Optimizing the Care of Patients with Localized Colorectal Cancer (CRC) — Jeffrey Meyerhardt, MD, MPH

DR MEYERHARDT: Thank you for inviting me today. I’m Jeff Meyerhardt. I’m a medical oncologist that specializes in GI cancers at the Dana-Farber Cancer Institute in Boston, and I’m going to spend this time reviewing some concepts that are emerging and important in treating patients with early-stage or nonmetastatic colorectal cancer.

So the first few studies that we’ll review are really focused on circulating tumor DNA, which has really been of great interest for several years now, and there’s a lot of different platforms to be able to measure circulating tumor DNA in measuring what people describe as molecular residual disease or MRD, and it’s a very powerful tool, and the question is, is what can we do with that powerful tool and what settings it should be applied in, and which patients and most importantly how could it inform us regarding how to treat our patients.

So I’m going to first look at this trial. This is the Circulating Tumor DNA for MRD in Patients with Stage II and Stage III Disease, a Final Analysis of the BESPOKE CRC cohort. And so the BESPOKE effort is an observational prospective, observational study using 1 particular informed assay to understand both how that affects prognosis and potentially how it informs treatment decision in providers of patients with Stage II and Stage III colon cancer. So the way this is set up is, you can see here the different time points of when blood is collected in patients who had colorectal cancer. We’ll focus particularly in the after-surgery time period, where there was blood collected somewhere between 2 and 6 weeks after surgery, somewhere around 6 plus or minus 2 weeks after that, and then 12 weeks and 20 weeks, and then a surveillance program, so multiple time points collected. There’s also questionnaires that were sent to providers and part of the questionnaires is asking how did it potentially inform their decision making. And then there’s patient reported outcomes that were also collected in these patients.

And what we can see from these Kaplan-Meier curves is what we know really from most trials and we’ll see pretty replicated in several of the trials that we’ll focus on today, is that having positivity of ctDNA postoperatively is very prognostic in terms of disease-free and ultimately also overall survival. And so you can see here, in patients with Stage II disease compared to Stage III disease, a hazard ratio greater than 10, 11.23 for Stage II disease and 8.3 in patients with Stage III disease. So again, a very prognostic marker in terms of either recurrence or death, which is the definition of disease-free survival.

And what we can also see is that, if you look at the first surveillance time point, so again after receiving adjuvant therapy, that also is very predictive of outcomes. So people who either were negative and became positive, or positive and remain positive, is also significantly associated with worse outcomes, with in this case a hazard ratio of 20, so very significant effect.

What we learned from this is a few things. One is, again, they asked physicians how this affected their treatment decision. And what we learned from that data is about 16% of those decisions, on what to do about adjuvant therapy, were influenced by the data. It doesn’t mean, again, that it affected their outcomes or predicted the response of therapy, but 16% of the providers who participated in BESPOKE with their patients, changed their decision either escalating about a third of the time, or deescalating about 60% of the time, their adjuvant therapy, based on the results of the ctDNA. We also learned if you clear ctDNA, that’s associated with improved outcomes. And what we learned is that those who were positive, particularly Stage III patients, who received adjuvant therapy compared to observation, had an improved disease-free survival, and that those who were MRD negative, so after surgery, who did not, who received adjuvant therapy, didn’t seem to have a significant benefit in terms of outcomes. But I think the important thing, and this is really the issue with the field overall, is these are observational studies. So what, though we learned that people may have changed their decision, but what really influenced giving some patients adjuvant therapy is undoubtedly confounded. And so you can’t really take the data and say that, because the positive people who got adjuvant therapy seemed to do better, that definitively says that that should define who should get adjuvant therapy, because that decision for therapy is not controlled and it’s not a randomized study. But again, certainly suggestive data of the potential benefit of adjuvant therapy in positive patients.

This is a second study by a group in Japan, that looked at MRD and survival in resectable colon cancer. And again, very similar results that we saw with the other study, that both, at the MRD window, so after surgery, highly predictive of disease-free survival, very significant hazard ratio, 11.99, and similarly a highly statistically significant overall survival. And when you look at surveillance you can see, again, that having positivity after the end of treatment, in first surveillance, is also highly predictive of outcomes, both in terms of disease-free and overall survival.

They looked at how can we think about this particularly with certain molecular alterations. And this is looking at BRAF V600E, which we know from other studies that patients who have early-stage disease, that have a V600E mutation, do have a worse outcome. And you can see though, if they are ctDNA-negative, they still have some recurrences. There’s clearly recurrences that are occurring within the first few years after surgery. But those who are ctDNA-negative have a much better outcome than those who are positive after surgery, that unfortunately have a very significant drop in their curve. Now, I think the other thing to note is these are subgroup analysis. This is an observational study. You can see that the ctDNA-positive patients is only 11 patients. So it’s a very small number of patients in this endpoint, but again, suggestive that this is important in multiple different type of mutations, including those who are MSI high, those who have certain KRAS mutations, those who have ERBB amplification with HER2-amplified tumors, et cetera.

And what we also saw, which was interesting, a little expanded upon in this study, is that if you take patients who were positive at the MRD timepoint, so soon after surgery, and those who did not clear their ctDNA, those patients had a significantly worse outcome. Those who sustained clearance, so looking at surveillance endpoint, seemed to do much better. And then there is an intermediate group that has transient clearance, so became negative and eventually became positive, who do relatively better. But ultimately again, it’s defining those patients as undoubtedly having metastatic disease, that whether detectable or not initially, will ultimately have metastatic disease, and that also translates to overall survival.

