNOTE-091 trial, is one of the two adjuvant trials that have delivered results today. It is randomizing patients with completely resected non-small cell lung cancer with an optional chemotherapy - one year of pembrolizumab or one year of placebo. There are dual primary endpoints. Dual means that if one or both prove positive, the trial is positive. One is disease-free survival in the whole patient population. The other endpoint (and probably the one we expected to be the easiest) is DFS in the high PD-L1 (more than 50% TPS score) population. While in the first presentation we could see a benefit that is statistically significant and clinically meaningful in the intent-to-treat (ITT) population, surprisingly, we could not see the same benefit in the high PD-L1 population, which was supposed to be the easiest to reach. Tomorrow, I am doing a presentation to try to explain why. We tried to look why this specific population showed no improvement. First we looked at patient characteristics. Are they different between the ITT and the high PD-L1s? The answer is no, baseline characteristics are exactly the same. Second, is the toxicity different - more or less? No, toxicity is the same. Third, is the drug exposure different? No, the drug exposure is the same. So we had to go back to Kaplan‐Meier curves, and the good news is that pembrolizumab works as expected. The more PD-L1, the more benefit. The benefit is incremental in terms of the Kaplan‐Meier curve in the high PD-L1 subgroup, as is expected. The unexpected surprise is that in the placebo arm, the high PD-L1s were also doing better, which is contrary to the data, because until now, we knew that if we use chemotherapy or nothing, usually PD-L1 is not prognostic. If anything, it indicates a poor prognosis - it doesn’t make it better and it can work worse. There may be a combination of factors resulting in the over-performance of the control arm, specifically in these high PD-L1 patients. Therefore, the conclusion we make is that pembrolizumab is a valid option across all subgroups of PD-L1 patients.
Oncology Frontier: You just made the presentation about the application of ctDNA in the adjuvant setting. So, If PD-L1 cannot be used as a biomarker for adjuvant immunotherapy, will ctDNA be potential, or do we have other available biomarkers?
Dr Peters: Thanks a lot. There are plenty of trials, as I discuss in my talk (and I am happy to share those slides with whoever writes me an email). I did review what we have. We use ctDNA in the specific context of what we call minimal residual disease (MRD). Most of the time, you use it after surgery to try to classify patients between having residual disease or not by looking in the blood. Most of the tests we use today (but not all) are what we call tumor-informed, meaning that at the time the tumor was present, there was a whole genome sequencing performed and compared to the blood to try to see the signature of the tumor. Later on, you try to look for this signature down the road. It can be a set of genes that are known or a personalized signature of specific mutations of that tumor. But that is how we do it. We try to follow through so we know if the patient still has the molecular signature after surgery. We have a dozen trials showing that if you try to understand what happens to the MRD-positive versus MRD-negative, then this will be prognostic. If you are MRD-positive, you do worse, that is for sure. There are two or three things we still haven’t shown. The first thing is what methodology should be used. The second thing is how to improve sensitivity, because in all of the trials we have been seeing, some patients who relapse were MRD-negative, meaning we are not able to detect MRD in many patients. That is a big problem. Third is whether this is predictive. If you detect MRD and you add-on treatment, are you able to revert the course of disease? Are you able to improve survival, or is it too late, and whatever you do is not useful? I take the example of IMpower010. In IMpower010, when they had done surgery and before starting atezolizumab, they did MRD with a very nice technique. When we look, the MRD-positives do worse than the MRD-negatives. When looking at the MRD-negative, a lot of them relapse, almost 25%. So we suppose they are not actually MRD-negative. Importantly, atezolizumab still has the same benefit in these MRD-negative patients as well as the MRD-positive patients, meaning that this is not predictive of additional treatment. Prognostic, yes. Predictive, still not. So to answer your question, we need better sensitivity with the establishment of a predictive ability. This means it is now not ready to use. We need to have more large prospective trials to find the ways to use this. It will not take a long time. It will take months or years, but it is not ready to use yet.
Oncology Frontier: The KEYNOTE-091 study has no clear requirements for pembro and chemotherapy cycle, only limited to ≤18 cycles and ≤4 cycles. As you participated in another discussion, immunity for lung cancer: the longer the better? The more the better? How do you view the course of adjuvant immunotherapy?
Dr Peters: That is the one-million dollar question, right? Is more better? Usually, we should not think this way. The reason why we think this way is because most of the clinical trials are driven by the pharmaceutical industry, and of course, there are always marketing issues. Usually, more is not better. We need to deliver just what is needed. Today, we have a dilemma. You can give a short course of neoadjuvant chemo-IO based on CheckMate 816 - three cycles of chemo-IO, then surgery, and its done. But you can also give surgery and one year of immunotherapy, which takes way longer. What we will discuss tomorrow will be all of it - neoadjuvant chemo-IO plus one year of adjuvant therapy - a maximal therapy regimen. Which one is better - neoadjuvant, adjuvant or perioperative (the whole cycle)? We don’t know. I am not sure we will get a very clear answer, because we are not supposed to compare, but in time, we will have so many datasets that I hope we will be able to decipher which is the most convenient (in terms of sustainability, cost, and so on) and also in terms of potential really significant differences in outcomes. Right now, we cannot say. We have to leave this to the regulatory environment and the investigators. But I am not sure in the end that more will be better. I think we will need to subdivide patients into risk categories. I come back to your question about ctDNA, at some point we might do ctDNA to know if we need more. And this is really where we need ctDNA.
Oncology Frontier: The KEYNOTE-091 study has reached the DFS primary endpoint. Do you think it is possible to reach the OS key secondary endpoint? Is it possible that OS tailing of immunotherapy will appear in early treatment?
Dr Peters: All of these immunotherapy trials are challenging, because we don’t know if they will lead to an overall survival benefit. We have been seeing overall survival early curves from IMpower010, and there are signals that OS endpoints could be met in the future. These results need maturity, which will take 2 or 3 more years, but I expect that the immunotherapy strategy can lead to an OS benefit. So I will say wait for a little bit, and look at the IMpower010 as a signal of OS being reached that could lead to an OS benefit. The question is, is this in the ITT population or only a subset? We will have to wait for more data. I would guess that for immunotherapy, as you said, based on all immunotherapy trials, the impact is greater on OS than on PFS - the tail of the curve. If we wait long enough, I hope this will lead to the same phenomenon.