Introduction

DR LIEU: All right. Good evening, everybody, and welcome. Thank you for taking time out of GI ASCO to be here today, and thank you to all those that are online joining as well. Welcome to “Expert Second Opinion: Optimizing the Use of Immunotherapy, MRD Assessment and Other Novel Approaches for Patients with Localized Colorectal Cancer.” My name is Chris Lieu, and I'll be moderating the meeting today. This is our faculty for today. We're very thankful for their work and effort to be here. This is Dr Stacey Cohen from the University of Washington, and we're very thankful to have Dr Jenny Seligmann from the University of Leeds. So thank you for coming all the way from the UK.

These are our relevant faculty conflicts of interest for Dr Cohen and Dr Seligmann and myself. And here's the commercial support for today. Research To Practice CME planning committee staff and reviewers have no conflicts of interest to disclose. And as a reminder, this program does contain discussion of non-FDA-approved uses of agents and regimens. So as always, please refer to official prescribing information for each project, product for their approved indications.

All right. So what are we going to be talking about today? So this is where we're heading. We're going to do this as if we're doing treatment in terms of lines of treatment. So we're going to start with neoadjuvant treatment for localized colorectal cancer. And Dr Seligmann will bring us through her slides. Then we're going to talk about the adjuvant therapy and some novel emerging approaches for adjuvant therapy for localized colorectal cancer. And then Dr Cohen will lead us through the role of ctDNA and ctDNA testing in localized colorectal cancer. And with each of these modules, we'll actually be discussing the slides and the data behind neoadjuvant, adjuvant and ctDNA testing.

But we're also going to go through a lot of cases. And so what you're going to see is that a survey of 50 community-based general medical oncologists was performed from December 22nd through January 7th. And these oncologists definitely did not go easy on asking the hardest questions possible. So you're going to see some real-life examples, because we're going to show you data and all clinical trial data fits into a nice little box. But those of you in the room and virtually certainly understand that our patients rarely fit into those boxes. So we're going to go through a lot of cases. We're going to talk about how to apply some of this data in the neoadjuvant space, the adjuvant space and then what to do with ctDNA testing, because we know that there's a lot of testing going on. What does the data tell us about how we're supposed to be using this in clinical practice?

Neoadjuvant Treatment for Localized Colorectal Cancer (CRC) — Dr Seligmann

DR LIEU: So with that, I'm going to turn it over to Dr Jenny Seligmann from the University of Leeds to discuss neoadjuvant treatment for localized colorectal cancer. Thank you, Dr Seligmann.

DR SELIGMANN: Thank you. Thank you for the introduction. And thank you for the opportunity to speak to you all tonight. So I am going to talk about locally advanced colorectal cancer. Now, this isn't a small brief. This is a very big brief, actually. So we've got a tale of 2 tumor sites, first of all. So we've got differences in neoadjuvant treatment between locally advanced rectal and colon cancer, but also how we select patients for neoadjuvant treatment, what the relevant endpoints are for trials in these settings and also response assessment. So I'm going to go through some of those and the differences between the 2 tumor sites.

Additionally, we have another main thing to consider, is that we've got a tale of 2 biomarker groups. And, in fact, we really need to consider these groups very differently. First of all, we have the MSI-high deficient mismatch repair group, which is a small but important proportion of localized rectal cancer and a slightly larger proportion of localized colon cancer, which I think is really important. And again, we've seen, unless you've been living under a rock, you will have seen the New England Journal papers. So a huge impact, great successes, great efficacy. But remember that the larger proportion of patients are microsatellite stable. And it feels slightly like being a Scottish sport fan here, where we're continually watching the MSI-high groups and going, oh, come on. But I think there's reasons to be optimistic, and I'll discuss those today. Like Scotland qualifying for the World Cup, for example.

So let's talk about neoadjuvant approaches and why is this a potentially promising avenue. And remember, we're trying to cure people in this setting. So this is really important that we're pushing the boundaries. So what we know, there's positive experience in other cancers. It allows early treatment of micrometastases. It allows better surgical outcomes by downstaging, particularly in advanced bulky tumors. There's now the potential for organ preservation, as we'll discuss. And we also see, particularly for immunotherapy, but for chemotherapy first, there's a stronger and deeper immune response with immunotherapy when the primary tumor and its intact tumor microenvironment is in situ. However, it's not quite as straightforward, and there's maybe some potential downsides.

So there's a concern when introducing neoadjuvant treatments that patients may not proceed to surgical resection, either due to progression in the neoadjuvant window or, indeed, due to toxicity. Additionally, is there going to be concerns around perioperative complications, remembering that surgery is, of course, the really key part of this treatment pathway? And then the third thing that there's concerns around is, can we select appropriate patients using mainly radiological staging assessment? And I'll touch upon that as well.

So let's start with locally advanced rectal cancer. Now, I'm not going to go through all of the data from the trials, because really now, total neoadjuvant treatment with chemotherapy and radiotherapy; i.e. TNT, is international standard of care in locally advanced tumors to reduce local and distant recurrences. And we've got 4 trials that I've shown here, RAPIDO, PRODIGE, STELLAR and OPRA, that have all told us slightly different information with relatively consistent data on the surface.

What we know about rectal cancer, however, is that this is a lot of treatment. We're subjecting, well not subjecting our patients, but we're asking our patients to have usually nearly 6 months of treatment ahead of even having an operation. And multimodal treatment is lifechanging with risk of bladder, sexual, bowel dysfunction. However, this need has led to the OPRA trial showing that in complete clinical response, organ preservation is viable, with no detriment in disease-free survival compared with TME.

But interestingly, how then do we schedule it? So some clues from OPRA seeing the up-front chemotherapy, then consolidation chemotherapy, led to higher 3-year TME-free rates. We also saw this between the PRODIGE 23 and the RAPIDO trial, and again, this adds the question, do you need chemoradiotherapy or short-course radiotherapy? And of course, we have all options available to us in the guidelines.

So I think a really key question at the moment is, how do you select the TNT strategy? So number 1, something that may guide you is the anatomy of the tumor, the proximity to the CRM and the peritoneal reflection. The treatment intent. Are we trying to aim for organ preservation in our patient? Is that the most important thing? And also, we have varying degrees on what locally advanced means. And we certainly, in my view, have the good, the bad and the ugly. And we do need to think about things like, well, what are we most concerned about? Is it local recurrence in a low tumor or are we more concerned about distant recurrence, particularly in particularly node-positive patients with EMVI? Again, I'm not going to go through this in detail, but this was a really helpful article from the team at MD Anderson, maybe helping us think about the factors that may guide which TNT strategy that we would use.

MSI-high locally advanced rectal cancer. Again, this is very well-known data and really, really exciting data. So this is the data that has come from MSK where 6 months of dostarlimab was given for locally advanced rectal cancers. Many of them bulky T4 patients. And what they reported is that all patients managed to achieve nonoperative management just with dostarlimab. No radiotherapy, no surgery. This was really astounding data. Also, there's no recurrences at this point, and they saw the fall in ctDNA. So you can see the forest plot for response, but for me, the actual endoscopy pictures are really something, showing really nasty tumors and then normal rectal mucosa.

And I think very much the human side of this. So again, we were talking about this lifechanging multimodal treatment, and here's Andrea Cercek, the chief investigator, alongside one of the patients who was able to proceed to have a pregnancy. So this can be big, lifechanging data for many patients.

But we don't usually change our data based upon a study of 49 patients from 1 single center with world-leading expertise. You have to ask about the generalizability of the patients and particularly the ability to assess the complete response. Delivering the dostarlimab is quite easy, but interpreting the endoscopy and the MRI findings is more difficult. However, this is a small and rare population, and we have to think about the ethics of randomizing to standard of care based upon the current data. So this is reflected in the guidelines, and we have the AZUR-1 trial, which is going to give more generalizable data. Again, it's nonrandomized, but it's a registrational study.

So neoadjuvant IO in MSS, locally advanced rectal cancer. Again, I'm not going to touch too much on this, but I think we're going to hear far more about this in the coming years. We often think that IO doesn't work very well in MSS tumors, but actually in the neoadjuvant setting, as I described, the biology is different. So if we think about a historical control CCR rate of 25%, we have 2 studies here, the TORCH and the PRIME-RT study, reporting CCR rates of 50 and 62%. So again, I think this may be a promising avenue for the future.

So we're going to spend the rest of the talk just touching on neoadjuvant treatment for locally advanced colon cancer, which is not as well developed as compared to rectal cancer. We know that there's needs in this setting. So there's been very few improvements in the care of colon cancer. We'll discuss some of them today. But ultimately, we have a group of patients who are failing and recurring despite current standard of care.