And the other thing to know about the GALAXY trial is it was a mix of Stage I to IV patients. The majority of this cohort did not receive chemotherapy. Again, it’s highly prognostic of outcome. Those who received adjuvant therapy, who were ctDNA-positive, which is really a small number, 181, had actually received adjuvant therapy, 38 sustained clearance, 31% transient, 31 no clearance. So it does suggest the chemotherapy for some people seems to lead to clearance, again, as was shown, a good portion of those patients will remain disease free but not all of them. And those who get no clearance, will have a significantly worse outcome. And again, it’s not a decision — it’s an observational study. And again, the issue of who actually receives chemotherapy is not controlled in this setting.

The next study will look specifically at a subgroup within the CIRCULATE GALAXY cohort and these are patients who had resected colorectal liver mets. So this study, which was published in November, looked at that subgroup of patients. And, again, a very similar — now these people, again, because they already had metastatic disease, will have a worse outcome almost by definition. So as opposed to when we saw for the earlier-stage patients, where ctDNA-negative patients seem to have what looks like remaining disease free long-term in the 85% to 90% range, you can see these patients in the liver MRD group do seem to have, even if they’re negative, disease-free survival of 2 years of only about 70% to 75%. But again, if they’re ctDNA-positive after resection of liver mets they do worse and similar results that we see for surveillance.

If we look at potential predictive value, again in an observational setting, of adjuvant chemotherapy in this setting, and I will make sure it’s noted that you can see from these Kaplan-Meier curves and the number at risk, that the numbers are pretty small here, right? So these are patients who were ctDNA-negative or ctDNA-positive. The ones that were negative, who received adjuvant therapy compared to observation, doesn’t have a significant difference, though there is an absolute difference. And there is some separation of the curve. So, again, in the population had resected liver metastases, receiving adjuvant therapy, not statistically better but does seem to have some sense of some improvement. And those who cleared seem to have a much better outcome, though only a small portion of them cleared. And you can see the at risk in the bottom left Kaplan-Meier curve, is a very small number of patients. And for those who were positive, again they do worse in terms of overall outcome either way, but those who receive chemotherapy seem to do better and that hazard ratio actually is 0.07, which is sort of a little unheard of in most medicine. But again, the numbers are very small. But clearly, the patients who are positive and only are observed, unfortunately have a very poor outcome. But again, that is likely confounded for some of the reasons why they did not receive adjuvant therapy.

And so there’s 190 surgically resected patients, none of them had gotten preoperative therapy, 32% were MRD-positive. Only a quarter of them actually received chemotherapy, which I think for most of us practicing, where you had a patient who had liver metastases that were resected and look on imaging to be disease free would be providing them adjuvant therapy. So it’s only a small number. And, as I pointed out earlier, though the MRD negative patients, the difference between chemotherapy and observation was not statistically significant. There’s still actually a 10% difference of 2 years. And so it should be certainly a consideration, those patients as well.

Our group presented these data from the Alliance CALGB/SWOG 80702 study at GI ASCO this year. And the study itself, just to give a background, it was a randomized clinical trial that compared 2 questions, 3 versus 6 months of adjuvant FOLFOX therapy, specifically in Stage III colon cancer patients, and celecoxib and placebo, that was started at the time of adjuvant treatment. The study itself, primary endpoint was the celecoxib question. The hazard ratio was 0.89. We published it several years ago in JAMA. We have since looked at several molecular biomarkers related to potential subpopulations that may benefit. We collected tumor blocks and blood in many of the patients and we did whole exome sequencing in over 1,000 patients and we matched their blood ... and we thus did an informed ctDNA assay in that population. So it’s about 960 patients that are included in this analyses. What was shown here is that, very similar to other results, and this is looking at MRD, so soon after surgery, before they started receiving adjuvant chemotherapy, the patients who were ctDNA-positive had a significantly worse outcome. Again, we keep seeing these hazard ratios in terms of a prognostic value of 7, 11, 13. So it’s really a very prognostic marker in patients who have Stage II and Stage III colon cancer. And that also translated into overall survival. But then, when we looked at the subgroup of patients who received celecoxib versus placebo, we saw a result that frankly was a little unexpected. What we see here, just to get a sense, is the more solid line, are those who were ctDNA-negative, again at the time after surgery before starting chemotherapy. And you can see there’s really no difference between the patients who receive celecoxib versus placebo. But those who were ctDNA-positive, you can see there’s clearly a separation in the curve, and the hazard ratio for that subpopulation, 0.55, was statistically significant, with a significant p-value.

Now, the caveat of these data are that when you look at an interaction, and the way interactions work, which I know is not necessarily always intuitive, you’re basically — when you’re looking at associations between 2 things, 2 subgroups, you’re basically creating curves and lines and you’re really deciding how different those 2 lines are, and is it enough of a difference to say those 2 subgroups really have something different happening to them. So in this case, is the group that are negative versus positive, is the differences we’re seeing between celecoxib and placebo just by chance or there truly are different subpopulations that are defining the relationship or association differently? And so there’s not a statistically significant interaction between those 2 groups, and this is looking at overall survival, which again shows a significant difference in the subpopulation. But the interaction term, 0.2, is not statistically significant.

