Molecular Testing Needed to Inform CRC Treatment Decisions

Michael J. Overman, MD

Physicians must perform molecular testing in each patient with metastatic colorectal cancer (mCRC) as early as possible in order to identify subsets within the disease, as this information can have a prognostic impact in terms of treatment, said Michael Overman, MD.

“Microsatellite instability (MSI) status has to be obtained as well as BRAF, KRAS, NRAS, RAS, and now, HER2, is one that I would say is key,” said Overman. “We have to obtain [information regarding] all of those alterations in our patients and, if found, all have treatment-relevant implications.”

If a patient’s tumor is MSI-high (MSI-H), immunotherapy might be an effective treatment approach, whereas the same choice would not be made with a patient who has microsatellite stable (MSS) disease.

“It is frustrating when we look at tumor types like melanoma or lung cancer where you have this great PD-1—based response, whereas in MSS CRC we don't have those agents right now,” said Overman. “There is no standard immunotherapy option. The only differential is if [a patient has] MSI-high [disease]—which should be tested [for] in all patients with CRC—then there is an immunotherapy option.”

Detection of RAS alterations can also be used to inform treatment decisions, especially if patients end up developing resistance to agents that had shown prior success. For example, an anti-EGFR rechallenge approach may be beneficial for RAS wild-type patients with CRC who have developed resistance to therapy, according to data from the phase II E-Rechallenge trial presented at the 2019 Gastrointestinal (GI) Cancers Symposium.

The study examined the predictability of efficacy for cetuximab (Erbitux) rechallenge by liquid biopsy and showed that the rechallenge approach resulted in activity in the salvage setting, in patients who previously responded to cetuximab. With the use of liquid biopsy, physicians can identify who may benefit from this approach, which could potentially become a standard of care in the future, said Overman.

OncLive: What are the clinical implications of the CCTG CO.26 findings?

In an interview with OncLive^® during the 2019 State of the Science Summit™ on Gastrointestinal Cancers, Overman, a professor in the Department of Gastrointestinal Medical Oncology, Division of Cancer Medicine, at The University of Texas MD Anderson Cancer Center, highlighted data from pivotal trials presented at the 2019 Gastrointestinal Cancers Symposium and discussed where research efforts should be focused to improve the treatment of patients with CRC.Overman: This was a randomized clinical trial from Canada that examined the idea of combination immunotherapy in MSS CRC. Specifically, it looked at the combination of durvalumab (Imfinzi) and tremelimumab—a PD-L1—targeted agent and a CTLA-4–targeted agent—and compared that with best supportive care.

Patients with MSI-high CRC account for 4% to 5% of all mCRC cases, and we have tremendous immunotherapy activity there—those patients are a really unique subset. But the rest [of these patients], the 95%, have MSS tumors. We really had a challenge getting immunotherapy to [be effective] in that group.

There have been several trials looking at more single-agent checkpoint inhibitors and they haven't worked, and so, this is one of the first trials where we're looking at a big data set [with] combination therapy. The trial was a positive clinical trial. The endpoint was OS difference between the combination of durvalumab and tremelimumab versus best supportive care and that marker was met.

There are some caveats to the trial because the trial was designed to be a little bit more exploratory in nature, and there were some [small] differences between the groups in how the study was performed. The most unique aspect that really is the key that needs further confirmation is that there was a difference in OS, but we didn't see any difference in regard to PFS or in terms of response rate. Therefore, there's some discrepancy in regard to the efficacy endpoints we classically look at. Immunotherapy does have some different mechanisms of action, and you can see this component of maybe not as much antitumor obvious effect but still a benefit in OS.

The IMblaze370 and MODUL trials did not see the same kind of success with immunotherapy in MSS patients with CRC. Could you expand on these studies?

