Round 2: Are We Adequately Targeting the TME in mCRC?

Watch Round 3 here.

Aparna Parikh: Next, we'll discuss with my colleagues here today whether we are adequately targeting the tumor microenvironment in metastatic colorectal cancer.

Nataliya Uboha: I'm on the affirmative side of the argument, though I think a lot of good information will come out during our discussion. As we discussed in the first session, several studies support the use of multitargeted tyrosine kinase inhibitors in colorectal cancer, and several of these agents target the tumor microenvironment. Regorafenib is a multitargeted tyrosine kinase inhibitor that targets kinases involved in VEGF receptor signaling and has demonstrated activity in patients with heavily pretreated colorectal cancer in a randomized phase 3 study. Fruquintinib is the most specific inhibitor targeting VEGF-R1; two and three have also demonstrated activity in more recent studies. In patients treated with more modern therapies, FRESCA was a phase 3 study that randomized patients to Fruquintinib or placebo, and we observed improved overall survival with Fruquintinib.

And of course, we have been using Bevacizumab in combination with chemotherapy for many years during different treatment lines for patients with advanced CRC. And again, it's important to note that Bevacizumab has limited activity as a single agent. So while these drugs work, they have modest activity on their own. Bevacizumab is combined with chemotherapy, but it does have activity in patients with this disease. And many mechanisms have been proposed for how these agents help treat colorectal cancer.

Currently, we don't have clear biomarkers for patient selection. We offer these agents to patients regardless of biomarker status, and all have been shown to improve overall survival in patients with this disease. One of the studies to note regarding Bevacizumab is the SUNLIGHT trial, a phase 3 study that randomized patients to Lonsurf alone, chemotherapy alone, or Lonsurf plus Bevacizumab and showed improved overall survival with the addition of Bevacizumab in a late-line setting. So again, supporting the fact that we are: when we target the human microenvironment and add anti-angiogenic agents, even in a heavily treated patient population, we are seeing benefits.

Kanwal Raghav: Nataliya, I think I would disagree with that a little bit, not because the evidence is incorrect, but I do think that the issue is a little bit more complex than has been presented. As CRC advances and metastasizes, I do feel that the tumor microenvironment, including the immune environment, does become spatially and temporarily complex. We know that liver metastases are particularly hostile environments for immunotherapy. Some data show that liver metastases respond poorly to immunotherapy compared to, say, lung metastases. And we know that refractory tumors are characterized by T-cell exclusion, myeloid-mediated immune suppression, and stromal barriers that limit effective immune trafficking. MSS disease is also not biologically uniform. Distinct microenvironments and phenotypes exist within these tumors. They change as treatment progresses. Some of these physiologies do have an immune desert or immune-excluded state, so I'm not quite sure how immunotherapy would add to achieving responses in that setting, but it does reflect the heterogeneous pattern in which not just the tumor but also its microenvironment and the immune microenvironment evolve over treatment.

Furthermore, the fundamental barrier to immunotherapy response in refractory disease is truly the absence of a T–cell-inflamed microenvironment. Even in first-line or earlier-line settings where treatment has not had a detrimental effect on the immune system, you don't see the responses we hoped for, with multiple trials showing no clinical activity. Most later trials do enroll this broad MSS population. So I agree with you that we are not stratifying appropriately. We don't have any biomarkers. And while VEGF blockade has been shown to have this effect so far, despite VEGF being on the landscape for a long time, we have not identified any biomarkers that allow us to select patients who would respond better to VEGF blockade. And this stratification by immune contexture, hypoxia, and stromal phenotype is critical, but it may dilute the subset. A small proportion of these patients are immunoresponsive, but we won't ever reach that proportion if we keep doing unselected trials.