So these data, again, support the highly prognostic value of ctDNA. Celecoxib seems to benefit patients who are ctDNA-positive, but the p interaction was not significant. There is a benefit in patients with, the data I didn’t show here, is we looked specifically at a subgroup of patients who are PIK3CA wild type and who are mutated, and that benefit is seen in both of those subpopulations. And I’m going to talk a little bit more about what is PIK3CA, why it’s important, in a later study. And again, the data are suggestive but not definitive for using celecoxib when patients are ctDNA-positive.

One other study that we’re going to look at, focusing on ctDNA, is this study, the DYNAMIC study from the group at — Peter Mac in Melbourne, and this is looking at the 5-year updated results of the DYNAMIC trial, and again that was specifically in patients with Stage II colon cancer. So the way this study worked is, it was patients who had Stage II colon cancer, they underwent a resection, they had staging scans within 8 weeks. They had a tumor informed ctDNA analysis done at week 4 or week 7, and they had bloods drawn at those 2 endpoints but then they were randomized to either using the results of their ctDNA to guide management or doing what you would standardly do. So the way we all make decisions on who receives adjuvant therapy for patients with Stage II colon cancer, where we use, if it’s a T4 disease, if there’s perforation or other clinical factors, to determine who should receive adjuvant therapy. But those who were CT-guided management, the recommendation was, for those who were ctDNA-positive either at week 4 or week 7, to receive adjuvant therapy, either fluoropyrimidine base or oxaliplatin based or, for those who were ctDNA-negative to have observation. And then they were followed by the way we standardly follow patients in surveillance.

So using the ctDNA results to determine whether you receive adjuvant therapy or not, versus what you would do clinically, there was no difference in terms of recurrence-free survival at 5 years between those 2 groups. And similarly, there was no difference in overall survival or disease-free survival. What we can see is that the patients who were ctDNA-negative versus ctDNA-positive, you can see that there is, again, a prognostic value to ctDNA, where the people who were positive, and again, some of those would have received adjuvant therapy either because that was the recommendation because they were using ctDNA-guided management, or they were ctDNA-positive because they had just clinical management. You can see that those people do worse if they’re ctDNA-positive, both in terms of overall survival and then if we specifically look at a T3 versus T4 for those who were negative, they again are prognostic in terms of what we know clinically, that a T4 disease does worse even if you’re negative. But the ones who are T3, ctDNA-positive do quite well.

And so it was noninferiority design for a standard of care to ctDNA management, which can either be interpreted as using MRD to determine whether you received adjuvant therapy, doesn’t change outcomes, because at the end of the day the group that had ctDNA management did the same as if you didn’t get the ctDNA, or you could interpret it as by checking MRD we use less chemotherapy in patients with Stage II disease. The issue is it actually is not that much less chemotherapy. In fact, the less patients on the ctDNA management actually received chemotherapy as we expected but they actually — more of them got oxaliplatin because they were ctDNA-positive, a small percent, 9.5 versus 2.7, of the overall population. But again, using the ctDNA management, they overall got less chemotherapy, but more of them got oxaliplatin, which is honestly the chemotherapy that can be more problematic for patients because of the neuropathy. And the median time to start therapy actually increased by about a month. And that is some of the issue with checking ctDNA in these patients, to use it for a decision regarding adjuvant therapy is you have to obtain the tumor, you have to obtain the blood, you have to send it in, it has to get analyzed. And so you can see the median time from surgery for those who did receive chemotherapy in the standard no-ctDNA-checking arm was 53 days, and those who had ctDNA guiding management it was 83 days for those who received chemotherapy.

This is the last topic I’ll talk about related to ctDNA, is the current ongoing NRG study. So the NRG study is actually asking 2 questions. So it’s going to take patients with high-risk Stage II and Stage III disease who had a resection, and ctDNA’s going to get checked. And those who were ctDNA-negative are going to be randomized to a potential de-escalation plan, which is either receiving what we would standardly offer these patients, which is capecitabine/oxaliplatin or FOLFOX, or half surveillance with serial ctDNA measured, and if they became positive then go on to a randomization of then receiving chemotherapy or chemo escalation to FOLFIRINOX. Those who’re originally ctDNA detected, those are going to get randomized either to oxaliplatin and fluoropyrimidine or FOLFIRINOX because of the concern that they have a much worse prognosis and trying to look if that increasing the intensity of chemotherapy will overall improve outcomes. So again, this study is set up to look at those 2 questions, comparing disease-free survival in ctDNA-negative patients following resection of colon cancer treated with immediate versus delayed chemotherapy based on serial ctDNA or, and then the second question is looking at disease-free survival in the positive cohort, comparing FOLFOX to FOLFIRINOX. The study, as of February of 2025 is one third accrued and I would strongly encourage those who have this open at their sites to continue to enroll patients because it really is an important question in how we can think about this.

The second topic that I’ll focus on is related to immunotherapy in early-stage, both colon and rectal cancer. So these are the more updated results from the group at Memorial, that looked at giving a checkpoint inhibitor to MMR deficient, locally advanced rectal cancers. The drug used in this trial was dostarlimab, patients who had Stage II or Stage III rectal cancer, or found to be deficient mismatch repair proteins by immunohistochemistry, were treated with dostarlimab. They were then evaluated and if they had residual disease the plan was for them to receive chemoradiotherapy and then potential surgery versus nonoperative management, for those who had a complete clinical response to move on to nonoperative management. And there’s serial biospecimen collections along the way and a very close follow-up of these patients. It really has been a stunning result so far. So you can see here, people at different timepoints of where they are, but 100% complete clinical response. Forty-two patients have already, at the time of this analysis, had already completed their dostarlimab.