It is a very interesting hint that there may be a role for combination therapy in MSS CRC, but it's a study that needs more confirmation. I don't think that we're ready to take this into clinic any time soon in regard to standard of care, but it’s one of the first few hints that we've seen of potential activity in the MSS CRC population, which has been a tough population for us [to treat].IMblaze370 and MODUL were large randomized clinical trials that were in patients with MSS CRC and looked at the idea of single-agent PD-1¬—/PD-L1–based approaches. The IMblaze370 study looked at the use of atezolizumab (Tecentriq), a PD-L1 agent, versus atezolizumab plus cobimetinib (Cotellic), a MEK inhibitor—there was a lot of theory that this might help improve the immunotherapy response—versus standard regorafenib (Stivarga). All those arms were basically the same; there was no real benefit activity seen with monotherapy or with this combination of [atezolizumab and] cobimetinib. It was a negative study showing that monotherapy doesn't appear to have a role in CRC.

The MODUL study was similar in that vein where [investigators] looked at the combination of atezolizumab, plus an anti-VEGF agent, bevacizumab (Avastin), in a little bit of a different setting. The question was, “Would that combination have activity versus a fluorouracil/bevacizumab comparison?” Both were negative studies. In CRC, PD-L1 therapy—and we have a lot of PD-L1/PD-1 therapies are on the market that are out there and are used across many different tumor types—it's been pretty clear that in MSS [tumors], they don't appear to work.

What is the role of immuno-oncology in this space?

Is there a combination that we can get to work with immunotherapy? That's where the exploration is, and as we've talked earlier about the potential with the CCTG CO.26 study from Canada, which has given some hints that maybe a combination of a PD-L1 and a CTLA-4 inhibitor may have some activity, although again, that needs further confirmation.Immuno-oncology in general is kind of exploding across the cancer space. What potentially is the right combination to use? There are a lot of different resistant mechanisms at play and it does appear that with CRC, we have a lot of stromal microenvironment factors that are potentially hindering our success. We do have a pretty macrophage- and myeloid-infiltrated tumor environment. We do have T cells, so there's always been a lot of theories that we would have success with immunotherapy, but there appears to be something that's hindering those T cells from being active.

That’s why many of the efforts are starting to move beyond PD-1 or CTLA-4 [inhibitors]—which are really T-cell—targeted therapies—to considering targeting some of the other immune subsets within that tumor microenvironment. Could you combine therapies that target myeloid cells or macrophages with T-cell targeting? Would that be a more successful effort? There are a lot of combination studies out there.

The challenge is, there is a lot of complexity to the immune system and there is a lot of different reasons to look at different combinations. We’re still having a lot of challenges with getting our initial model systems in the laboratory to reflect patients. In our lab model systems, we can see a lot of immunotherapy activity in colorectal model systems, but it doesn't [translate] to the clinic. We’re having trouble modeling it well, so there are a lot of efforts in the clinic based on theories rather than good model system-based evidence.

Could you discuss the concept of an anti-EGFR rechallenge in CRC?

However, we are still in an early exploratory phase. There are several ongoing trials looking at CRC, so my hope is that we're going to have some hints of activity from new agents that will potentially lead us down a path of success. The agent that has the most hints of activity that I've seen recently is a bispecific antibody targeting CD3 and CEA. The idea is that this molecule binds a T cell, binds the tumor, brings those forcefully together, and then generates an immune reaction; that has shown activity in early studies. There have been some toxicity management issues [with this approach], but that agent is moving forward, so we do have some compounds that we have some hints of success. I hope that, over time, we'll see some of those pan out.In regard to interesting spaces that are evolving, there was some work related to the concept of anti-EGFR rechallenge. There was the E-Rechallenge study, which was the Japan clinical trial looking at the idea that anti-EGFR therapy can work in a subset of RAS wild-type patients who then develop resistance. There were some good data demonstrating that the resistance appears to be mediated by subclones of the tumor becoming resistant—we often see that related to developing RAS mutations. We know a RAS mutation predicts that anti-EGFR therapy won't work. Could a RAS wild-type tumor then be forced or selected to have mutations develop? We see that in patients, and then those mutations actually decline over time.

With our anti-EGFR therapy, we're selecting out these resistant groups of the cancer and then when you take the anti-EGFR therapy away [the resistance] tends to go away. There is a lot of an interest in whether you could actually retreat patients. They develop resistance, but maybe that [resistance] goes away over time.

The E-Rechallenge study [showed that] you can do a rechallenge by selecting patients who have prior benefit with anti-EGFR therapy, given them time off therapy, and then rechallenge them; patients can have response and benefit. Those response rates were around the 20% range—so reasonable response rates.