So one approach then is to think about giving the chemotherapy ahead of surgery. We have 4 main trials which have reported these. And this is the schema of the FOxTROT trial, just one of them, they're all very similar. And the trials are very simple. They give a proportion of the planned chemotherapy ahead of surgery. So in FOxTROT, it was 6 weeks, but with a total duration of 24 weeks compared with up-front surgery. And again, there's been a recent meta-analysis of neoadjuvant chemotherapy, including over 3,000 patients. And what they found was up-front surgery, compared with up-front surgery, neoadjuvant chemotherapy led to significant improvements in 5-year disease-free survival and in overall survival, leading to meaningful reductions in recurrence and in death.

Additionally, there was very reassuring data in all of the trials around surgical safety and in positive surgical margins. So again, consistent data over 3,000 patients. And in particular, when we limit this to the MSS patients, i.e. the group of patients that we're most interested in delivering chemotherapy to, we're seeing a 5% improvement consistently between the 2 trials. And remember, this is the same difference that we get from adding oxaliplatin to 5-FU from the MOSAIC trial. So in my view, very clinically relevant, particularly in the most locally advanced tumors. So neoadjuvant immunotherapy in MSI-high locally advanced colon cancer, again, I've highlighted the major trials. I'm not going to go through them all. They all give slightly different things in terms of the duration.

Do they give single or doublet immunotherapy? But there's some patterns that come out here. So safety and deliverability is demonstrated. We see consistent, impressive efficacy. I mean, these PCR rates are unlike anything we've previously seen in colon cancer. We see some hints of maybe how we should be implementing into standard of care. So when giving single-agent, it appears you need to give longer. Remember in rectal cancer, they were giving 6 months of dostarlimab, but in the NICHE-2 trial, they gave combination nivo/ipi, and 6 weeks duration was sufficient to lead to 68% pathological complete response rates. But the heterogeneity in the study design limits the interpretation.

What a lot of people then say to me is, well, these are good prognosis tumors. Why do we need to give them immunotherapy? I would disagree with that. So we've recently done an analysis looking at comparing with NICHE-2 with the FOxTROT trial. So what you could see here is that there are no recurrences. At 3 years, not 1 patient with a very locally advanced bulky tumor has recurred, whereas the same patients who were either given up-front surgery or neoadjuvant chemotherapy, 20% had recurred. And if we limit that to T4 patients, then actually this is a very big difference. So 30% recurred when given either up-front surgery or neoadjuvant chemotherapy.

And also we haven't cured metastatic disease. Yes, we've made big dents in our outcomes, but there's still patients who are doing badly with metastatic disease. I think something we can talk about in greater detail is, well, with the ATOMIC data coming out with the adjuvant setting, has that now killed neoadjuvant IO in MSI-high? And maybe we can discuss that more around the cases.

Similarly to rectal cancer, what we're seeing is neoadjuvant immunotherapy appears to be active again in colon cancer. Again, this is from the NICHE-1 data. We're seeing PCR rates of 30%, and that's compared to 12% in FOxTROT and OPTICAL. And again, the NEST-1 trial with bot/bal showing really impressive responses. So I do think we're going to see more of this in the next few years.

A few things to be wary of. We are seeing strictures when you have a primary tumor. So when tumors respond, they appear to respond in the way, and that can lead to obstruction. And we're seeing this as we're giving more. And we also have to think about the impact of low-grade endocrinopathies, however, generally well-tolerated.

Patient selection and changing patient pathways is something that we need to work on. However, in both rectal and colon cancer, the trials were based and were positive trials when selecting patients using the radiology. But is there more work to be done? Absolutely. I suspect ctDNA will play a role in this.

So finally, we need to be able to deliver neoadjuvant treatment that often requires changes in your patient pathway. You need up-front MSI testing, reflex testing and you need a radiologist to call TNM stage. We need team working with informed and supportive surgical colleagues. We need a pathologist to tell us if our neoadjuvant treatment has worked. We need to be able to review patients and treat them rapidly. We need to coordinate all of this and make sure everybody knows where the patient is, but really importantly, patient choice. And again, I'd be really happy to discuss this outwith.

So in conclusion, we need to know MSI or MMR status. I think neoadjuvant IO has been transformative in rectal cancer, and I think it has the potential to be in colon cancer. But there are great opportunities, particularly within the MSS population. We need to work on patient selection. We need to keep treating people. I'll stop there. Thank you.

DR LIEU: Thank you for that outstanding talk. And it's amazing how the field of neoadjuvant chemotherapy and immunotherapy has really changed. Obviously, a lot of our questions are going to be about MSI-high and deficient mismatch repair rectal cancer and colon cancer. And I think you did a great job really looking at how diverse the data is. And a lot of these are single-arm Phase II trials with different durations and doublet therapies and single immunotherapies. And then also, really a wonderful review on proficient mismatch repair and obviously the need. But still a lot of neoadjuvant data available for review. So yeah, thank you so much for that, Dr Seligmann.

So as I said before, there's a survey of 50 general medical oncologists. And these are just real-life cases, and this is what we're all seeing in the community and at academic centers too. And so really highlights the fact that this data is present, but how do we really put this into practice? And so we'll start with this, which really comes down to asking the question, has the paradigm totally shifted away from chemotherapy for these patients? So for this 54-year-old gentleman with rectal cancer, MSI-high, deficient mismatch repair, has some comorbidities here, oncology team is considering neoadjuvant PD-1 inhibitor instead of chemoradiation.

For these patients with locally advanced MSI-high colorectal cancer, for you, Dr Cohen, at the University of Washington, have you completely done away with chemotherapy and chemoradiation for these patients and moved strictly to immune checkpoint inhibition? Tell us how you might approach this patient with rectal cancer.

DR COHEN: Yeah, thank you. I would say that we really have made some major changes to the paradigm and really, I think, Jenny, as you pointed out, I mean, the outcomes are just tremendously better with immunotherapy when it works. And so we have really shifted that all of these MSI-high rectal cancer patients are getting neoadjuvant immunotherapy. I think we also have to keep in mind that there is retrospective data that the MSI-high patients might progress through standard TNT. So I think it's not equivalent. I think it could be potentially harmful to go on, especially the neoadjuvant chemotherapy portion. So if someone is considered to be a candidate for immunotherapy, doesn't have a comorbidity that would preclude it, I would absolutely give them immunotherapy.

DR LIEU: Absolutely. And Dr Seligmann, in the UK, has there been a paradigm shift and a movement away, obviously from chemotherapy/ chemoradiation therapy for these patients?

DR SELIGMANN: We want to. Availability is a bit more difficult. So we will probably need the results of AZURE-1, but we have been able to use compassionate use. We've been able to enroll patients into trials. I mean, I just want to, I've given a very positive view of the data, and the data is very positive, but having recruited quite a few patients and treated quite a few patients, there are patients that won't achieve a complete clinical response. And we have seen this, and the difficulty is not for oncologists in delivering the treatment, mainly. It's usually very well tolerated. The difficulty is interpreting the endoscopic and particularly the radiological findings.

So things like, we're being asked to think about new things. So what does that mucin pool mean? It's not an entirely normal MRI. And actually, there's very little in the literature telling you that that abnormality is okay. So surgeons are understandably nervous about leaving a tumor in situ, where you can see probably a post-immunotherapy change in an already mucinous tumor. But there are lessons that we as a community are going to have to learn, particularly after AZURE-1 is published. The stenosing complication is a real thing. I've had a patient who stenosed in front of their primary tumor. So you couldn't actually see, you couldn't do surveillance of it. So it necessitated them having an operation. So the data looks — 49 out of 49 patients, I suspect it's not going to be quite so straightforward.

DR LIEU: Yeah, I think we always want that 100% CCR rate, and that's what made the New York Times. But we also know that that's just not going to be realistic. And one of the things, we certainly at the University of Colorado, we have definitely moved to a immunotherapy neoadjuvant approach, which certainly makes sense in rectal cancer, with the chance to have a CCR. But again, sometimes you're just not going to get a full CCR. And so at that point, you may have to move on to other standard of care therapies. As a practitioner though, you want to hold on to that hope that they're going to get that CCR after a period of time, but we know it's not going to always be there.

So in regards to some of these other questions that are coming up regarding duration of therapy, the role of combination IO/IO in rectal cancer. So for this patient, a 62-year-old gentleman with Stage III rectal cancer, no evidence of metastasis, again, MSI-high, deficient mismatch repair, just as we had talked about neoadjuvant immune checkpoint inhibition. What is the correct duration of therapy, right? And this is a question that comes up all the time and I'll go back to you, Dr Cohen, and just say, is there a strict amount of IO that you're going to do before making a call and what is the right number there?

DR COHEN: I think we can say from the reported data that at 3 months some patients have a complete clinical response, but far more likely at 6 months. I'll say in my own practice, if I'm treating someone off trial, I don't feel obligated to stop at 6 months if I feel like they've had a very good response and maybe as you're describing, you can't quite tell is it or isn't it having a complete clinical response. In those patients, I have sometimes kept going for a little bit longer while we're still trying to decide if we feel like overall we've had a good outcome. Because at least here, I would say it's relatively easy to get immunotherapy approved.