The median follow-up in this cohort so far has been a year and a half but there are multiple people, quite a few, years out even now and continue to not have a recurrence in this population, which is really sort of unheard of in this population. So, again, these are patients only receiving dostarlimab, not receiving chemotherapy, not receiving surgery, and not receiving radiation. And those patients will continue to be followed. The caveat is it’s a small subset of rectal cancer, so only about 5% to 10% of rectal cancers. Again, nobody has received any other therapy. The sample size is still modest but continues to hold up.

You know, the questions of this is, is this applicable to other checkpoint inhibitors? Most of us think probably yes. How many doses are really needed? In this study they receive up to 6 months of therapy. Will there be later recurrences that we just are getting delay. We just don’t know and longer follow-up is needed.

And the question of how much therapy is needed, I think particularly comes out when you look at these 2 studies from the NICHE group, that looked at immunotherapy in colon cancer, again, mismatch repair deficient colon cancers. And this set up, and this is looking specifically at the NICHE-2, is that patients receive 2 doses of treatment and then underwent surgery and you can see the vast majority of them are having pathological responses, and 68% have actually complete pathological responses at the time of surgery. And you don’t really ever see waterfall plots like this in really no other setting. So, again, a remarkable finding of how affective immunotherapy is for early-stage colon and rectal cancer. They do actually collect ctDNA in these patients and you can see here that, at baseline, the vast majority of them are ctDNA-positive. And after the 2 cycles of therapy, about half of them become negative. But right before surgery, 80% of them are negative and then after surgery 100% are negative. And so far, the 3 year disease-free survival is 100%. And it’s really, again, an unprecedented finding and, in this case, ctDNA does seem to predict what they’re seeing, what they learned about the pathological findings at the time of surgery.

So they only received 2 doses of nivolumab. They actually only get 1 dose of ipilimumab. 2.7% received chemotherapy after surgery, 3 of the 14 who had positive lymph nodes, and 1 of them actually had residual tumor in the specimen. So it’s a very small percentage of patients even getting chemotherapy after. They really didn’t delay surgery for most of these patients. The plan was within 6 weeks from study enrollment. So it really didn’t delay these patients actually getting surgery. What we also learned is radiology didn’t actually predict response very well. So there was only a radiological response seen in 2 of the 75 patients with a path CR. So that’s the 1 caveat of this, is really the challenge both in identifying patients who are the right stages for this, and then being able to predict who potentially maybe surgery shouldn’t be necessary. And I think that’s the challenge. Because, again, it’s not 100% path CR rate. It’s a 68% path CR rate. So that is a challenge for these patients, but they do avoid chemotherapy, the difference in response and duration for localized therapy compared to metastatic disease. And I think that’s really an important other thing. What we learned both from the rectal study and the colon cancer studies, is that patients who have just localized disease seem to have a much better response to immunotherapy versus those who have metastatic disease, where the progression-free survival of metastatic disease combined with chemotherapy is 71%. It’s not anywhere near — and not just a response of at least 30%.

The NICHE group also then proceeded with a second study looking at another combination of chemotherapy, a LAG-3 inhibitor with nivolumab, a very similar result to what we saw with the NICHE-2 data. So again, a very high pathological tumor regression and complete response rate in patients receiving nivolumab and a LAG-3 inhibitor. These were, again, 2 treatments. They were 28 days apart. Very similar results. Unclear if you really need the combination. So this really raised the question, what would be the results of nivolumab alone? It raised the question, is a LAG-3 inhibitor better than a CTLA-4 inhibitor? We don’t really know. It did seem to have a higher rate of some immunological toxicities, particularly thyroid dysfunction, adrenal insufficiency in colitis compared to nivolumab and ipilimumab. But again, there needs to be some caution with cross study comparison.

And this is looking at some ongoing studies that are looking at PD-1 antibodies, either as single agents or in combination with chemotherapy or without in this setting.

There is a study looking at single agent dostarlimab for mismatch repair deficient tumors across multiple cancer types, and you can see here ones that are deficient in mismatch repair proteins, colorectal cancer, gastric cancer, small bowel, pancreas. It’s a whole host of different studies. They have variability in terms of response rates in these populations. And you can see here when we look across all different cancers that are dMMR-positive, you can see here the progression-free survival across those different cancer types and it really confirms what we learn from other PD-1. The other thing about this study is it included POLE E, those who actually have molecular alterations that lead to high TMB, tumor mutational burden, that they’re responding to chemotherapy and there’s a long tail to the curve.

So the last study I’ll present is a study of aspirin in this population. The background of this is that there was an observational study published in the New England Journal several years ago that looked at patients who were receiving aspirin, most of them for reasons of cardiovascular reasons, who had colon cancer, if their tumor had a mutation of PIK3CA, an active mutation, that they seem to have a markedly improved or more significant benefit to taking aspirin than those who did not have a mutation, and the hazard ratio was about 0.5 in that study. And that’s been replicated in a few other studies. In the CLGB 8702 that I showed earlier, with celecoxib, the subpopulation that had PIK3CA mutation also had about a 50% improvement in outcome in terms of disease-free survival for this subpopulation of PIK3CA mutated versus those who were wild type. And those data led to now several studies, and this is the ALASCCA study that was presented at GI ASCO this year, specifically in colorectal cancer patients who had a PIK3CA mutation.

