ROBERTO SALGADO:
Dear colleagues, dear friends, many thanks for attending this special session on breast cancer, and we have the pleasure here to have a real expert, Marleen Kok, from the Netherlands Cancer Institute, a world-renowned breast cancer oncologist. And she has an amazing talent for innovative trials, and that’s actually what we’re going to talk about right now. There is this emerging topic, Marleen, about deescalation therapy. We heard it at several meetings already this morning here that ESMO Breast in Berlin. What is exactly deescalation therapy?

MARLEEN KOK:
Yeah, good question, Roberto. It’s kind of a scary terminology, right?

ROBERTO SALGADO:
Of course, yes.

MARLEEN KOK:
In oncology. What we in general consider deescalation is giving less systemic treatment, or even less surgery or less radiation. So it can apply to all kinds of modalities, but as a medical oncologist in general, we are talking about giving less systemic treatment, and it usually applies to early setting, because in a metastatic setting, there’s another goal for a patient living longer. And in the early setting, yes, living longer, but also with less toxicity.

ROBERTO SALGADO:
Interesting, interesting, because we heard several lectures this morning about potential use of deescalation in triple-negative breast cancer, which is actually what we know, one of the most aggressive subtypes in breast cancer. And still there, we tend to think, let’s give less chemo.

MARLEEN KOK:
Yeah.

ROBERTO SALGADO:
How dangerous would that be? Can you explain, what is the evidence already that we can identify a subset of those patients who maybe we may be overtreating? That’s what you’re saying, right?

MARLEEN KOK:
Yeah, yeah, yeah. Coming back to your point on aggressiveness, I don’t agree that as a total, the TNBC is aggressive. If you look at a longer term outcome, yes, the TNBC patients, if they recur, then it’s usually in the first couple of years, first two years or very early, which is of course dramatic. But if you look at the long term, the outcome is as good or as poor as with ER-positive breast cancer. That’s, I think, good to say. And then, maybe I’m a bit a teacher here, but also coming back to the definition, triple-negative breast cancer, right, it’s a poor definition, meaning we measure three things, and if you don’t have it, you come to the basket of the rest. And I think that also explains the heterogeneity because, you see, the outcome really depends on subtype of breast cancer, and for example, whether you have immune cells in the breast tumor, and of course, on tumor size and lymph node status. Very heterogeneous disease.

ROBERTO SALGADO:
That’s extremely interesting, because we are now in the era of immunotherapy, certainly triple-negative breast cancer, certainly when there is an increasing tendency to treat this patient before surgery. Chemotherapy and immunotherapy. And now you’re saying there might be a subset of those patients with a lot of immune cells already present, even if you didn’t treat them before with anything, they have already, intrinsically, a lot of immune cells present.

MARLEEN KOK:
Yeah, yeah.

ROBERTO SALGADO:
And just to raise awareness to the patients, because they might get scary to say, “Will I really get less chemotherapy than other patients?” What is the evidence already in, for example, what we tend to call the very aggressive subtypes in BC in the very early, young age group of our patients? Is there any evidence that the amount of infiltrate of immune cells, the strength of the immune system, is actually identifying a subset of this patient, might do excellent?

MARLEEN KOK:
Yes, yes, there is, and thanks to pathologists like you, Roberto. If you go back in history, and I think you’re more aware of that than I, there is this subtype of breast cancer reported, I think, already in the ’80s or ’90s, lymphocyte-predominant breast cancer. So, it’s not really a novel finding, it’s more that now that we have treatment options available, we are able to subclassify and sub-group, and thereby, I think, this feature comes back, and indeed, so already present. So, what at least I see, and I think many people agree, it’s a readout for an endogenous immune recognition of the tumor. Basically, the immune system is already busy trying to get rid of the tumor, but it was not successful, otherwise you don’t get the diagnosis of cancer. And yes, for the young patients, because that’s especially the group where you are even more scared of a relapse, there have been first data in the area of chemotherapy, so if you have high number of lymphocytes, or we call them TIL, tumor-infiltrating lymphocytes, high level, you’re way more likely to respond to chemotherapy. And maybe, instead of four chemos, you can give two, so that’s the first hint, but very important, and thanks to the work of Sabine Linn, a professor at my institute in NKI in Amsterdam. She had this vision to go to the Dutch Cancer Registry and collect data from TNBC patients of younger than 40, treated in the past, because, if you can recall, chemotherapy used to be, when it was introduced in breast cancer for lymph node-positive disease, lymph node-negative disease, and that’s exactly what we’re talking about. TNBC, which you know, lymph nodes not involved, where we have a good prognosis, and there you can think of deescalation, in that group, she collected over 500 cases were tumor blocks. And pathologists, thanks to you, Roberto, looked at these TILs, and remarkable results. So, in that group, lymph node-negative of younger than 40, if you have high TIL, your prognosis is excellent, even on overall survival, long term outcome. And that’s really questioning if this patient comes to me like tomorrow, my clinic, and she has high TILs, and she’s young, and I will prescribe her anthracycline four cycles, carboplatin, paclitaxel. That’s over treatment, if we now see that without chemotherapy, they did excellent.