The more interesting thing is, at the time that you rechallenge them, if you do circulating tumor DNA (ctDNA) and [RAS mutations are detected], then the rechallenge doesn't really work, but if you have patients who have RAS wild-type status, the rechallenge appears to work well and the response rate is even higher than 20%.

What recent data were presented from the ReDOS trial?

This E-Rechallenge study is confirming that, and it puts this idea of an anti-EGFR rechallenge as a real therapeutic option for us. The question is, “How do we do that? When do you draw the ctDNA? What are the best criteria to select patients?” There are several other studies in this space that are active and ongoing, so over the next year or so, we are going to have more studies. The space is going to get better defined with a potentially standard approach. However, I don't think it's quite there yet regarding how to clearly define the exact population who will likely have the most benefit and how to do the approach.The ReDOS trial looked at the optimization of regorafenib regarding toxicity management to some degree, as regorafenib does have some significant toxicities related to hand-foot syndrome and fatigue. The idea was, can we try to get to the optimal dose in the most efficient way possible? Classically, we start at a very high dose and then often come down. A lot of the clinical feeling is that we often end up with a dose that's lower than the starting dose.

In the ReDOS trial, investigators started at a lower dose and ramped up very quickly. Over the first few weeks, they ramped up to full dose—they started at 80 mg, went to 120 mg, then 160 mg, and it was compared with a more standard approach. The question was, “Would the toxicity tolerance be better?” The primary endpoint was evaluation of the number of patients who [continued] after the first 2 cycles and started the third cycle of therapy. They found that more patients started the third cycle of therapy, so patients were able to stay on therapy for longer.

It's a smaller randomized study, so there weren't robust endpoints regarding OS and PFS. However, the trial definitely looked at toxicity, and it did appear to be favorable to the idea of ramping up more than starting high and coming down. You saw that with later lines of therapy.

Where should future research be focused?

I do think that the ReDOS is a good new approach to how we dose regorafenib and a lot of us in the community feel that the FDA-approved dose, in a lot of situations, is a tough dose to tolerate and potentially may not be the optimal dose for a subset of patients. Clearly, there are patients who can tolerate it, but the idea of ramping up makes a lot of good sense to maximize quality of life and efficacy. ReDOS gives us pretty good data that there is no real hint of any decrement to efficacy with that ramp-up approach.The model system for immune prediction is a huge area that should be worked on regarding development. Again, we're seeing an ongoing success we've had lately in CRC, which has been mutation- or subset-based, so one of the areas we'll see is more refinement in that space.

For BRAF-mutated disease, we have vemurafenib (Zelboraf), irinotecan, and cetuximab, but there are other great combinations out there; we are going to see some other great combinations coming down the pike. We found a combination that works, but can we refine it? HER2-targeting combination therapy appears to be successful and, in time, we're going to see an FDA approval for combination therapy for HER2-targeted therapy in CRC. In some sense, one of the [research] areas would be doubling down on the successes we've had and trying to make those better.

It's the same thing in the MSI-H space. We have amazing outcomes with immunotherapy for MSI-H CRC; we have this dramatic number of patients who get durable benefit, potentially even curing patients with metastatic disease. And yet, there are still patients who don't respond. Can we do better and look at combinations that double down on the success seen in that space? Those are active areas where we can go further to make bigger headway. The MSS immunotherapy challenge is what we need to solve.

What is your take-home message?

I would vote for us to take a step back. There are many different exciting drugs out there, almost too many in some sense for the various combinations [available], so how do we move forward through that and do that efficiently?In CRC, the take-home message is [to focus on] molecular testing and identifying subsets within [the disease]. That's where we make tremendous headway and tremendous advances. We have to test our patients with mCRC for a number of different molecular alterations right out the gate, because it has prognostic impact. It changes your thoughts on treatment options, and we have to do that with every patient with mCRC.

Ohhara Y, Shinozaki E, Osawa H, et al. Liquid biopsy for optimizing the rechallenge of cetuximab in metastatic colorectal cancer: additional study of E-Rechallenge trial. J Clin Oncol. 2019;37 (suppl 4; abstr 585). doi: 10.1200/JCO.2019.37.4_suppl.585.