They were a population that received chemotherapy with aspirin or placebo, or if they weren’t going to get chemotherapy they only received aspirin, 3500 patients were screened. They created 2 populations, the hot spot mutations of PIK3CA in exon 9 and 20, and then a whole host of other mutations that also lead to activation of the pathway. In each of those groups they were randomized to either aspirin or placebo. The primary endpoint was disease-free survival. The preplanned analysis was focused particularly on those in the hot spot mutations, but then also looking for secondary analysis in the other mutations. And what you can see here, both in the hot spot and the other mutations, there’s a significant benefit in terms of disease-free survival. And again, that 50%, which has really been consistent across multiple studies, a hazard ratio of 0.49 for those taking aspirin versus placebo. This is looking at cumulative incidence of colorectal cancer recurrences. The dose of aspirin is 160 mg daily. Their conclusions was it can change clinical practice for about a third of the patients with nonmetastatic colorectal cancer. It’s repurposing a safe, inexpensive, globally available drug. It does point out, which is what we don’t normally do, is looking at molecular alterations in early-stage disease, but potential need to do that to be able to identify these patients. So it’s the first randomized study to show benefit of aspirin in a subset of colorectal cancer. And as I said, the hazard ratio is very similar to what we’ve seen in a randomized study related to celecoxib that uses a secondary endpoint as well as observational studies for the PIK3CA-mutated patients.

So that is a little bit of a wrap-up of how we think about ctDNA, immunotherapy, and now aspirin in early-stage colon and rectal cancer. Thank you so much.

Advances in the Management of Metastatic CRC (mCRC) — Scott Kopetz, MD, PhD

DR KOPETZ: Hi, my name is Scott Kopetz. I’m a Professor and GI medical oncologist at MD Anderson Cancer Center. Delighted to be able to participate in the Year in Review and really highlight the advances in the management of metastatic colorectal cancer. So I hope to be able to share with you the real advances that we’ve made. So this does feel like an impactful year and I think we’re leaving the year with better outcomes for our patients than when we came in.

So the first, just to jump into an area that has been really kind of provocative and thought-provoking, is this idea of transplant for liver-limited metastatic colorectal cancer. So there’s been a series of smaller reports over the years that have commented on and shown outcomes of single-arm series of liver transplantation. But this was a large, randomized study called the TransMet study that Rene Adam and the TransMet group led. And so we started off the year at ASCO with this really kind of convincing data.

And so this was a randomized study. There were 157 patients. They had to be reviewed by a panel, this validation committee to make sure that they indeed were felt to not be resectable by traditional single or 2-stage hepatectomies, and not have tumor progression going into this. So really, these are kind of good biology patients, disease control, liver-limited metastatic disease. There were 94 patients randomized, so a heroic effort to get to that, and evenly distributed between the chemotherapy and a liver transplant and chemotherapy in an intent to treat manner. And you can see of those, that we’re looking at small-ish numbers, but in the 30s number of patients included per protocol. And you can see there was some crossover as shown in the slides here.

The primary endpoint was the 5-year overall survival. And as you can see, this is the intent-to-treat data. And what you can appreciate is a striking separation of the curves favoring patients who were randomized to the liver transplant. And this was a hazard ratio of 0.37. You can see the p-value there. And I think strikingly, seeing a 5-year overall survival of 57% versus 13% with the caveat that these were small numbers going in and only a few that had made it to the 60-month time point. So what do we make of that?

Well, this is very impressive results, I would say, in highly selected patients. But I think a study that was well executed. This was done in France where there was a high priority that these patients were placed on the donor list. So because of the research component, they were really put to the top of the list for organs. And at the moment, that’s not really where cancer patients fall in the US. So it is provocative, but maybe not directly applicable. However, one of the changes that we’re seeing is this real increase in living related donors, so the idea that it’s not an unrelated donor as was done in the TransMet, but this idea that family members, related donors may donate part of their liver to the patient, and that may improve feasibility in the future. So there’s studies ongoing, cohorts being done in the US for this now to get better understanding of this. But I think really, an area to watch. But it is still very much research related at the moment.

The next study is the CheckMate 8HW study. This is a game changing study, and I’ll share why I think so. This is a study that randomized patients with MSI-high metastatic disease that had not received prior treatment. And they were randomized in a 2:2:1 fashion where 1 out of every 5 patients received investigator’s choice chemotherapy alone, no immunotherapy in the first line, and then randomized either to a PD-1 or a PD-1/CTLA-4, which are really kind of the key questions that we’ve had in the field.

Now I’ll note as an aside that one of the interesting areas here is that there’s still this bit of issue with MSI-high testing. And I think this is demonstrating in this case that patients who had central MSI testing demonstrated MSI-high or deficient mismatch repair, 16%, 1 in 6 patients did not have that confirmed centrally. So I think just a little bit of a caveat to keep an eye on and think about in those patients that have features of the immunohistochemistry that may not quite fit. But an area for development.

But back to the study. This first report that came out was demonstrating the impact of the doublet, nivo/ipi, versus the chemotherapy. And that was a striking homerun, of course. And this is following the pembro data compared to chemotherapy. And what you can see is a really convincing 72% versus 14% 24-month progression-free survival. You can see a hazard ratio and p-value there. Now the real question though has been, does CTLA-4 add anything to the PD-1?

And that was answered then with the recent report of the nivo/ipi versus nivo arm. And again, PFS is the primary here. And the randomization then is as shown. And the results are as shown here. And I think consistent with what was seen in these cross-trial, cross-arm comparisons, for example, CheckMate 142. What is seen here is a convincing improvement with the addition of those 4 cycles of the CTLA-4 with a hazard ratio of 0.62 and the p-value as you see there. This is amazingly an area where median progression-free survival probably isn’t the best metric, but it’s really looking at that tail with a 36-month progression-free survival rate at 68% versus 51%. So convincing benefit. This was published in Lancet just a few weeks ago.

