Molecular Residual Disease (MRD) Testing in Colorectal Cancer (CRC): Neoadjuvant Setting (Round 1)

To hear the arguments related to the adjuvant setting, see Round 2.

To hear the arguments related to the surveillance setting, see Round 3.

Dr. Jason Zell: All right, thank you. Yeah, very nice to be here. So, I'd love to introduce this topic, and let's see if we can go. I do believe ctDNA, especially in the MRD setting, is a game changer. So, I want to go through why I think that. Do I think it's, you know, the final word is set in stone? No, definitely not. So, we'll go through where we have holes in the data and where we're headed. But I was very honored to participate in this joint colorectal task force paper. It was a white paper in 2020. And the interesting thing about this paper for me is, at the time we wrote it, we really put about two sentences on that section to the left of this curve, which was colorectal cancer screening. And to my great shock, that is the one test that's FDA approved and out there is colorectal cancer screening. So, we really didn't see that happening in five years. So, you really never know how fast things move. The talk right now, we're going to focus on the neoadjuvant setting, which is shown here. Everything above the line of clinical detection is, you know, in red on this graph. And it shows where we have a patient with disease. We want to use ctDNA if we can to make some decisions in this neoadjuvant setting. So, what do the data look like? There's really three advantages, and they all point towards the last one. Number one would be improved risk stratification. Obviously, if we could identify a cohort of patients that are at higher risk for complications, or maybe need dose intensification or dose de-escalation, this would be wonderful, right? This is what we're all looking for. It gives us the opportunity to intervene. When we're talking about neoadjuvant setting, a lot of that is not colon cancer. It's always rectal cancer. So, when we talk about—I mean, rectal cancer has undergone a transformation in the last five years. We're not even operating on patients now, certain patients, right? So, everything— some we don't give radiation. Some we don't give chemo. Some we don't give surgery. It's incredible. But all of this is pointing to a personalized treatment approach. And so, hopefully, we'll see where we are and where we are not. This is what we have. So, we have a large—for rectal cancer, this is the largest clinical trial of rectal cancer in the neoadjuvant setting. And the reason why I want to focus on this, despite FOxTROT and colon neoadjuvant, or NICHE trial, which is MSI high patients primarily, really in this setting, most of us are treating rectal cancer patients in the neoadjuvant space. The other thing I'll say, and it's a nod to my challenger here, I don't think there's a role for ctDNA in MSI high patients right now. And the reason is there's such an overwhelming response of immunotherapy. So, if we look at NICHE trial for colon cancer patients, most everybody responded to upfront immune checkpoint inhibition. You can do ctDNA or not, but it's not probably going to change your treatment. And that was just after four weeks. So, by the time you get the ctDNA back, the patient's already had surgery and had an excellent response. So, we're going to focus on MMR-proficient rectal cancer, right? MMR deficient, we're going to treat with immune checkpoint inhibition anyway. So, but this is the population where we have data. Locally advanced rectal cancer, stage II or III, T3, T4, N+. These are treated in the standard approach in Australia. This is Dr. Jeanne Tie’s group, an outstanding group from Australia. What kind of assay do they use? They use something that they call the Safe-SeqS assay. It’s their proprietary test, was developed at Johns Hopkins through their GI spore. So, the important thing for this group is we have two main blood types in this area. One is a tumor-informed assay where we're sequencing the patient's tumor, picking the top hits, developing a blood test for that patient. And then, only looking for those hits in the blood. The tumor agnostic approach is another one. And we'll talk about both types in here where we have data. So, this was a proprietary type of tumor-informed assay. Patients were randomized in a two-to-one fashion to either ctDNA informed management or standard management. And what does that mean? So, if they're in the ctDNA-informed group and they had positive ctDNA, they would be assigned to get four months of adjuvant oxaliplatin-based fluoropyrimidine therapy. So, either FOLFOX, 5FU, something. If they were ctDNA negative and had no evidence of pathologic lymph nodes at surgery, they were put in the observation group. And if they were ypN+, these are people you treat with chemo upfront or chemo