And the current data shows that, at least as far as immune checkpoint inhibitors are concerned, that subset is very small, if any. Furthermore, when you look at hypoxia and RTK-driven vascular signaling, they can promote immunosuppression, fibrosis, and adaptation, ultimately leading to tumor progression. So, I think that if angiogenic modulation would explain everything, it would, again, lead to deeper and more durable responses, which we have not seen, at least clinically. I also do feel that there is a dynamic evolution that influences these responses. Clearly, we've seen data from the cricket and Kronos studies: whenever you detect KRAS mutations in circulating tumor DNA, you have inferior progression-free survival. And even in Kronos, where you excluded patients with resistant mutations, you found increased responses. So, I think all therapies modulate the tumor microenvironment to some extent, but they may not always be favorable to an immunotherapy strategy.

Aparna Parikh: So thanks again so much for that great discussion. I think we've heard some great points and counterpoints regarding the ability of various antiangiogenic and RTK-targeted strategies to consistently modulate the tumor microenvironment. On the affirmative side, we have heard discussions that we do consistently see survival benefits across refractory colorectal cancer patient trials. On the skeptic side, we have seen discussion and data suggesting that without consistently integrating various tumor microenvironment biomarkers into clinical trial designs, we really are probably overestimating the extent of TME reprogramming.

So, in terms of where we go as a field going forward, I would love some thoughts on possible measurable biomarkers that could consistently demonstrate reprogramming of the TME. I think we've seen attempts at this with the immunoscore. We've seen attempts with TMB and MSS patients as a surrogate for some possible signal of what tumor types may be a little bit inflamed. This is particularly for immunotherapy efficacy, but I’m curious if either of you has any thoughts on measurable biomarkers we could use to demonstrate that we are having a lasting impact on reprogramming the TME.

Kanwal Raghav: The field of data mining for biomarkers has been rapidly evolving for some time. We have relied on a very binary look at the cancer. It's either sensitive or resistant, but now we are slowly starting to appreciate that there are transitional states between those two extremes, and how tumors adapt to treatment changes is as important as what you see at the end. So, I do feel that the biomarker field needs to go beyond what we can easily measure, because your resistance mechanisms may be more driven by genomic biomarkers than by the ones we tend to measure with, say, circulating tumor DNA or baseline biopsies. So, incorporating treatment biopsies and a transcriptomic and metabolomic perspective might help us better understand this.

Nataliya Uboha: I was silent at the beginning because I couldn't think of a better biomarker than a paired biopsy, looking at what's happening under the microscope in the tumor microenvironment, but these are so tough, especially for patients with advanced disease. But as we know, especially in heavily pretreated patients, the tumors are heterogeneous. One side may not have the same composition as the other side. So, really looking, I think paired biopsies are the best way. I mean, the other field that is evolving is with the functional imaging, the PAT imaging, radiopharmaceuticals, that may be another way really to look at what's happening as a surrogate with these new, both theranostics, radiopharmaceutical theranostic therapies, but also the different imaging modalities to look at what's happening in the tumor microenvironment without invasive procedures. I was wondering what you both think about those modalities entering practice?

Aparna Parikh: No, I mean, I think we've had some early experience with the FAPI PET. And I think FAPI, for example, is, as you mentioned, advanced molecular imaging to really look at activated fibroblast. And this was initially for diagnostic purposes, but now we're seeing FAPI-targeted approaches therapeutically as well in GI cancers. So, I think some novel approaches to better assess the baseline and post-treatment tumor microenvironments, and then to therapeutically target the TME as well. So I think FAPI is just one example of that, and we're seeing a lot of buzz, I think, in the theranostic space and GI cancers, which is great.

Any other final comments? I think the take-home from this section is biomarkers, biomarkers, biomarkers. I think binary response and lack-of-response assessments aren't going to really help us understand what is happening better. And so, encourage us all, as a field, to really think about what we'll need from the biomarker space. And I'm with you, Dr Uboha, I think if we could get paired biopsies on every patient, and on many patients, I think it would be informative but challenging to do. So, look forward to not only imaging but also to more advanced blood-based biomarkers that might help us tease out what is happening across metastatic patients.