Endoscopy and ctDNA Assays May Help Facilitate, Optimize Nonoperative Management of Select GI Cancers

Michael J. Overman, MD, MD Anderson Cancer Center

Cutting-edge research in gastrointestinal (GI) cancer management is shedding light on more reliable tools for guiding nonoperative approaches, particularly in localized mismatch repair–deficient (dMMR) rectal cancer, according to Michael J. Overman, MD, who adds that ongoing studies like the phase 2 Phoenix trial (NCT05961709) at the University of Texas MD Anderson Cancer Center are set to explore the role of endoscopy and circulating tumor DNA (ctDNA) assays in evaluating tumor responses with immunotherapy regimens.

“From our experience, we have found that endoscopy is the most consistent [tool] to assess tumor type and tumor responses to therapy, ” said Overman, who is the associate vice president of research in the MD Anderson Cancer Network, a professor in the Department of Gastrointestinal Medical Oncology of the Division of Cancer Medicine, and chair of the Division of Executive Committee of the Medical Staff at MD Anderson, in Houston, Texas. “The key is going to be laying out the structure prospectively and showing that [endoscopy] is able to identify patients who really do have a complete tumor response and can [receive] nonoperative management.”

During an interview with OncLive®, Overman expanded on key topics from his presentation on nonoperative strategies in GI cancer at the City of Hope Annual Advances and Innovations in Endoscopic Oncology and Multidisciplinary Gastrointestinal Cancer Care meeting. He explained how endoscopy and/or ctDNA assays could address challenges with assessing responses to immunotherapy in select GI cancers, shared insights from prior and ongoing studies conducted at MD Anderson regarding the predictive value of endoscopy and blood-based ctDNA biomarkers in guiding therapy decisions, and detailed the importance of understanding and modulating the microbiome to optimize immunotherapy responses in cancer care.

OncLive: What are some potential tools for improving the utilization of nonoperative approaches in patients with GI cancers?

Overman: Classically, when you do therapy for cancer, we tend to use imaging. We perform computed tomography [CT] scans, MRIs, and PET CTs. The issue that comes up with nonoperative management is we’re often talking about luminal tumors, such as colon or small bowel or rectal [cancers], which are a little bit harder to image. Thus, it’s a lot harder in colon cancer, for example, to really be certain of what’s really happening with regard to response. The question is, are there other ways to better assess response in a localized colon [cancer] setting? We have 2 other options: endoscopy and ctDNA. We do use endoscopy for rectal cancer, and it’s actually a pretty good kind of approach there. So [the next question becomes]: Could we use endoscopy for other luminal cancers like small bowel cancer, gastric cancer, or colon cancer, and how well does that predict the response of immunotherapy?

We’ve done some work here, at MD Anderson, and published some retrospective studies looking at this, and we see a real discrepancy between endoscopic response and CT scan response. We see that the [sum of lesions on a] CT scan generally [cannot be] 0, so you don’t have complete responses [CRs]. You often have scarring left behind or some thickening; there’s something still on the scan. Endoscopy seems to be a lot better at assessing whether the tumor is there or not there. That’s kind of an area where we’re seeing some encouraging data.

The other [question] is: Could some of this new technology that we’re using more in the metastatic setting be used in the adjuvant setting to guide therapy, such as blood-based ctDNA biomarkers? We’ve used that historically, for metastatic disease settings, just to look at various mutations. We’re starting to see data in the adjuvant setting to assess whether cancer is left behind in patients with resected cancer. Could we use that in the neoadjuvant space to assess the disease response, and does negative ctDNA tell us that we’re at a point where we’ve treated enough? Could that be used as the marker to guide when we give enough therapy and if we’re done with therapy?

What have current data shown about the utility of these assessments in clinical practice?

We have some pilot data showing that [ctDNA levels] seem to correlate well with responses. With neoadjuvant therapy, we [want to know whether] it is working early, because these are patients [for whom the] alternative to immunotherapy will be surgery. We don’t want to treat [a patient] for a long period of time and have that [approach] not work, so we want to have an early assessment [to determine] that the therapy is working. Can we rely on our CT scans early? Can we get a blood-based biomarker early? Can we get a scope early? What’s the best way to make sure we’re on the right path?