RT, and they still have a persistent lymph node positive? That's a very high-risk group. We've always considered that to be a high-risk group. That was left to the clinician's choice, probably wisely, because the investigators did not know what to do with that group. So, this was the design. The primary endpoint is very interesting. It's the proportion of patients receiving chemo, which as you'll see in colon, they did a similar design from the same group. But the key secondary endpoint is the three-year RFS. And I would love to be able to point to positive results here, especially in a debate, but unfortunately, this trial did not complete accrual. Their target size was 408 patients, and it was cut early at 230 due to slow accrual from COVID. But probably just as much because of the changing paradigm where everybody gets total neoadjuvant treatment, everything upfront. We're not treating with this approach anymore, with chemo rounds followed by surgery, followed by adjuvant chemo. It's just not done as often. So, with that, the last year at ASCO GI, investigators presented their data. The two groups, although they're two-to-one randomization—so the ctDNA-informed management group is larger—they were very similar in terms of their basic characteristics. And there were some interesting findings. So, I think the first thing to note is that ctDNA positivity, like what percent of rectal cancer patients would have ctDNA if you look for it in this population? It was 28%. That's the percent with their assay, which is a unique assay. And it was strongly associated with lymphovascular invasion, as you might predict. And what did they find? Actually, this looks really good. If you look at the slide, it looks like they proved their primary endpoint. There was much less chemo given in the ctDNA-directed approach. Only 46% of these patients had adjuvant chemo versus 77%. The other thing of note here, unlike the colon cancer study, the oxaliplatin use was about the same in both groups. So, there wasn't more oxaliplatin given to the ctDNA-informed group. So, this was really excellent results. The problem is, number three on here, small sample size precludes any conclusions to be drawn about noninferiority where you need to be kind of overpowered. This study was underpowered. So, you know, we can't say much more about this. But I would say, still, this is game changing in the fact that it shows us the trial can be done. There's a unique space to look at MRD assays in rectal cancer. And, unfortunately, this one due to power just is not going to give us the final say on the answer. Okay, second, and the last trial I'll talk about in this space because we don't have that many. This is conducted by my colleague Hagen Kennecke and the Canadian group CCTG up in Canada. The Neoadjuvant Excision Observation trial or so-called NEO trial. And this is replaced by the current NEO RT trial. For those of you who are accruing, you can put patients—we put several on, in my own institution. This trial was a unique trial, T1 to T3, N0. So, think early stage, even stage I rectal cancer. Well, we don't usually, you know, do anything but surgery. But, in this approach, the idea was for organ sparing. And that's where all the treatment in rectal cancer has gone, trying to either avoid the APR or LAR even because—or go into resection or an APR—because patients have terrible life consequences from losing their rectum, right? This is—the goal here is to preserve function of the rectum. So, patients, you can see 53 patients with ctDNA were given three cycles of FOLFOX. It's actually six cycles, so three months of FOLFOX. Then, they went to transanal excision, which is through the rectum instead of going through the abdomen. So, you know, the patients come out with no scar. And interestingly, if they had evidence of incomplete response or some kind of progression, 10 of them, just 10 of them went on to TME. So, this study on face was a success. In the ctDNA group, what they found was this. Overall, a full 46% had positive ctDNA pre-chemo, and then only 8% afterwards. So, that's really a big difference. It shows that, yes, this can be a game changer in this space trying to identify groups that respond and groups that don't. And the other thing you'll see, and I'm going to point to it here on this slide, is that there are only four ctDNA-positive patients at the end of the study, out of the 53 that started. Well, what happened to them? All four of these had to go on to TME. Out of the 10 that had the actual resection, four of them were those with the ctDNA-positive. So, yeah, definitely, it's a small trial. We're aware of that. The authors conclude detectable ctDNA after a neoadjuvant chemotherapy is associated with poor response and persistent cancer. I think that's a fair conclusion, even though a small population.