ROBERTO SALGADO:
Interesting. In order to have more convincing evidence, because I do understand clinicians and also patients, that they may still be a bit reluctant to use this concept of the immune infiltrate, mostly because it’s morphology, it’s what you see through a microscope. So, it’s not a genomic essay. An oncologist told me a few days ago, “If you would quantify the immune system with a genomic method, we would already use it in clinical practice and there would be five to six clinical trials already running, but because it’s morphology, people become a bit reluctant.” So, my question to you is to convince the patient that indeed, there might be a substitute to what patients call, not deescalation, but optimizing therapy.

MARLEEN KOK:
Yes, optimization. Yes.

ROBERTO SALGADO:
What are the future trials that are now being developed to have this extra additional information to convince gleesome patients that effectively, if we give you less toxic treatment, you do actually as good as before.

MARLEEN KOK:
Yeah. Yeah, I think, first of all, there’s a lot of discussion going on in the fields.

ROBERTO SALGADO:
Which is good.

MARLEEN KOK:
Which is good, to really pinpoint to what evidence do we have and what is the right trial to do? I think there’s not only one trial that we can do, and I think, if you look at all things that made it to the clinic, there are several trials that provide accumulating evidence.

ROBERTO SALGADO:
Interesting.

MARLEEN KOK:
But first of all, we need to assess the TILs, and not only in a clinical trial, but really when we set a diagnosis, because if it’s really in a standard report, it’s way easier to select patients for a trial.

ROBERTO SALGADO:
Indeed.

MARLEEN KOK:
But maybe not for all countries and all centers that will be possible, because you need a pathologist that’s also willing to do so. But first of all, I think, for those patients with super high TIL, the cutoff that can be used, about 50%, even 75%. So, give no chemotherapy or less chemotherapy, and maybe the less chemotherapy is more appealing, so do it step by step. And then, the discretion is focused on, do you need to randomize? Because if you do a randomized trial your numbers you need, huge. And I think the majority of the people is not convinced that a single-arm trial for TIL-high patient would be an option.

ROBERTO SALGADO:
Okay, but that would be a phase two trial, or would that be two?

MARLEEN KOK:
Single-arm, phase two. But we have examples, for example, in the HER2-positive space, which is an example, I think, because HER2-positive used to be aggressive, but now we deescalated the anthracycline treatment, for example, and now we are discussing in that field, can we really leave out the chemotherapy for selected cases? So, the HER2-positive is an example, and in HER2-positive, we now prescribe a little of chemotherapy, paclitaxel, or trastuzumab, based on a single-arm phase two trial.

ROBERTO SALGADO:
Okay.

MARLEEN KOK:
Another trial that’s important, and it’s a debate whether it should be done for all-comers or select based on TIL, is leaving out anthracyclines. Remember, the anthracyclines were there first, and it’s the backbone, so we put things on top. We put the taxanes on top carboplatin, pembrolizumab, and now the PCR rate, it’s 65%, but we don’t know the contribution of this one. And this one is most toxic. So, what will happen if you save the anthracyclines, start with the first three compartments, and if you don’t have a PCR, then give the anthracyclines, as kind of like a selfish treatment?

ROBERTO SALGADO:
That’s intriguing. Intriguing because you always hear back and forth, this argument, that before they introduce a biomarker in daily practice, you need to have the highest level of evidence, which is level 1A, prospectively randomized trials, biomarker-driven. And what pathologists tell me, and I’m a pathologist myself, for many prognostic biomarkers that we use in our daily practice already today for decades, which have been in guidelines for decades, we don’t have this level 1A evidence, because we use them combined, and that gives the clinician an indication about the risk of recurrence, the clinical high-risk profile of their patient. Do you think that we should use the TILs today, already in such a concept, namely, not as a binary variable, but together with all the other variables informing the clinician, “This is the risk profile of your patient,” and without formally stating, “Now you can safely omit chemotherapy,” but just to inform the clinician that if he has a patient in front of him with a lot of comorbidities, that he might say, “Okay you have a high number of immune cells, you will do excellent, so we might give a little bit less toxic treatment.” Is that pragmatic and feasible, or is that already a dangerous evolution? What do you think?