And I've typically done single-agent immunotherapy just with the idea that perhaps the toxicity is a little bit less and we haven't necessarily had any head-to-head data, but I think that'll be very interesting is at some point would we move to a shorter duration of a doublet rather than the single agent for 6 months.

DR LIEU: Yeah, so to that point, Dr Seligmann, you had brought up this idea of being able to access some of these immune checkpoint inhibitors. When you're able to do that within the UK, have you chosen to go more single-agent checkpoint inhibitor or are you trying more doublet?

DR SELIGMANN: I mean, I think this is where the data is. It's very difficult when there is, there's far more data, ironically, in colon, but there are more patients with colon. Remember, we don't see these people very often and so it's very difficult to generate this data. The only thing I would say is we give 6 months and you often then see patients moving from near complete response to complete response 2 months later. So I think just hold your nerve. If you can deliver more IO, I think that's very reasonable. If you can't, then hold your nerve and do a repeat endoscopy, repeat MRI 2 months later and many of those patients will have developed into a complete clinical response at that point.

DR LIEU: Absolutely. And Dr Cohen, not to jump onto your talk, which we're going to hear about, and this is a question that's going to come up and you can see how many providers are asking this question about ctDNA. And this is something that's going to come up in case after case that we discuss today in today's program. So do you utilize ctDNA as part of the surveillance while you have somebody on IO therapy for in the neoadjuvant setting?

DR COHEN: I think you can. I don't think you're mandated to. Certainly the data does not say that we are wrong not to do it, but I think the data also suggests that ctDNA is exquisitely sensitive, especially for immunotherapy response. We know that there's going to be fewer patients who are positive who have localized disease, but I do think that it can be a very powerful indicator for response. And outside of MSI-high cancers, we see that the rectal cancer patients that are still ctDNA-positive tend to have more localized disease as opposed to necessarily having concurrent metastatic disease at that time. So I think it can be a very helpful indicator. Perhaps you have that near CR and you're trying to decide is it or isn't it. I think that having a ctDNA and following that quantitative level may be very helpful in that situation.

DR LIEU: Absolutely, yeah. And again, I think we'll learn more as we get more and more data, but I know it's just a field that's evolving very, very rapidly, but in some ways not rapid enough. So Dr Seligmann, we had talked about complete clinical response and certainly when we were able to get that, that's a phenomenal thing, especially when patients are going to be given an APR and they can avoid that. They have a CCR after neoadjuvant immunotherapy. Tell us about your protocol locally in terms of a watch and wait and how that looks if patients are able to avoid surgery.

DR SELIGMANN: I mean, what we know is that patients are most likely to recur within the first 2 years, but of course that's extrapolated from MSS. So actually we don't really know what the, because none of the patients have recurred. So we don't know when patients are most likely to recur, but we are basing the protocols for surveillance on the MSS patients. So they're quite intense in the first 2 years and then it drops off. So at the moment we've got 3 monthly scopes and I think the imaging is every 6 months from the top of my head. And is that overkill? Maybe, but I think whilst we're developing the evidence base, I think we need to be cautious because really we're keeping potentially a tumor in situ which could be easily treatable by an endoscopic resection. So I think we do need to be conservative.

The hope would be that maybe ctDNA will allow us to back off a bit, but I don't think we're there yet. So you do need, the other thing you need is a motivated patient. This is not the right thing for all patients. Some patients will say, actually, you know what, I'd rather have an operation and get it out rather than coming back every 3 to 4 months concerned about and living with the anxiety of not having had an operation.

DR LIEU: Yeah, it's interesting because the responses to immunotherapy can leave behind a certain amount of scar tissue. And you had brought this up in terms of the obstruction that you may see sometimes with IO therapy. I had a young woman that had MSI-high rectal cancer, received single-agent immunotherapy and was left with this mass, residual mass that MRI called a locally advanced rectal cancer. Scope biopsy, that even though the biopsy was negative, was like, this is definitely cancer. So she did end up going to surgery and of course there is no cancer there. We've been fooled multiple times in this surveillance period and even with MRI and flexible sigmoidoscopy can still be fooled.

DR SELIGMANN: But we will learn from these patients. That's the point. I think we've had a patient as well that we were nervous to call a complete clinical response because of some mucin pools, which apparently can take 2 years to resolve. So we will learn from these patients. We need to come together as a community, and we need to be very open about our outcomes, and we need to publish and talk about this.

DR LIEU: Absolutely. In terms of patients that, for this question, this is an interesting situation. A 44-year-old gentleman with Stage II MSI-high colon cancer wants to avoid surgery, right? Similar to what we've seen with rectal cancer. Open to close monitoring with surveillance scopes, ctDNA and scans. And you have to keep in mind, what this person is asking for is potentially q3 month colonoscopies, right? Which is something that I'm going to talk about a little bit in my talk. But Dr Cohen, any data to suggest that patients with MSI-high colon cancer can avoid chemotherapy in this setting?

DR COHEN: I think it's really tricky. We know that, I mean, doing a colonoscopy at that frequency I think is not practical or viable or at some point safe to undergo that level of anesthesia and prepping and everything. But what the data says is, we can look to Memorial Sloan Kettering again and in their all-comer trial, they saw instead of the 100% responses that we were seeing for rectal cancer, they saw about 83% for colon cancer. And of the ones that failed, at least on what looked to have failed, I think half of them actually had viable tumor. So I think we know that there's no MRI in this space unless you have a very distal sigmoid cancer. So there's not really as good of imaging modality and you have endoscopy that's not as good.

And so given the fact that not every complete clinical response is truly a pathologic complete response, I would say that I am at this point uncomfortable with nonoperative management, unless of course there's going to be that patient that just is too sick for an operation and we will end up learning through those patients. But in general, I don't recommend it at this time. And even ctDNA alone is not enough because I think that we don't have a way of really proving the cancer is gone and monitoring it at any reliable frequency.

DR LIEU: Yeah, Dr Seligmann, your thoughts?

DR SELIGMANN: I think this is an area that will develop. Patients will ask, patients are asking. Patients don't want to have an operation when they have no viable cancer cells. But again, it's the process of response assessment and you've touched on this in terms of it is more difficult with a colonoscopy. CT colon is much better actually than CT itself. But similarly to the, we need to have to look at this in a clinical study that looks at it in a practical way and actually gives real-world data to then actually say, well, this was our intent to go down this road and this is what happened. So there are trials that are actually ongoing in this space. And patients are asking for it. So I think it's our duty to generate the data.

DR LIEU: Absolutely. And we'll wrap up with this one last one just because I think it's such an interesting case. A 56-year-old male with colorectal cancer, Lynch syndrome and 2 isolated liver mets. Any data for immune checkpoint inhibitor or is it better to try and do a chemotherapy/ immunotherapy combination? Now there's very, very little data certainly understandably to answer this question. But Dr Cohen, very briefly, any thoughts about this?

DR COHEN: I think there's many right ways to treat this and it's going to be different for different patients. I think when you have resectable disease, you could rationalize that you could just go ahead and do the operation and be done with it. I think that you could give immunotherapy. I think we'll find out on Saturday if chemoimmunotherapy is the right answer or is an answer here. But when you have resectable disease, I think it becomes a lot more complicated. But you also have to think about this patient having Lynch syndrome is at risk for other primary cancers. So I don't know. I think I would end up talking it through with the patient. I don't think I have a blanket answer in this scenario.

Emerging Novel Approaches to Adjuvant Treatment for Localized CRC — Dr Lieu

DR LIEU: Absolutely. Okay. Very cool. Well thank you guys so much for that great discussion on neoadjuvant therapy. So I get to move on to the adjuvant part of the talk. And so we're going to talk about emerging novel approaches to adjuvant therapy for colorectal cancer. And Dr Seligmann got to talk a lot about some very, very cool data. She gets to talk about 100% complete clinical response rates in New York Times and dostarlimab in that setting. And my talk is a little bit more old school. So this is what we're going to talk about over the next 10 minutes.

We're going to talk about the historical outcomes for our patients with MSI-high deficient mismatch repair colon cancer. I'm going to bring you back in the time machine, right? Because so much of this data came out in the early 2000s. And John Marshall likes to point out that our care or adjuvant therapy for colon cancer has really not changed since oxaliplatin was approved in the adjuvant space in 2004. That is a really embarrassing statement, right? When you think about how our care has evolved over the course of 2 decades. We're going to talk about historically, how do these patients do. We're going to talk about obviously the newest thing that's come up. And Dr Seligmann has really teed me up for this and she had brought up ATOMIC. So we're going to talk about FOLFOX in combination with atezolizumab for Stage III colon cancer. We're going to talk briefly about some of the trials that are coming down the pipeline.