And my take on this is this is really the new standard of care for first-line patients. Now the CTLA-4, again, just a few doses of it did have slightly higher rates. This is low-dose CTLA-4.

But I think an important point is that these curves, this 68% and 51%, the vast majority of these patients are cured of their disease. So the treatment has stopped, they have not progressed on the disease on scans. And the data really suggest that very few of these patients have any viable disease remain or will progress. So what we’re looking at on these plateaus are cure rates for patients.

And so this is likely improving cure rates in this situation. Now I’ll put some asterisks on that word, cure. It’s a big statement to throw out there. And we’ll look for the longer-term follow-up. But all the data that we’ve seen with prior cohorts really suggest that these plateaus are stable in the majority of patients.

Alright. Shifting gears. A reminder, the BEACON study. This is a previously reported study looking at encorafenib and cetuximab, so cytotoxic chemotherapy as a comparison to encorafenib and cetuximab in these BRAF V600E-mutated patients, a second- and third-line study. The report here was looking at the translational endpoints and really trying to better understand what is the mechanism of resistance that we see in these patients.

And this was reported out in Nature Medicine last year. Here, it’s a busy slide, but pay attention to the red, the red hashes. These are acquired alterations. And what you can see is in the doublet or the triple arm which included MEK, which is not part of the standard of care, it didn’t add anything to the efficacy. What you see is a higher rate of KRAS, NRAS mutations, MET amplifications, MEK mutations or MAP2K1 mutations that are present. And so these really are not seen in the control. So really, the initial concepts here are that these mutations are occurring and may be contributing to resistance.

Now one of the key findings though when you start to look at these in more detail and actually ask how much, and you can estimate the percentage of the tumor that actually contains these mutations. And so some of the provocative data coming out of this was that when you are controlling for the amount of DNA being shed, you can estimate that less than 1 out of 100 cancer cells in the tumor are actually carrying these resistant clones at the time of clinical progression. So there’s some discussion in the literature and in the field now to say, well are we really overestimating the contribution of these secondary mutations to clinical progression? And when we’re dealing with something that’s occurring in less than 1% of the tumor, is that really why the tumor is growing on the scans or are these just happening in parallel? So we don’t have answers, but I think some provocative data coming out of that.

So the conclusions here. These are common after targeted therapy when given alone. We see this after EGFR or KRAS/EGFR inhibitors as well. But again, the suggestion these are very rare, 1 in 100 or less in the tumors and they may not be driving resistance in the majority of patients. So more that we need to learn there.

So another update in the BRAF field was the report of the BREAKWATER study. So this was a large multicenter Phase III study, first-line BRAF V600E mutant metastatic colorectal cancer. This was the, a study that was executed under the FDA’s Project FrontRunner. So this was the first study to report out through this mechanism. So the Project FrontRunner is the FDA’s initiative to try to bring novel therapies into earlier lines. And what the FDA said is that we will provide an accelerated approval for the agent if you can show higher response rates and durability of response. And that we will then convert that to a full approval if you then meet your primary endpoint of the study, in this case, progression-free survival. And so this is designed to really accelerate innovation in the first-line setting and provide therapy to patients sooner. So this was the first study to report out that first result. So this is a randomized Phase III. It has a dual primary endpoint of progression-free survival and overall response rate per that FDA plan.

And indeed, I’d say that the response rate was increased. It went from 40% to over 60% with durability. And as a result, the combination of EC and FOLFOX was FDA approved in December. In this interim analysis that was presented at GI ASCO though, also showed that there was a really strong trend to improved overall survival. And you can see a hazard ratio of 0.47 and a p-value of 4.5 x 10^-5 at this really early look. So we’re looking forward to being able to share more data at an upcoming meeting on the overall survival and the progression-free survival. I will say that a press release saying that the overall survival has been met. So this is a statistically significant difference now. We’ll be able to report more on that later. So the anticipation then is that the encorafenib, cetuximab and FOLFOX, which has accelerated approval in the US now, will turn into a full approval to follow.

So what do we conclude here? I would say, with the caveat that I’m an author on that, so I may be a bit biased, but that encorafenib, cetuximab and FOLFOX is the new standard of care for first-line BRAF V600E mutated colorectal cancer. I will say that the degree of this benefit was greater than any of us had anticipated and really was impressive magnitude. This was following the really strong safety lead-in data that had been presented in 60 patients earlier. And so I think really is kind of compelling and follows some biology and really nuanced understanding now of how tumors are adapting to MAP kinase therapy and adapting to cytotoxic therapy. And the idea that we may be, by combining, not just 1 + 1 = 2, but really providing synergy that follows some of this biology of tumor plasticity. So more to come on this, but I think really compelling results. FOLFIRI hasn’t been forgotten, for those FOLFIRI aficionados in first line. There is a randomized EC + FOLFIRI arm that will be reported out later that was enrolled after the BREAKWATER study completed.

Alright. What about HER2? So MOUNTAINEER, to remind everyone, was the first study looking at tucatinib and trastuzumab dual HER2 inhibition. And this was second line and beyond. RAS wild-type, that’s an important point, RAS wild-type colorectal cancer. And the tucatinib and trastuzumab are blocking signaling through HER2. And we’ll come back to that and contrast that with the ADCs to HER2. But this is really blocking signaling; therefore, requires a RAS wild-type tumor. Just like an EGFR inhibitor requires a RAS wild type for signaling.