Also, how do we assess that we have completely eradicated the tumor and that we’re okay to do a nonoperative approach? Could we use both ctDNA and endoscopy for both of those thresholds? A CT scan can be utilized as well, but our experience has [shown] that CT scans don’t [accurately show a] CR; there’s generally something left on a CT scan. That does create some uncertainty. [If] I still see something on the scan and I don’t know what it is, how do I best assure myself that it is just scar [tissue] remaining and there’s not actually any cancer in that area? That’s where the idea of using these novel approaches becomes clinically relevant.

[It’s important to discuss] how to use [these tools], what the frequency is, and what the best assessment [might be]. This is an area of great interest. If a patient [would] rather keep [their] organ, that’s a reasonable request, [but] how do we [ensure] that [request] is fulfilled? We need to figure out what the recommendations are to do that. It’s going to take some prospective clinical trials to answer this in different areas.

Are any ongoing studies at MD Anderson seeking to confirm the predictive value of endoscopy and ctDNA assays when considering nonoperative management in this population?

There is a study we’re starting up at MD Anderson [called the] Phoenix trial of cemiplimab-rwlc [Libtayo] in localized dMMR colon cancer. Colon cancer is a space where we wouldn’t necessarily use [nonoperative management] as a guideline-based standard. We have serial endoscopies as our critical threshold for driving decision-making.

We use endoscopy as an early assessment of response, and then we use it as a final assessment of [whether] there is a tumor left behind. We weight endoscopy over CT scan findings, so the CT scan needs to show no progression, but as long as endoscopy shows a response, then the therapy is working. As long as endoscopy at the end [of treatment] shows a CR and CT doesn’t show progression, then we’re saying that’s a CR in the tumor. From our experience in small numbers [of patients] at MD Anderson, [endoscopy] appears to be the best assessment that could be done across the board and at many centers.

We’re going to open our study to multiple other centers to reflect standard clinical practice. They do get CT scans as an alternative approach, but there are some limitations there. Certainly, ctDNA is interesting. We did look at that in our initial study. There are 2 main distinctions in ctDNA [assays]: one is tumor informed, and the other is panel based. “Tumor-informed” means that we test tumor tissue and then look for those mutations in the tumor and in the blood. [For this approach], we need a lot of tumor tissue to test, and it takes time to do that. In a neoadjuvant setting, we don’t have a resection specimen, so we often don’t have a lot of tumor tissue. Then, if you want to assess response quickly, you need to have to [results] back quickly. [Because of these limitations,] we did panel-based testing, which worked well in the patients who had positive ctDNA. [However, approximately] 50% of patients didn’t have any ctDNA that we detected. [Accordingly,] that approach works well, but it doesn’t capture the totality of [patients with] localized [disease], so it wasn’t a useful marker to follow. Some tumors are more detectable with ctDNA than others. [Maybe that’s because] the tumor is on the smaller side, is more luminal, and doesn’t shed as much.

[Based on this] experience, [we feel that] endoscopy is the most consistent assessment that will be most responsive to assessing tumor status in this nonoperative management space for luminal colon cancer. That’s how our study is structured.

What future research directions in this space are you most interested in?

My interest is in immunotherapy, [creating] good assessments of response to immunotherapy, and [figuring out] how we can optimize these responses. One area where there’s a lot of interest from us is attempts to modulate immunotherapy responses with endoscopic techniques. We’ve seen some success here with microbiome modulation. This is [based on the] understanding that the microbiome interacts with the immune system. There are good microbiomes and bad microbiomes, [which can] interplay with how well immunotherapy works. We see this in our preclinical mouse models, and it is best demonstrated in [patients with] melanoma.

The studies that demonstrate this are amazing. [Essentially], you take the microbiome from a patient who has melanoma and responds to immunotherapy and infuse it into the colon of a patient who has melanoma and does not respond to [the same] immunotherapy. They do a colonoscopy and put that good microbiome in there, and then a fraction of those patients goes on to respond. It’s the same drug, and all we did was change the microbiome. We need to be capturing the microbiome more to [better] understand [and assess] it.

[Regarding] the modulation of [the microbiome] with endoscopic procedures, one of the big questions is what part of the microbiome is critical to [remove from responders], and that’s where a lot of research is being focused. That’s a space that looks interesting in the GI field. [Notably], this is also a proven technique for other infectious diseases. The oncology space is taking it and leveraging it to show that it does have potential in cancer care, which is interesting and is something that I’m sure we’ll be discussing [further in the future].

Editor’s note: This interview was conducted prior to the City of Hope Annual Advances and Innovations in Endoscopic Oncology and Multidisciplinary Gastrointestinal Cancer Care Meeting.