So colon cancer and colorectal cancer is not without its advances, right? We know that there are over 20 FDA-approved regimens for the treatment of metastatic colorectal cancer. But this is the slide that I get to show for adjuvant therapy, right? And so I've given this lecture recently to my fellows. And I said, listen, if you're in GI clinic and it's a localized colon cancer and the attending asks you, what do you want to give? Just say FOLFOX. And if the attending just looks at you and starts shaking their head and just, well, you can say CAPOX after that. And that's literally the 2 regimens that we have. So of the 20 plus FDA-approved regimens in the metastatic setting, these are the 3 FDA-approved drugs currently for adjuvant colon cancer. And it's FOLFOX and CAPOX.

And to really bring you back in the time machine here, this is from JCO 2009, which is an update of the MOSAIC trial, basically asking a very simple question. Stage II or Stage III colon cancer, you're going to get infusional 5-FU/leucovorin or FOLFOX for 12 cycles. These are the curves for disease-free survival. The top 2 curves are Stage II colon cancer, really a nonstatistical significant difference in disease-free survival, but you do see a separation of the curves between 5-FU, which is that red line, and FOLFOX, which is the gray line. Again, embarrassing that this is from 2009.

When you look at the overall survival benefit, you see that it's not really there with oxaliplatin in Stage II colon cancer, but for Stage III colon cancer, there's a 4.4% overall survival benefit in Stage III. And Dr Seligmann had really touched on this too, but, well, how do our MSI-high patients do compared to the general population? And we know that the prevalence is higher in Stage II and Stage III colon cancer. The prevalence is lower for MSI-high in Stage IV. We think that hypermutated cancers are harder or too deranged to metastasize. But when you think about the prognosis, and I'm going to show you data on this, for Stage II we know that MSI-high is a great prognostic feature, but that's not true for Stage III and it's definitely not true for Stage IV, where it's possible that the prognosis with just standard chemotherapy is actually worse.

And so this was incredible data presented in 2010 and published by Dan Sargent and his colleagues, showing that the yellow line there, Stage II MSI-high colon cancer, those patients are going to do great. The prognosis is excellent. Amazingly, if you give them single-agent 5-FU, for some reason, there's actually a survival detriment because you gave them single-agent 5-FU. There's no difference in proficient mismatch repair with chemotherapy or nonchemotherapy in Stage II disease. What about Stage III colon cancer that's deficient mismatch repair? At least we feel a little bit better that they're receiving some benefit of oxaliplatin compared to infusional 5-FU. That's that yellow line on top of the purple line. And so it does seem like patients with Stage III MSI-high colorectal cancer do receive some benefit from FOLFOX.

So what's new, right? If I'm telling you that nothing has really changed in 20 years, well, what's new in the treatment of adjuvant colon cancer? Well, I'm going to have to go back to some really old school things and talk about exercise and talk about aspirin. So we can thank our Canadian colleagues that did the CO21 study answer a very, very simple but critical question. What happens if you take patients with Stage III or high-risk Stage II colon cancer and you randomize to receive health education materials, what we may give every single one of our patients in clinic, or actually give them a structured exercise program with a personal trainer? So what is a MET?

The goal of the study was to increase physical activity by 10 MET hours per week above baseline. You guys are getting 1 MET in just by sitting here for an hour, so congratulations. If you had 1 hour of brisk walking, that's 4 MET hours. So really, what is it that we're telling our patients to do? Go for a long walk 3 to 4 times per week. That's it, right? And so what's the difference in disease-free survival? There is a 6.4% improvement in disease-free survival if you get the structured exercise program versus getting health education material. For every 16 people, exercise prevented 1 person from recurrent or new cancer.

What about overall survival? An incredible difference between 90% and 83% and that overall survival driven by reduction in colon cancer death. And so when you put this effect size in context, there's our guy, adjuvant oxaliplatin, on the second row, and that's 5% improvement in overall survival after 10 years. Adjuvant exercise beats it, 7% improvement in overall survival at 8 years. And you can see how that compares to not only other things in colon cancer, but in non-small cell lung cancer and breast cancer as well.

Dr Cohen is going to talk a little bit about aspirin and colon cancer recurrence, but this is the ALASSCA study where patients were randomized to receive placebo or aspirin if they had group A, PIK3CA hotspot mutations, or group B, any other somatic variant in PIK3CA, PIK3R1 or PTEN. And this was the improvement in disease-free survival just with aspirin given 160 mg daily for over 3 years. And you can see an improvement in both groups in disease-free survival with aspirin. You can see that there are several more adverse events related to aspirin, but still an impressive disease-free survival benefit.

So what are some of the take-home points? Well, the take-home point is to exercise, and aspirin should be given to patients with a PIK3CA mutation. So the questions that I have for this group are, well, should we be hiring more personal trainers and giving less oxaliplatin? That's one. By show of hands, how many providers in the room are currently testing for PIK3CA in early-stage colon cancer? I see a couple of hands in the room. This is a paradigm shift, and it actually now shows up in the NCCN guidelines as I highlighted here in the red box. And of course, what's the optimal dose and duration of aspirin? We've asked this question even with neoadjuvant immune-oncology or immunotherapy in MSI-high colorectal cancer. We don't really know, but at least in the ALASSCA trial, we know it's around 160 mg.

Okay, well, that gets us to the meat of the matter. So really, what is new in adjuvant colon cancer? Well, of course, adjuvant atezolizumab for our patients with MSI-high colon cancer. So the ATOMIC trial, a very, very well-designed study, large trial looking at FOLFOX versus FOLFOX/atezo. In this trial, FOLFOX was given for 6 months. There was no variation off of that. We're going to talk a lot about how long we can give chemotherapy for different sets of patients. But in this trial, 6 months of FOLFOX, an additional 6 months, so 1 year total of atezolizumab.

The baseline characteristics of this trial are listed here. I just make the point that there were about a third of the patients that had what we consider to be high-risk disease, T4 or N2 disease. When you look at the primary endpoint, this is the major take-home point. There's a 10% difference in disease-free survival between the FOLFOX/atezo arm and the FOLFOX arm, suggesting that this is now the standard of care. For all the talk that we've had about neoadjuvant therapy, for many of those that are in this room and online, these patients are going to show up to your clinic after they've received their surgical resection. And this answered a question of what we should do for these patients if they come to us in the adjuvant setting. Across whether it was high-risk or low-risk, it seems like there was benefit that favored FOLFOX and atezolizumab.

As you might expect, whenever you add an immuno-oncology agent to chemotherapy, you're going to see a certain higher rate of adverse events. And so here, you can see that there's a slightly higher rate of Grade 3 and 4 treatment-related adverse events. And so to dive a little bit more into this, when you look at some of the toxicity that patients receiving FOLFOX/atezo versus FOLFOX, you're seeing pretty similar rates. There's just a little bit more adverse event rate in the FOLFOX/atezo arm, looking specifically at specific toxicities, mainly diarrhea and abdominal pain. And so when you look at some of the immune-related AEs, I think we're all worried that whenever we give an immuno-oncology agent that we're going to blow out somebody's adrenal gland. Luckily, that was something that was very rare in this trial. But you can see that there's a certain rate of diarrhea, colitis, hyperglycemia, hypothyroidism and rash that you would expect to see with atezolizumab. Luckily, there are no clinically significant differences in Grade 3 to 4 immune-related adverse events.

So here's your take-home point, right? I mean, and I agree with Dr Seligmann, and I really do believe that neoadjuvant immune checkpoint inhibition should be considered for high-risk disease up-front. But again, knowing that many of these patients are going to show up to your office following their surgical resection, FOLFOX/atezolizumab is the new standard in patients not receiving adjuvant therapy and you can kind of see some of the NCCN guidelines there.

Some of the questions that we've already even addressed with one of the patient cases, should we consider nonoperative management for MSI-high colon cancer if you're going to give them neoadjuvant immunotherapy? And I would make the argument, as Dr Cohen did, are serial colonoscopies better or worse than a hemicolectomy? I don't know if I could handle a colonoscopy every 3 months. That sounds like something that would be pretty difficult. What's the best duration of immunotherapy after resection? In the ATOMIC trial, it was 1 year. And obviously, is there a role for immunotherapy in proficient mismatch repair, MSS colon cancer? Dr Seligmann did a great job reviewing some of that data.

I'm just going to wrap up with 2 trials that kind of step on my colleagues' other trials. So I'm going to talk about a ctDNA trial, and I'm going to talk about a neoadjuvant trial. So the AZUR-4 study is a randomized study of neoadjuvant dostarlimab plus CAPOX versus CAPOX for high-risk Stage II or Stage III proficient mismatch repair or microsatellite-stable colon cancer. So here, it's really interesting. They're getting a lot of up-front therapy with chemotherapy and immunotherapy. This is a randomized trial. The primary endpoint's a very unique one. Its major pathologic response defined by less than or equal to 10% residual viable tumor.

To step on Dr Cohen's talk, you're going to see more and more of these trials in the pipeline and hopefully close to where you practice. And so what you're going to see are a lot of trials targeting the MRD space. What happens, and you're going to see this in our cases too, what happens when patients finish their adjuvant therapy, they request a ctDNA test and they're MRD-positive? What do you do for those patients? And we're going to talk a lot about this in our cases, particularly after Dr Cohen's session. But this is a trial of a personalized mRNA cancer vaccine. Essentially, patients have to be ctDNA-positive following surgery. After they complete their adjuvant therapy, they're then randomized to a personalized cancer vaccine or an observational arm.