MARLEEN KOK:
Yeah, it’s good that you emphasize that the TILs is not a standalone as it is right now. So, in the tumor boards, we make the decision based on tumor size, lymph node status, age of the patient, comorbidity, et cetera. So, the TILs is not a standalone, but it’s an addition to what we have to better tailor prognosis. And you’re right, in those difficult cases, with comorbidity or patients at certain age, or some patients really have this wish not to be treated with chemotherapy, it can help you. So, it’s not like a decision, yes or no, it contributes. May I ask a question to you?

ROBERTO SALGADO:
Of course, of course.

MARLEEN KOK:
Because you ask me, “Can we use it?” But the discretion that usually is among medical oncologists, “Is it reproducible?” So, how reproducible are TILs? If you score it, and then a pathologist in U.S. and in Asia scores the same slides, so can you tell us about that?

ROBERTO SALGADO:
Well, if you compare, and that’s the thing about being pragmatic, most of the biomarkers that we analyze in our daily practices are estimates. Estrogen receptor, progesterone receptor, HER2, Ki67, great, atypia, lymphovascular tumor invasion, they’re all estimates by the pathologist. And the pragmatism in that context is, it’s good enough for clinical practice. If you read the concordances of all those variables, it’s always moderate to good, rarely excellent. And if you compare the concordance analyses on TILs done by powered studies by the TILs Working Group, it’s even better than most of these other biomarkers. And it gets even better if you train pathologists. Now, because you asked me this question, I take this opportunity to throw out to the community a concept, which I think is crucial, if we want to implement biomarkers in daily practice. Namely, how has it been done so far? You have the pharma industry, and which we have excellent collaborations with. They do the phase three clinical trials with the biomarker, they push a button, and then everybody expects that all pathology labs in the world will be able to implement a biomarker. And we have seen with PD-L1 that it doesn’t work. That you need to train and retrain and develop preference materials to help those pathologists assess that biomarker. So, the only way, I think, that we can introduce the TILs safely is that we start today, systematically adding the TILs in the pathology reports.

MARLEEN KOK:
Exactly, exactly.

ROBERTO SALGADO:
That we continue to develop preference materials and training materials for the pathologist. And that’s the message, which I think is an important one for the community, that we start to use local testing for trial inclusion, and not central testing. Because, how does it work in in big pharma-driven trials? You have a serum, with just a few pathologists who do the biomarker analysis for that trial, and then the trial gets positive and everybody assumes we will all be able to find those results by those limited number of pathologists. And we have seen, that’s not the case. And the only way in which we can enhance the concept of biomarker introduction is if we use local testing by trained pathologists.

MARLEEN KOK:
Maybe I can share the experience we have at NKI.

ROBERTO SALGADO:
Please.

MARLEEN KOK:
It’s a pleasure to really work in a research institute, because we try and we do the following. If a pathologist that is not well trained on TIL scoring, but wants to use the TIL scoring, or when a patient seems to be eligible for a trial where TIL scoring is required, we do the following. The H&E slide that’s necessary for TIL scoring, it’s scanned and uploaded to a certain software system, and of course, the pathologist in charge can provide a TIL score, but maybe that pathologist is not really certain. But then an expert panel of pathologists really well trained also provides a TIL score within 48 hours. So then, the starting pathologist will have a reference, and thereby is kept trained on the spot, and also thereby, we can already run a trial with TIL as entry criteria. It’s the BELLINI trial. It’s a small trial, but it’s a start to get experience with using TILs as entry criteria of vital importance for deescalation purposes, and also for future IO trials, probably.

ROBERTO SALGADO:
Excellent, excellent, I think I read that manuscript by Nature “NPJ Breast Cancer,” which I strongly encourage the community to have a look at it because it’s a new concept, which I think will solve a lot of issues in the field of biomarker implementation in daily practice. Many thanks, Marleen, for this excellent discussion.

MARLEEN KOK:
Pleasure.

ROBERTO SALGADO:
I learned a lot, once again,

MARLEEN KOK:
Thanks, Roberto.

ROBERTO SALGADO:
and thank you all for your interest and patience. Thank you.

MARLEEN KOK:
Thanks.

Recording date: 04-May-2022; Feature publication date: 20-May-2022.

Roberto Salgado reports no relevant disclosures.

Marleen Kok reports no relevant disclosures.