So the final analysis was reported there. Response rates of 39%. And I think, as you see the spaghetti plot on the right, really some impressive durability. And you can see a number of those out well past a year. Some of them even into 2 years. And so I think just acknowledging that some of these responders can be really durable. And I think we’ll continue to watch the tails of some of those amazing responses.

So this is a standard of care for second line and beyond as it has been already. And this just gives us reassurance. There’s a subset that have a really long benefit. But again, RAS and BRAF wild-type patients and HER2 amplified.

So what about the DESTINY-CRC01? So this is now looking at trastuzumab deruxtecan. So this uses a topoisomerase payload with the trastuzumab antibody. But predominantly, we think this is working through the cytotoxic payload. Trastuzumab alone does not have much activity at all in colorectal. It really needs the doublet or combinations. But we think it's through the payload that is predominantly doing the heavy lifting. Now what does that mean? Well, we’re delivering a payload and so patients can enroll and may be deriving benefit even if they have some mutations. We’ll talk a little bit about that and the emerging data there.

So DESTINY-CRC01 was looking at the first cohort here and some updated response and biomarker data. What you can see is the responses are predominantly in the IHC 3+. The 2+/ISH-positives, maybe some activity but not quite as high there was seen. Now you’ll notice that there were some RAS mutations, PIK3CA mutations, HER2 mutations kind of scattered throughout. So really, no strong correlation with the mutation status.

Now one of the questions was, how are the tumors progressing? If this was a targeting therapy, maybe we would have expected these secondary mutations to appear. But we didn’t see any. Really no meaningful number of acquired MAP kinase alterations. And there’s also discussion about, well perhaps the tumor could lose its HER2. We see this in gastric cancer, for example, and others. But really, that was not seen. It was not a very high rate of that. It was a modest number of patients that that occurred.

Now the CRC02 now is the randomized study. This was a dose optimization looking at 2 different doses. Here, again, RAS mutated or wild type allowed in. IHC 3+ or 2+ with ISH-positive.

And the data was presented by Kanwal Raghav and published in Lancet Oncology, 37% response rate with the 5.4 mg/kg, slightly lower response rates and less favorable toxicity profile in the 6.4, so the 5.4 is really the recommended dose. Now you can see a scattering of the mutations there. And, in general, you can see on the bottom that the RAS status local test here, really no difference statistically in the response rate by the mutation. But interestingly, very low rates of the 2+/ISH-positive. So you did not see a ton of activity there. Really, the activity was manifested predominantly in the 3+ population. So reiterating that this is not signaling, right? Signaling and ADCs. So it’s really an expression that’s driving this. The more expression, the more payload that you get in.

So what can we conclude about this? So again, trastuzumab deruxtecan remains the standard of care for second-line and beyond. HER2 amp without regard for RAS or BRAF. Predominantly, resistance is not through loss of HER2 amp. There were some, but not a big mechanism and acquired RAS mutations weren’t seen. Because questions have come up, how do I sequence my HER2 therapy? And if I have mechanisms of resistance through one that may impact sensitivity to the other, maybe it matters in what order I do them. My conclusion of this is that we should do what’s best for the patient, but we don’t have to be wed to a certain sequence.

What do I do? I tend to do more targeted first followed by ADC second. And a lot of times, that’s to give a little bit of space between my cytotoxic payloads. I’m always a little concerned about giving a topoisomerase payload right after a patient is progressing on irinotecan, for example. And so having a little bit of time in between. That’s a bit of a nuance. I think this data gives you the flexibility to really pick the sequence that you would take a patient through.

Alright. What about KRAS? KRAS mutations, common in colorectal, of course. We’d love to have inhibitors of all KRAS. Maybe we’ll get there. We’re not quite there yet. But where we have made progress is with the G12C. So remember that G12C is just a few percent of colorectal cancer. Unlike lung cancer where it’s a more common mutation, it’s less common in colorectal. But absolutely seen. So the current data is with sotorasib and panitumumab or adagrasib and cetuximab. And so we’ll just review some of the update data on those 2 approved agents. So this is the CodeBreaK 300 study. This was looking at sotorasib at 2 different dose levels versus the investigator choice. Prior data had suggested the response rate and the activity was higher with the 960 mg daily dose compared to the 240. So that is the approved and recommended dose. And this was asking the question in a Phase III study that was looking at progression-free survival, but looking at a key secondary of overall survival. So this was in patients who had had prior 5-FU/oxaliplatin/irinotecan. So you think about traditionally, a third-line and beyond setting. And they were randomized to trifluridine/tipiracil or regorafenib as the control arm.

So the protocol specified final overall survival in intent to treat, as I’ve shown here. A little disappointed there wasn’t more of a survival benefit seen. You can see the hazard ratios there were trending in favor of the KRAS inhibitor. But remember, access. There was access to KRAS inhibitors on crossover. And about one third of patients in the control arm went on to get other KRAS inhibitors. And this, again, small numbers, as you can see, wasn’t really powered for overall survival. But this was the update that we received on this. But again, active regimen, good response and progression-free survival.

The final is the adagrasib/cetuximab. This is an update of the KRYSTAL-1 study that led to FDA approval of this regimen. And you can see here in the study that continued to demonstrate a nice response rate, 34% by blinded central review, 43% by investigators. Overall survival is 16 months. This was not randomized, so we don’t have a comparison. And really reiterating the benefit of this in this second or third-plus-line setting. There is a KRYSTAL-10 study that has finished enrollment that is randomizing in the second-line to adagrasib/cetux versus a FOLFIRI/bev or standard of care regimen. So we look forward to the randomized data there.