I do want to highlight the crcmrd.com website. This is a really, really nice website. This is becoming an increasingly serious and common clinical scenario where we have patients that have completed their adjuvant therapy and are ctDNA-positive. What do you do for those patients? Well, we need to be able to offer those patients a clinical trial, and crcmrd.com is a great place for that. And with that, we're going to move on to our cases, but thank you very, very much.

Alright. So now having talked about aspirin, exercise but mainly FOLFOX and atezolizumab, we're going to start with this 58-year-old male, and we're going to talk about some of the high-risk features. So this patient had a resected colon adenocarcinoma, MSI-high, BRAF V600E mutated. What are some of the clinical pathologic features that most strongly influence your recommendation for adjuvant therapy in MSI-high disease? And does it really matter? Are you going to give the same treatment no matter what? And so for this first patient, we'll start with Dr Seligmann.

DR SELIGMANN: So first of all, I mean, I think that actually breaking down TNM is really important. I think that considering a T4 as a Stage II patient is very misleading. T4 is bad, and I think we need to be really careful about that, that we're not reassured thinking this is a Stage II patient with a T4 tumor. So I think we need to just be a little bit careful with TNM. But again, I think all of these features influence the recommendation. I mean, for Stage III combination, ATOMIC would guide what we do. The big question of ATOMIC is, what is the contribution of the chemotherapy? I can see why they couldn't, at the point, I think ATOMIC was, I echo your points, it was a wonderful trial, and it would have been too far at the time that the trial was being delivered to having an IO arm alone, but that's the real elephant in the room. Do we really want to give 6 months of FOLFOX along with immunotherapy?

DR LIEU: Yeah, we're going to have some cases —

DR SELIGMANN: I don't.

DR LIEU: So we're going to have some cases that ask that question. And I think that that is the missing arm. And I think you're right. I think when ATOMIC was developed, there was just no way that there was ever going to be an IO only arm, but that's the one that I think that we all really want. So in terms of skipping adjuvant chemotherapy, to that point, this is a 71-year-old female with uncontrolled diabetes, Grade 2 neuropathy, right? And so again, the concern there obviously is with oxaliplatin, has Stage III sigmoid cancer, again, MSI-high. This patient did check a postop ctDNA. It was negative, so that certainly makes us feel better. But are there situations where you would completely skip adjuvant chemotherapy and give only immunotherapy? Very, very difficult question. Dr Cohen, let you get a stab at it.

DR COHEN: Thanks so much. So if we, let's say we said that this was an MSS patient and we didn't have the option of immunotherapy, I guess the question would be, if you were looking just at the ctDNA as a prognostic risk factor, we would argue that that makes them less likely to benefit from chemotherapy, but not yet at a point where we would not recommend chemotherapy based on ctDNA alone. So I would say I still would ideally give that patient chemotherapy, though with that level of neuropathy, I don't know that we can give them oxaliplatin. If you switched it around and said they were ctDNA-positive and MSI-high, I would happily give them just immunotherapy and call it metastatic disease.

I think the hardest situation is this one in the middle where they're ctDNA-negative, MSI-high, you really don't want to give them oxaliplatin. I'm not sure I'm quite there for immunotherapy alone, but I could see the rationale. I could also see the rationale in observation in someone who's just not appropriate for chemotherapy, and I think that we will be data free for a while in this area.

DR LIEU: Would you ever give infusional 5-FU and atezolizumab? I see shaking of the head over to your right.

DR COHEN: I don't like that opportunity.

DR LIEU: Okay, so, and Dr Seligmann, you're shaking your head. Tell us a little more.

DR SELIGMANN: You showed us the data. You showed us that for some reason, biologically, 5-FU alone does not appear to be effective. So I think the tempting thing when ATOMIC came out was, oh, we'll just drop the oxaliplatin, but the data in the adjuvant setting has never supported 5-FU alone.

DR LIEU: It's a couple levels of not having data, right? So the 5-FU is one part of it, and then is 5-FU plus atezo is another layer of data that we just don't have. So but what I'm hearing is significant concern about single-agent 5-FU, even with the addition of atezolizumab. Okay, I think that's very, very helpful. Thank you for that.

Okay, so this is an older patient with several different comorbidities, postop ctDNA-negative. And so what you're looking at here is actually a practice pattern from a general medical oncologist. So this is a very, very interesting way of treating these patients, and would love to hear your guys' thought on this. So this practitioner does not offer patients with Stage III MSI-high colon cancer adjuvant therapy if ctDNA is negative, especially if older than 70 years old. And so what this practitioner does is actually wait for them to have metastatic disease and treat with an immune checkpoint inhibitor at that time. Luckily in the 6 years that this practitioner has been doing this, none has had disease progression. But this is a very, very different, and again, kind of evidence-free zone. Dr Cohen, what do you think about this practice pattern?

DR COHEN: I mean, I think it's great that it hasn't failed thus far. I'm not sure I'm ready to recommend it. I would say that in terms of chemotherapy, there was an analysis of MOSAIC that suggested that a 70-year-old did not really benefit from oxaliplatin, but it was still recommended to give them fluoropyrimidine. In IDEA, the 3 versus 6-month trial, they enrolled patients up to 85, so there wasn't a 70-year-old cutoff. And then there's been analysis, I think in the ACCENT database, that suggested that older individuals still benefit from chemotherapy. So I tend to still recommend sort of standardly if someone's fit enough, I give them whatever the recommendation would be for their stage.

That being said, I do also agree that sometimes the older individuals have more indolent disease, especially the MSI-high ones. So in that specific population, I would say for me, maybe patients in their 80s, I have omitted chemotherapy and just given them immunotherapy if they were ever ctDNA-positive or metastatic, but that also predated ATOMIC. So I wouldn't say that there's a standard approach to older individuals not getting them chemotherapy, but I think it is smart to be aware of the toxicities and what your patient can tolerate.

DR LIEU: Outstanding answer. Thank you. Another question, in regards to POLE mutations, in the NCCN guidelines, in the metastatic setting, immune checkpoint inhibitors are obviously approved in MSI-high, but also in hypermutated POLE mutations. And so that, but that's certainly in a metastatic setting. In this particular question on the bottom, is a 55-year-old gentleman with a POLE mutation and oligometastatic disease to the lung. So Stage IV, but now without evidence of disease. And so I, the question here is one that's actually very complex. Now, POLE mutations are rare. And so I've only seen a handful, even in my entire career, but would you consider that to be similar to MSI-high, knowing that that wasn't exactly studied in the ATOMIC trial, Dr Seligmann?

DR SELIGMANN: I think that's probably quite reasonable. So it's the hypermutations and the ability of the MSI-high tumors to become aware, to be recognized by the immune system that makes them able to use immunotherapy. And so it's exactly the same principle comes into POLE cancers. So I think we're never going to have perfect data for these patients because they are so rare. So I would be very comfortable in treating them in the same way as an MSI-high patient.

DR LIEU: Yeah, I agree. And I think the biology is similar. And for this patient, you know what the biology is like. So it does make sense to treat it much like MSI-high. So yeah, really, really great question. Dr Cohen, you already had addressed some of this with older patients, but, 92-year-old female or an 83-year-old female with Stage III MSI-high colorectal cancer, would you, we had had this discussion of, do you really need the 6 months of FOLFOX and obviously the 1 year of atezolizumab? Would you treat with a single-agent immune-oncology agent?

DR COHEN: Yeah, I think that in some ways, as we already had talked about, we're looking for excuses to use immunotherapy or to have an indication for immunotherapy. And to me, positive ctDNA could stand for micrometastatic disease and I would use that as a rationale to treat this 92-year-old individual if they wanted ongoing therapy. Depending on their health status, you could argue maybe you don't do anything, but we're going to assume that this person was fit enough that they underwent a surgery. So maybe they'd be fit enough to have some systemic therapy. So I would say that I have definitely given older individuals immune checkpoint inhibitors and typically would do single agent, again, with the slight rationale that maybe there's less toxicity involved.

DR LIEU: Dr Seligmann?

DR SELIGMANN: I think that you can, and I think patients, older patients can. I'm just thinking of the IMHOTEP trial. So this was a neoadjuvant IO trial with single agent. There were 5 deaths in that trial, not so much related to the primary tumor. So I think we do need to have a bit of caution. They are generally well tolerated, but honestly, I don't think I would be offering a 92-year-old adjuvant anything.

DR LIEU: Anything.

DR SELIGMANN: But I think when we push the boundaries and we're starting to treat PS3, PS4 patients, even when there's a potential of really good responses, there does need to be some caution. And certainly when they looked at the IMHOTEP data in detail, they saw that the patients that did die were out with the recommended inclusion criteria. So it was pushing the boundaries because we're so excited about immunotherapy. We just need to remember, we can still cause toxicity and problems. I mean, there's some Grade 5, particularly things like myocarditis that we're starting to see as well. So they're not without any problems.