Dr Namrata (Neena) Vijayvergia: Thank you. I can take this. 

Dr Yelena Janjigian: So, before you—I know we've talked about MSI case [audio cuts out]. You know, yes, recognizing that less and less MSI tumors present, you know, in adjuvant setting. But we do see some especially if the testing wasn't done or the patient had primary surgery. So, in our—for our violinist, what would you have recommended?

Dr Jason Zell: So, first of all, in New York, I know that's a good violinist, right? The New York violinist is a good one. In California, this would be a surfer who doesn't want to have neuropathy. It's the same population. Yeah, I would say this is my least favorite consult because I wish they had upfront immunotherapy prior to surgery. But, it still happens, you're right. What do you do? I'm old school, right? So, I still remember the article from Ribbick in New England Journal 2003 that showed that you can harm stage II patients by giving chemo. And I don't do it. And I don't think there's a role for ctDNA in that population. We might agree on this one. Your thoughts?

Dr Namrata (Neena) Vijayvergia: And I can—we can talk because we're doing the adjuvant setting next. And after that, we can answer it better after we both have discussed, too. So, thank you, Dr. Zell. And here, I am—I'm talking about the con side as whether we should do it in the neoadjuvant setting or not. And the question is—I'm not discussing the potential. I'm talking about is it ready now? So, that's—please bear that that in mind. I really believe that at some day—and I really hope that we would be able to use it in all our patients and help guide their treatment, but I think, at this time, I don't think we are ready for the prime time. So, let's start. So, here is the same picture that I'm talking to you, but we are looking at the red area where we know that there is cancer. So, the neoadjuvant setting is there is cancer that we can see and we're trying to treat it. The problem with using ctDNA, the biggest problem in this setting is—you know, the goal of ctDNA is to identify micrometastatic disease, right? We all want to avoid that cells that will go and cause cancer other places. The problem in the neoadjuvant setting is majority of the ctDNA you see in the blood is coming from the primary tumor. So, while we're going to remove it anyways, right, this neoadjuvant setting, we're going to remove that patient's tumor anyways. So, how much does that ctDNA positivity actually affect how this patient's going to do is hard to assess. The other aspect is, you know, the cost of it, adding it to every patient, the technical limits of our testing. And we're going to talk about some of this as I move on. And also, the question of overtreatment, right? Again, the same issue that if I'm seeing it, if I'm going to remove that tumor, am I just getting stressed out about a positive ctDNA in the neoadjuvant setting? So, I always—you know, my son keeps asking me in the car when we are going on a car ride, are we there yet? Are we there yet? And because we're really excited to be where we want to. But I don't think we are there yet. And the reason being, we just talked about it. I know Dr. Zell presented a really nice study about the DYNAMIC-Rectal trial and how adjuvant therapy was decided. But it was not a neoadjuvant study. That was, again, sort of an adjuvant study because these patients had had their tumor removed. And then, we were deciding whether to give them adjuvant therapy or not. So, it probably doesn't hold true in this particular setting. Now, to look at ctDNA in the nonoperative setting, which, you know, we have just heard—actually, is the most important potential benefit of this is, like, where I'm trying to preserve the rectum, can I use ctDNA as a surrogate? We all know that patients who have clinical complete response, not all of them actually develop a pathological response. And many of them recur and have complications and have disease-related progression, et cetera. So, can we use ctDNA in that setting? Yes, I would love to. You know, I would really want to do that. But there's not a whole lot of studies. There was—I found one study, which was about 90 patients it's in. And what they did was every—all of these patients got total neoadjuvant therapy, where they get chemotherapy and chemoradiotherapy before surgery. And then, we decide, if they have a, you know, complete response, we don't take them to surgery. And what they found in those 90 patients, there were 20 patients who were ctDNA positive, okay? So, after completing their neoadjuvant therapy, 20 patients had ctDNA positivity still. Out of them, four actually had a complete response, okay? And out of those four, only one developed recurrence, okay? So, basically, they had positive ctDNA. Sixteen also had no clinical response. So, we would already—we didn't need ctDNA for them. We would take them to surgery. And of the three patients who were complete clinical response but had a positive ctDNA, there was one patient who actually recurred. So, the other three, even though they were ctDNA positive, did not recur. Now, again, the follow-up was only 20 months in this study. We need more follow-up. But, again, did we just—like, would we treat all four of them because they were positive ctDNA and take them to surgery or do something extra for them? So, is there a concern for overtreatment? But maybe not, right? Because what we need to do is we need to use ctDNA in conjunction with other things, right? Is there MRI findings that suggest one way or the other? Can we use it as a combination, not as one thing? I think there is potential, but we have to learn how to do this. So, I don't think it's ready for prime time in this setting alone. The NEO trial, we just talked about, right? We just studied. This was a study in early stage rectal cancer. So, the difference between the other study and this study is the DYNAMIC study was, you know, stage II and III. This is predominantly stage I and early stage II patients of rectal cancer. Again, nonoperative treatment. You get chemotherapy and a lot of them convert to negative. Everything is good. But here comes the question of the technical limitations of the test. If you look here, 48.8% of the patients never had ctDNA positivity from the get-go. So, those 40%, 50% of the patients who never had ctDNA, that test is useless, right? They were negative at baseline. What happens afterwards if they're negative positive? Does that really mean that there was never cancer that was spread or were we doing that test for no reason? We need to get better at doing, in this early state setting, trying to improve the sensitivity of the test so that we can detect what these 48% or 50% of the patients are going to do. If you look at this last curve, the CT. You know, this is basically the percentage of patients who had recurrence. You can see the prechemotherapy ctDNA status had no connection, which if you were positive, because it probably was coming from the primary tumor, it didn't help much. And then, postchemotherapy, there were still 10% of the patients who had recurrence where no ctDNA was detected. So, if we just use one, you know, I don't know if you can just trust ctDNA for that. So, again, there is a role, there may be, but it's probably in conjunction with a lot of other factors, MRI features that you see. You know, we do a lot of tumor board discussions about it. And I think all of these together one day will help us, and it will become a part of their group. I really hope so, but I don't think we're there yet. And then, this is a good picture of, like, you know, our enthusiasm is ahead of the evidence, right? Basically, we had a plan. We think it's going to work. We thought we're going to get to success right away. But it's never a straightforward path. There's ups and downs, ups and downs. But I really hope, ultimately, we get to the success.