But we can say at the moment that both adagrasib and cetuximab, and sotorasib and panitumumab are standard of care options for KRAS G12C tumors that have previously been treated. The efficacy may differ between these. And we can look at the cross-trial comparison of response rates or progression-free survival. The divarasib data also has shown some impressive efficacy for G12C in colorectal with an EGFR inhibitor. So I’d say that the final chapter has not been written on this. But that for the moment, we have FDA approved therapies that really give us some flexibility and certainly something to think about. I think most excitingly is the roadmap that this provides for the G12D inhibitors or perhaps the pan-RAS or pan-KRAS inhibitors that will be following.

And stay tuned next year for the 2025 Year in Review. And hopefully, we’ll have some more data to share in that space as well. Thank you.

DR LOVE: Alright. Awesome job. Just a couple of quick follow-up questions. One, getting back to the last thing you were talking about, these KRAS G12C situations, inhibitor plus EGFR. I’m just kind of curious how much you think it actually offers a patient. We’ve spent so much time over the years talking about regorafenib and TAS-102, and I always wondered, do you really see a benefit? And I wonder the same thing here. I mean, have you seen patients have, you know, what seems to be not just stable disease, but good responses to these therapies? And how much quality-of-life issues? If you randomize versus observation, what do you think you’d see?

DR KOPETZ: Yeah. I don’t want that overall survival data to really kind of sour — that was data that was updated this year. But really, we do. And I think that’s a bit of the outliers. So these are tumors that are responding really well and patients are feeling much better. The G12C inhibitors are really well tolerated. We do have the cetuximab rash, but that’s no worse than you would if you’re doing a single-agent cetuximab, so certainly manageable in the majority of patients there. And we do have durable patients. We have patients that are on 2-plus years with these inhibitors. So there are a subset that are getting really good activity. But the median duration of response is in that kind of 6- to 8-month range. And so acknowledging that, again, these are late-line patients, but that most patients are finding — tumors are finding ways to progress despite the therapies.

And so that really is the challenge of how do we do a better job of addressing that resistance. There are strategies to think about. Combining with immunotherapy. So there’s a Stand Up To Cancer study ongoing in this space to look at adagrasib, cetux and a PD-1 inhibitor that looks, you know, we’ll be waiting to see some data on, just given the role that RAS plays in immunosuppression. And then there are other strategies to think about understanding the mechanisms of resistance to these agents and thinking about how do we intervene.

DR LOVE: That’s really helpful. One final question. You referred to one of the trials of MSI high to the fact that there were this 16% of people who were thought to be MSI high based on the community testing that were found to be not MSI high. What’s going on there?

DR KOPETZ: Yeah. So what we’re seeing is that a number of these are some immunohistochemistry tests gone bad. And so there are some subtleties. And sometimes, if you don’t have good controls, you’ll misread some loss of protein expression on some of these mismatch repair proteins. Remembering if you just have loss of 1 out of 4, then of course, you call it an MSI-high tumor. So things to watch out for, isolated loss, for example, PMS2. Most of those are kind of technical flukes and not true MSI highs. So there’s been some discussion about whether MSI high should really be confirmed through 2 methodologies. Through the IHC as a quick screen, but then if there’s anything that feels a little off, that’s where looking at the NGS to see what the mutation burden looks like. Many NGSs are now calling MSI through colors. And so I wouldn’t be surprised if we don’t see guidelines evolve over the next year or two to really encourage people to do kind of confirmation testing in some situations.

DR LOVE: I was just kind of flashing on some of the cases I’ve heard over the last few years where somebody had MSI-high disease and just did not flat-out respond to an IO. And the doc was like, why? Why don’t they respond? And I wonder if some of them are really not MSI high.

DR KOPETZ: Our estimates are it could be as many as half of the nonresponders are just misdiagnosed.

DR LOVE: Wow. And that kind of ties into, in general, the role of IHC nowadays. That’s always been kind of an issue out there. Any attempts using AI? I know AI has been used for pathology reading and radiology reading. Do you think that could be helpful?

DR KOPETZ: Yeah, it certainly can. MSI is one of the easier histologic features, even on an HNE, to pick out. And so the first AI examples a lot of times were ability of AI to call MSI high off of an HNE. Mostly, because you get this really robust lymphocytic infiltrate. And so that can be something that’s detected by AI. So I do wonder if better awareness, first of all, of kind of the quality control and the metrics and when something would be a little questionable on the IHC would be one component. But I do wonder if additional diagnostics, whether confirmatory NGS or, if we get there in the future, the AI could really make a difference.

DR LOVE: I’m glad you reminded me of that. I sort of forgot about that. I remember hearing stories about ‘Oh, the pathologist called me and said, is this an MSI patient because of the histology?’

DR KOPETZ: Yeah.

DR LOVE: Do you see that in general? When you have a patient who is MSI high and you check their histology, is it usually that’s the way it’s read?

DR KOPETZ: Yeah. So you can indeed, you know, the sharp-eyed pathologist will sometimes note that. It is one that’s not perfect.

DR LOVE: Right.

DR KOPETZ: And so the absence of a call of a lymphocytic infiltrate shouldn’t dissuade you from doing the IHC testing anyway. But yes, it certainly can be reassuring.