DR LIEU: Yeah. This is a question that's going to come up a lot about ctDNA-guided therapy. And so Dr Cohen, we're going to get to your talk here in just a bit. And this is going to be similar to many of the questions that we're going to see after your discussion. But a 37-year-old gentleman with Stage III colon cancer wants to be aggressive with treatment, is interested in considering immunotherapy. He asks if we can skip chemotherapy if his ctDNA remains negative. He does want benefit of immunotherapy maintenance. So based on the ATOMIC trial for a young patient who wants to be aggressive, would you give atezo with chemo if ctDNA remains negative in a postop period at 1 month and 3 months? Would you skip additional chemotherapy and continue only with immunotherapy?

And this is actually a really, really great case because I actually had a young patient getting FOLFOX/atezolizumab and really had a problem with oxaliplatin with worsening neuropathy. And the question was, well, can we stop at 2 months or 3 months? And obviously these are really, really hard questions to answer, but, in your experience, in the real world, right, because in the trial, 6 months of FOLFOX, 1 year of atezolizumab, how does that look like in your clinic?

DR COHEN: I mean, I think it's an easy answer because we just simply don't know. And so the way I view Stage III with ctDNA is you still follow stage recommended adjuvant chemotherapy. And then if someone is positive, you're more likely to want to push them through and get the full extent and see if you can help clear that ctDNA and get better outcomes. If they're negative, you may be more inclined to decrease or dose reduce or stop early in the setting of toxicity. I don't otherwise do it, but I think to some degree, you're never required to have adjuvant chemotherapy. We can never prove that that person's going to benefit. And so in a person who understands the choices, I think you talk to your patient, but I wouldn't routinely stop at a specific time period just based on the ctDNA.

DR LIEU: Yeah, and I think that we always feel a little bit more comfortable, I know we're going to talk about this with escalation decisions based off of ctDNA, but really have a hard time with de-escalation decisions just because we just don't have that data. And I know that you're going to talk a little bit about that.

Role of Circulating Tumor DNA Testing in Localized CRC — Dr Cohen

DR LIEU: All right, so with that, we're going to actually move on to ctDNA, and then we'll have a lot of questions based off of what to do when you have patients that are ctDNA-negative and ctDNA-positive. So go ahead, Dr Cohen.

DR COHEN: Okay, well, I appreciate all of the setup, and I do only have 10 minutes, so I did not do every single thing that you told me that I was about to talk about, but I did some of these things. So we're now going to talk about circulating tumor DNA or ctDNA for localized colorectal cancer. So I'll go over ctDNA as a biomarker and then walk through some of the data that we have and future directions of ctDNA technology.

So circulating tumor DNA or ctDNA is a biomarker for disease activity. So we all have cell-free DNA that's free-floating in our bloodstream. It's due to cellular breakdown from a variety of different causes. But when you look at the specific fraction that's from the tumor, you're looking for mutations that are specific to the tumor, that's ctDNA. And so we look at genetic changes, epigenetic changes, difference in fragment sizes. And so there's a lot of different hallmark changes that allow the different testing platforms to figure out what is ctDNA and what is the regular cell-free DNA. And it has a very short half-life, and so that allows it to be a very dynamic biomarker and to give you a real-time look at what's going on.

There's sort of 2 major approaches. One is called tumor-informed, and one is tumor-agnostic or tumor-naïve or tissue-free. All are sort of synonyms. But the idea is that in a tumor-informed approach, you're actually taking a specific patient's tumor and you're looking for specific genetic alterations in that tumor that you then try to match up in the fragments in the blood. Whereas in the tumor-agnostic approach, you are actually looking for more general changes that you expect in tumors overall. It may be cancer-specific or broadly across cancer types. And this is relevant not just in terms of the sensitivity of the different approaches, but also the turnaround time. So clearly the blood testing takes some amount of time, usually about a week, but if you add that tumor approach in, you're going to add several weeks of time, meaning that that first result is going to have a longer turnaround time. Thankfully, subsequently, if you do longitudinal testing, it'll be pretty similar.

And so we know that single-timepoint testing is helpful, but we can also use this as a biomarker over the course of disease. And so this is showing 2 different scenarios, a smaller tumor on the top and a larger tumor on the bottom. And that's relevant because the amount of shed of ctDNA is tumor-burden dependent. And so you can imagine that if someone has a higher cancer burden, that they are going to recur in a more rapid timeframe, but also may be more likely to have a ctDNA test that is detectable at an earlier time. And you can look at those quantitative levels, you can look at the doubling time to try to get an assessment of disease biology.

And this leads us to the idea of MRD, which is minimal residual disease or molecular residual disease or the idea that you're looking for a small volume of cancer not appreciated radiographically or with other clinical measures, with the idea that ctDNA could potentially pick up recurrences faster than would be detected by imaging by CEA. And so we're going to talk about some of the retrospective and then moving into prospective data.

I'll skip over our retrospective, we'll come back to that, but I want to talk about 3 large observational cohorts, GALAXY, BESPOKE and INTERCEPT. So there have been many very large skill efforts. There have been tens of thousands of patients who have had ctDNA testing, and some of that's been within real-world practice, but some of it has been within observational studies for which they would enroll patients who are fully resected and get serial ctDNA monitoring for up to 2 years. So we have the GALAXY study as a substudy of CIRCULATE-Japan with over 2,000 patients in Stage II and III. And then we also have, sorry, I should go back and say that there were opportunities of interventional trials that we'll come back to within this study for patients who are negative or positive. And then we have the BESPOKE study, which was done in the United States, which again allowed 2 years of monitoring as well as PROs and other time points to truly try to evaluate ctDNA testing.

And from this we can see Kaplan-Meier curves that are never really seen in terms of interventions, which is that ctDNA-negative patients do much better than ctDNA-positive. We see this on the left with a tumor-informed assay and on the right with a tumor-agnostic assay, but showing that we really see a polarizing, prognostic implication of ctDNA. And it matters at that single time point, but also for longitudinal testing. So on the left I have GALAXY, on the right I have BESPOKE. And both of them show that patients who are negative and stay negative do the best. Again, patients who are positive and stay positive do the worst. And then you're going to have some patients who have a dynamic change, a transient clearance. They're going to do a little bit better, but if you become positive, those patients still had poor outcomes.

But as I said, what matters is what we actually do about this testing, because it really doesn't matter to just have a prognostic test that we can't act on. We want to actually help patients and be able to treat them in a better and more efficient manner. So the MD Anderson INTERCEPT study is a single institutional study collecting patients with ctDNA evaluation after surgery from a variety of different stages. A portion of these patients went on to become ctDNA-positive and they noted that their provider would typically order subsequent testing, especially CT, MRI or PET scan. And when they did this, half the time they found that there was evidence of metastatic disease and the other half of the time they were MRD-positive or no evidence of radiographic disease.

And so when you look at this in terms of lead time, the median lead time was 3 months. But if we took out that half of patients who basically immediately were found to have radiographic disease, the lead time was about 5 and a half months, suggesting that ctDNA, again, is prognostic, but may also highlight patients who are going to go on to fail radiographically. But again, none of this matters unless we can actually act on that and really try and change outcomes based on ctDNA. So we're going to talk about how to deal with this in the adjuvant setting and in the post-adjuvant setting.

So the first major, and really probably the major study that has been shown prospectively is the DYNAMIC study. And this was done for Stage II colon cancer, randomizing 2 to 1 to ctDNA-guided management or standard management, meaning that these resected patients had blood draws at 4 and 7 weeks after surgery. And in the ctDNA-guided arm, if they were positive, they went on to have adjuvant chemotherapy. And if they were negative, they were observed. Whereas in the standard management, it was up to the provider to discuss with their patient whether or not chemotherapy was indicated. And they would follow them with a primary endpoint of 2-year recurrence-free survival.

And this was a non-inferiority study, and it showed that clinically, there was non-inferior outcomes between the 2 groups. There was similar recurrence-free survival in patients who had ctDNA-guided management and patients who had provider-guided management. But interestingly, the ctDNA-guided arm had half as much chemotherapy as patients in the standard management. But yet, the ctDNA-guided arm had much more oxaliplatin therapy as opposed to standard where most patients who got chemotherapy got fluoropyrimidine alone.

And so we have to keep that in mind that chemotherapy also does not fix the problem. So we know that adjuvant therapy is a risk reduction. It is not absolution of the risk. And so what they showed is that when they look in the ctDNA-positive patients, patients that cleared, did better, but not 100% of the time were they disease-free. And patients who were persistently positive certainly did worse. But that means that not everyone benefits from the ctDNA, excuse me, from the adjuvant chemotherapy. And so that then suggests back to what you've already heard about, that maybe we need to pick a better adjuvant chemotherapy.

Now we also know that not all studies have shown these amazing outcomes suggesting that we could really guide chemotherapy based on ctDNA. So the DYNAMIC study was done with a tumor-informed approach. The COBRA study, also done in Stage II colon cancer, was using a now older tumor-naïve assay. They saw in this very low-risk patients, just T3aN0, they showed that a very small amount were ctDNA-positive, but some of those patients spontaneously cleared. And they ended up stopping this trial early for futility because they weren't really able to show that ctDNA-guided management was better.

Now moving forward from the Australian group, again using the tumor-informed assay, we have the DYNAMIC-III study where they randomized Stage III patients. And again, these are those folks that we would pretty much recommend chemotherapy to everybody. They randomized them 2 to 1 to ctDNA-guided management. But here, being positive was a cue to escalate the chemotherapy, but being negative was a cue to deescalate. And one issue with the study, even though again, it was a very amazing study and really groundbreaking for ctDNA research with a survival endpoint, which is what we need, the problem is they didn't mandate the type of chemotherapy. And so it was a step up from whatever was recommended, but that could be anywhere from no chemotherapy all the way up to 6 months of a doublet.

And here, when they looked at recurrence-free survival, even though again, we were looking at patients having a slight modification of chemotherapy, patients who were ctDNA-negative did better with standard management, patients who were ctDNA-positive did better with standard management. So again, this gives us pause that we cannot just use this biomarker by itself. We need to still remember all of our clinical risk factors and think about ctDNA as being one of them, not the only one.

And so is it the assay? Because this is not our standard tumor-informed assay that we're often using. Is it the heterogeneity of the cohort? Is it the specific regimen that was used? What is the reason why this was negative? We don't know. But thanks to Dr Lieu and others, we have CIRCULATE-North America where we're hoping this will also help answer the question. Again, a Stage III trial looking at patients for a de-escalation question for patients who are negative, an escalation question for patients who are positive. But here we have much more control of the chemotherapy regimen used, and the hope is that this will start to address some of the questions.

And as I hinted at, we also have the opportunities for after standard adjuvant therapy. So in the ALTAIR study, this was part of the CIRCULATE-Japan, they looked at patients who were postoperative ctDNA-positive. Many of these patients had had neoadjuvant chemotherapy, about one-third, and about half of them had adjuvant chemotherapy, and then they randomized patients who were still positive to FTD/TPI, which is trifluridine/tipiracil, or placebo. And so here, this was again turning out to be an underwhelming study where they did not see a statistically significant difference between the 2 arms. Maybe more so when you looked in Stage IV patients, but again, this gives us pause. Is it the assay? Is it the chemotherapy? I think a lot of this may be the chemotherapy that's chosen. Is it the heterogeneity of the patients? Because some of these patients would have been exposed to similar chemotherapies and others not.

So that then puts us back to saying, well, what are some new questions that we can answer? So let's go back and ask some of the survivorship questions that have already been asked. So you already heard that the ALASCCA study was a trial establishing the role of aspirin versus placebo for patients who have a PIK3CA mutation in colorectal cancer. But so we also have to think about other opportunities where we've evaluated this. So in 80702, this was a study of 3 versus 6 months, they actually, of adjuvant chemotherapy, they did a second randomization to 3 years of celecoxib or placebo.

And so on the left, you see that the patients who were ctDNA-negative on this trial are those solid lines at the top. They did the best, more so than patients who were ctDNA-positive. But it seemed like the celecoxib benefit was more so in the ctDNA-positive patients. And even when they restricted to PIK3CA mutant patients, which was about one-quarter of patients, and that's the graph on the right, you still see that that was really a difference that was noted. So this then questions, again, should we be making blanket statements for adjuvant therapy or should we really be further honing these questions?

But finally, ctDNA is not 1 test, it is not 1 platform, it's not 1 assay. We're seeing constant changes, and these are things that are both exciting but also make it problematic to do trials because the technology is changing faster than the research. So some of the tumor-informed approaches have used 16 variants, and now we're starting to expand panel sizes. We can do a different depth and breadth of sequencing, like going from whole exome to whole genome. We can look for the improvement or inclusion of epigenetics, especially in those tumor agnostic assays that has helped improve sensitivity. Looking for cancellation of background noise like CHIP. And then thinking about before, maybe originally we were saying positive or negative, but then maybe quantitative levels, and then really can you look at full differences and using it as a biomarker.

So I'd like to end by saying I think that ctDNA is part of the future. I don't think we totally know how to use it, and the problem is it's changing as fast as the research is changing. But the hope is that eventually we're going to really be able to deescalate those low-risk ctDNA-negative patients, then we're going to be able to escalate and more appropriately treat the positive patients. And then even more so, we're going to be able to watch this over time, and depending on how a patient responds to treatment, we can further tailor their therapy. Thank you.

DR LIEU: Great. Thank you so much, Dr Cohen. Alright. Outstanding talk, and a very, very complex field, and I love that point of the fact that the technology has literally outpaced our ability to be able to know what to do with some of the data that we're getting. But here we are. We know that the United States is a little bit of an outlier in the fact that many, many people are getting this test. Either they're paying for it out-of-pocket, or they're getting insurance approval for it, and I know it's a little bit different in the UK as well.

But these are real-life cases, and we're going to start really with Stage II. I'm going to ask both of you to answer this question. For a 65-year-old with Stage II colon cancer, how comfortable are we in recommending chemotherapy, presumably for a patient that we typically wouldn't give Stage II colon cancer chemotherapy, but what if they were MRD-positive? And Dr Seligmann, if you're kind of left seeing this patient, low-risk Stage II MRD-positive, would you allow that to guide your decision-making?

DR SELIGMANN: I mean, I think it's an extremely powerful prognostic tool. I think GALAXY is very convincing. It's a very powerful prognostic tool, and at some point, it will not surprise me if it's added to the TNM. And for that reason, I probably would be quite comfortable doing that. Are we going to get really robust data for this situation? Probably not, but if you really trusted the assay, and I think, again, we have to remember, we have to really trust an assay, then I think that's probably quite reasonable, because I think being MRD-positive is probably more of a predictor of relapse than nodal positivity, for example. So probably, yeah.

DR LIEU: Dr Cohen, same question, but also the last point on the slide is really great. What's the, tell us about the practicality of waiting for the test results. So first question, would you allow the ctDNA results to influence your decision-making? And then tell us what the waiting looks like in reality?

DR COHEN: So I would use, I think Stage II is probably like the perfect case point for using ctDNA, because I think our clinical risk factors are poor. We've lumped high-risk disease to include a number of different factors, some of which are more prognostic and some of which are more predictive, but we really don't know the differential benefits. So I think ctDNA is very powerful in that setting. I would say, as has been referenced before, a T4N0 patient I would treat with adjuvant chemotherapy. And so I would do that sort of regardless of the ctDNA.

For anybody who's T3N0, I feel comfortable getting ctDNA. I feel comfortable waiting for those results. Now, I prefer a tumor-informed approach right now, but if someone is needing a result faster, perhaps you're seeing them already 8 weeks after surgery and you're thinking, I really need to start them in the next couple of weeks, I've also sent more than 1 test, which I guess is maybe a US approach to the problem, which is to send a tumor tissue-free test that you're going to get more quickly and a tumor-informed test that may be more sensitive. So I think it can be helpful. Certainly, if you get a positive test, I would believe that positive.

DR LIEU: Yeah, absolutely. So it sounds like, for both of you, and I think it's true for me as well, if I see ctDNA positivity because of the prognostic implication, also with the data that Dr Cohen showed, that if you can clear it, and I also want to point out the fact that adjuvant chemotherapy in its current form isn't a cure-all or fix-all for all patients as well, but the idea of being able to clear it and clear it kind of in a sustained manner, we know that there's, at least from the GALAXY data, some survival benefit, at least disease-free survival benefit.

This is a very, very similar case, but the question being, if, assuming all things are approved and you're able to get whatever test, Dr Seligmann, for low clinical risk Stage II, would you check ctDNA on everybody in that scenario?

DR SELIGMANN: So, I mean, I agree. I mean, I think that the MRD is most relevant for Stage II. I think the data that you didn't have time to show was actually for the high-risk patients. You really would not feel comfortable using ctDNA to make decisions. You are going to give chemotherapy. It will be a prognosticator. But really, I completely agree with you. And the road forward for these decision-making tools is going to be with Stage II patients, because people do relapse. And so I think that would be quite a reasonable thing to do. It's very T3N0, of course, very borderline, and as you say, the other risk factors are really not that helpful, so MRD positivity in that situation, absolutely.

DR LIEU: Yeah, absolutely.

DR SELIGMANN: And combination as well.

DR LIEU: Yes. Agreed. And so, really, really great discussion on Stage II, and I think you're hearing a lot of alignment in terms of how we approach Stage II and agree about the point that you guys have made also about T4. It's just, it's such a high-risk factor. It's hard to make de-escalation decisions based off of that. And so, to that point, I'm going to ask you guys a question about de-escalation in Stage III, and I think I know what the answer's going to be.

But if you take a 72-year-old female with standard risk Stage III colon cancer that's ctDNA-negative, if the patient would prefer not to receive chemotherapy in this setting, do you think it's a reasonable option to omit chemotherapy if the MRD tests are negative? So, we talked about escalation. I think that's a much easier question as it pertains to Stage II in a situation where you typically wouldn't give chemotherapy. Well, what about de-escalation? And we'll start with you, Dr Cohen.

DR COHEN: I would recommend chemotherapy in that setting, and at this point, I still think it's the right recommendation. And as I mentioned before, I don't think adjuvant chemotherapy is for everybody, meaning that a patient still has the right to say, I hear your recommendation, I don't want to do it. In that scenario, I would feel more comfortable if the MRD test was negative, and I might, if they've basically decided that they are not getting chemotherapy, I would ask the reverse question, that if it was positive, would they consider doing chemotherapy? And I think, at the end of the day, that's a good way to discuss it with your patient who sounds like has sort of already made up their mind that they don't want to have chemotherapy.

DR LIEU: Right. And Dr Seligmann, I'm going to ask you the same question. I do love the point about, I have flipped that question on its head and said, it sounds like you don't want to do chemotherapy. If the tests were positive, would that then get you to actually get chemotherapy? And then if the answer is, yes, it would change my decision-making, then I usually do send it, so, Dr Seligmann.

DR SELIGMANN: Well, if the patient doesn't want chemo, they don't want chemo, and I think, well, how helpful is then doing the test if it's not going to influence what you do? So I think we could get to the point where we can do these tests all the time. There is a health economic question here as well. If it isn't going to change what you're going to do anyway, is this information either helpful for you or helpful for your patient?

DR LIEU: Yeah, no —

DR SELIGMANN: I think it's very difficult.

DR LIEU: Yes, I totally agree. Alright. So this is a patient with neuropathy and Stage II colon cancer. Oncotype recurrence score is high. I haven't seen that test done in quite some time. ctDNA for MRD is negative. Kind of similar question about omitting adjuvant therapy. I know we've addressed some of that. Here, the combination of ctDNA and MRD being negative, but with the oncotype recurrence score, there's just no data to guide how to use both of those in combination, but certainly there's a lot of data with the ctDNA negativity in Stage II colon cancer with the DYNAMIC trial that Stacey presented. And then a 68-year-old with left-sided colon cancer about to start adjuvant FOLFOX and initial ctDNA testing is positive. Should the positive ctDNA test at the initiation of adjuvant therapy impact what it is that you do? Dr Cohen, I know it's always scary when you see that, and certainly we have patients that test ctDNA-positive postoperatively.

DR COHEN: Right, I think we should make the distinction that if you have, and I know this isn't stated here, but if you had an initial test that was preoperative that was positive, that is not prognostic. But the MRD space is that time window after surgery and before you start chemotherapy. So that is the binary positive/negative that pretty much all of these studies are reporting out. So I would take a positive ctDNA at the start of adjuvant therapy to mean that this patient is at very high risk for late recurrence, and I would consider following that. We know that we can get false negatives on chemotherapy just from suppression of the ctDNA shed.

So I don't necessarily follow it on chemotherapy, but I have definitely checked it at the 3-month mark because most patients I'm giving 3 months of chemotherapy, and so maybe I'm checking it at 3 months to say if it's still positive, maybe that's going to have me have a conversation with that patient about would they be willing to consider 6 months of chemotherapy. So I think it can be useful in some ways, especially as you're weighing toxicity of treatment versus benefit, but I think at this point, every Stage III patient should be getting chemotherapy, including positives and negatives.

DR LIEU: Outstanding answer. And Dr Seligmann, would postoperative ctDNA positivity influence anything that you would recommend in the adjuvant setting?

DR SELIGMANN: I think, again, we're seeing, so from the IDEA data, we would be a bit nervous of giving, particularly if you're using FOLFOX, 3 months. So for T4 and N2, so the high-risk categories, surely ctDNA positivity falls into that category as well. So I think I would be planning to give them at least 6 months 5-FU, maybe dropping the oxaliplatin at 3 months. So I don't think a ctDNA at 3 months would influence at all what I did. And I think we have to be, again, just mindful of ordering tests when it's not going to change what we're going to do. But again, we get more experience as we go along the line of how we should be handling results and how we communicate those results to our patients. It's a scary thing.

DR LIEU: Yeah, and really what you're speaking to is this idea of we really need more data about ctDNA kinetics, rate of clearance, sustainability of that clearance and then from there, we can start making some treatment decisions. But I think part of the issue is we just don't know, right? Transient clearance we know is bad. Persistent clearance we know is good. But what about the rate and the timing of all that? And it really, really is a completely data-free zone. And I really appreciate your guys' answers because it really just shows the fact that you're really making treatment decisions based off of very, very limited data.

We had already answered some questions about older adults. So this is an 82-year-old that's worried about chemotherapy, what MRD assay results help you with treatment decision-making. And I see a little bit of a shaking of the head from Dr Seligmann. Any thoughts about that?

DR SELIGMANN: Some decisions don't need an additional test. A patient who is 82 with comorbidities, I'm not going to offer them adjuvant chemotherapy, full stop. It doesn't matter what the assay is. I'm not going to offer them adjuvant chemotherapy. That would be inappropriate.

DR LIEU: And what I really want to end on is a situation that I think more and more practitioners are really dealing with and that is these 2 cases, right? So if you have a patient with Stage III colon cancer and there are 2 scenarios here, 6 months of FOLFOX and they're post adjuvant therapy, they have a positive ctDNA. The second case is a patient that is ctDNA-negative at 1 year, and obviously that's always great when we get that, but then returns low positive, right? And so these are real clinical scenarios. I was actually dealing with this in clinic even this week. And I would love to hear from both of you what your approach to this is. When these patients come, either completed adjuvant therapy, remain persistently positive, or they're negative and then they turn positive, which is also devastating. So would love to hear what you guys do in this clinical scenario.

DR COHEN: So I would say for a ctDNA-positive, not necessarily in the MRD space, but now we're somewhere in the surveillance time window, if that patient had had localized disease, I would of course offer them a clinical trial for the ctDNA biochemical recurrence if that's available, because there are a few of those studies, and I encourage you to look for them if you're in this scenario. But excluding that, I typically do not give systemic therapy. And the reason why is because I feel like we need to figure out where that recurrence is actually happening. At this point, that patient has not had any documented radiographic recurrence of metastatic disease, so I would start by getting them an imaging study. Typically I would get their surveillance CT at that time.

If that still is unremarkable, I might get them a PET or a PET/MR, because we know that the liver has the highest shed. But then this would really prompt me to start getting imaging every 3 months maybe, instead of every 6 months, and trending that ctDNA result. We know that every once in a while, you can get a spontaneous clearance, but those are really quite rare, maybe 1 to 2%. So this would really just have me start to put eagle eyes on that patient and say, we think that this person is at very high risk of recurrence, and we need to figure out where, because maybe it's an isolated liver metastasis, and you just resect it, and then you go back to monitoring them.

So I'm hesitant to jump in and do systemic therapy, because then you don't know how long to treat them, what's your endpoint, when were you successful, did you just most likely just delay time to it showing up again? I would say the flip side though, is if someone had Stage IV disease, which you treated to NED, and then they had a positive ctDNA, and those patients with prior metastatic disease, I have given systemic therapy to.

DR LIEU: Yeah, really, really great answer. And briefly, Dr Seligmann, anything else to add to that?

DR SELIGMANN: Look, we don't have, I think the one thing that has been clear here is that ctDNA doesn't replace actually trying to find new strategies to cure more patients. I think it has a place in the treatment algorithm, but we have to remember, we need to push the boundaries more if there is more agents coming through, whether it's the ctDNA-positive space versus the neoadjuvant space, we will see. But I think the importance here is that we remember that a lot of the agents that were tested in the adjuvant setting were tested a long time ago, probably not using the right end points, probably not in molecularly selected populations as we can do now. So I think I'd be hesitant to say none of the agents work in the localized setting. They should work. We know that molecular status is important. So I think we need to continue to be ambitious for our patients, push the boundaries. I don't disagree with anything that you said. I think we need to remember the impact on our patients, though. It's very hard to basically say, we know that this is going to be a problem and these results mean something. So again, there's just, there's very complex.

DR LIEU: Yeah, and this is a field that will evolve tremendously over the next several years and really want to be able to offer every single one of these patients obviously a clinical trial because that's really what we need. I really want to just give a round of applause and thank you to our fantastic panelists.

DR SELIGMANN: Thank you.

DR COHEN: Thank you.

DR LIEU: And thank you all for taking time out of your evening to talk about localized colorectal cancer.

And with that, I hope everybody has a wonderful GI ASCO 2026, and we'll see you again soon. Thank you.