Expert Discusses Immunotherapy for MSI-H Colorectal Cancer
Recent data shows promise for the use of anti-PD-1 therapy in the second-line treatment for patients with microsatellite instability-high (MSI-H) metastatic colorectal cancer (mCRC), says Michael J. Overman, M.D.
Overman is the lead author of the phase 2 CheckMate-142 study, which examined single-agent Opdivo (nivolumab) or Opdivo combined with Yervoy (ipilimumab) in patients with recurrent or mCRC, who were either MSI-H or microsatellite stable. At the 2017 Gastrointestinal (GI) Cancers Symposium, Overman presented results from the trial for 74 patients with MSI-H mCRC treated with Opdivo monotherapy.
The overall response rate in these patients was 31.1 percent, the 12-month progression-free survival rate was 48.4 percent, and the 12-month overall survival rate was 73.8 percent. Results are not yet available for the cohort of patients with MSI-H mCRC who are receiving Opdivo plus Yervoy.
In an interview with CURE at the 2017 GI Symposium, Overman, associate professor, The University of Texas MD Anderson Cancer Center, discussed the significance of the ongoing CheckMate-142 study and the next steps in MSI-H mCRC.
How common is MSI-H in mCRC, and what does it mean for those patients?
MSI-H is a subset of CRC in the metastatic setting — about 4 percent or 5 percent of metastatic patients. We've known for a while that this set of patients have been immune-recognized — they tend to have a high mutation load based on the genesis of those cancers as a deficiency in DNA repair. This means they have a high level of mutations, likely conferring they have a high level of neoantigens. When something is mutated, it is foreign, meaning it hasn’t been seen by the patient’s immune system before so it can be immune-recognized. This idea of high mutations/high neoantigen load makes up MSI-H mCRC.
That’s the reason this study was initiated—based on the premise that patients would have a high level of mutations that would be recognized by the immune system. It’s also known that these tumors have high checkpoint expressions, so it appears that these tumors are trying to turn off the immune system and by providing therapy we hope to unleash an intrinsic immune response. I think our study demonstrates that the tumors are doing this.
Can you give an overview of CheckMate-142?
It is a phase 2 study that explores two different strategies. One is combination therapy with Opdivo and Yervoy and the other is monotherapy with Opdivo. It was designed as a two-stage study, so it has a first stage with monotherapy and then based off a high-level response rate, it proceeds to the second stage.
We showed results for the 74 patients who received monotherapy with Opdivo. The second cohort treated with combination therapy was still enrolling at the time of the database lock for this presentation. We presented the six-month updated analysis from the initial presentation at ASCO 2016 on monotherapy.
What did the updated findings show?
The updated findings reflect the further duration of evaluation, which supports some of our early findings where we saw a high response rate of stable disease in patients that we treated. We see approximately two out of three of patients having clinical benefit.
The exciting thing about this data is that patients who have stable disease or responses, appear to be durable. The data shows that over 80 percent of patients who had a response are still on therapy. We see the same thing with stable disease—if a patient has stable disease that lasts over four months, the majority of them are still on therapy. So, this follow-up demonstrated the durability of that benefit in these patients.
The second factor that we discussed that we haven't before is looking at some of the clinical factors of relevance and biomarkers. We showed that PD-L1 expression is seen in about one out of four cases (more than 1 percent), but responses are irrespective of PD-L1 expression. Additionally, we show that responses are irrespective of BRAF mutation or a clinical history of Lynch syndrome.
For MSI-H patients, there are two big categories. There is germline inherited Lynch syndrome and then there is sporadic, or acquired, MSI-H. Having either BRAF or clinical history of Lynch syndrome doesn't appear to make a difference regarding response. From our data, we would say, irrespective of the reason the patient has MSI-H, every patient will benefit the same no matter what. Any MSI-H patient has a chance of this two-thirds rate of clinical benefit from therapy. No population appears to have a higher or lower benefit, per our biomarkers.
Can you discuss the significance of these results in the field of CRC?
It is very exciting for the field because we see a high level of activity, as well as the durability of that activity in this study. It is kind of like what we've seen with immune therapy, but for GI cancers this is one of the first areas that we are seeing robust immune therapy activity. Looking at expectations for a refractory CRC population, the durability and benefit rate in this single-arm study is dramatically higher. Based off that, there have been changes with guidelines that have made the standard of care change for this population.
The current NCCN guidelines state that patients who have had one prior mCRC therapy for MSI-H should be given an anti–PD-1 based therapy, which would be Opdivo based on the data that we are presenting. Keytruda (pembrolizumab) has had similar data—both of those are options for these patients. This is a big change from how we've practiced before because now there is another subset in CRC that we have a unique treatment for.
What are the next steps, given these findings?
The next steps relate to a few things: developing a deeper understanding of what we are seeing and furthering ways to optimize what we've already seen. We could see activity with one anti–PD-1, but could combination therapy provide a greater benefit? We plan to have a data lock upcoming for the combination arm. Clearly, MSI-H is immune-responsive, so it may be possible for other combinations to get a higher bar. When we look at our data, we see this two out of three benefit that appears durable, but could we get an even higher bar? I think that is one area is of great interest.
The second big area, is understanding the reason why some patients do not respond. It is the minority of patients, but understanding them would help us inform some of the next steps. There is a large effort in place to understand the tumors that don't respond based on protein expression, RNA and DNA. If we understand the cohort, then we can figure out which combinations would elicit responses.
Is there anything else in the field of CRC that you are particularly excited about?
There are two areas of significant interest. First, we're talking about 5 percent of mCRC that is MSI-H, so the question for us is what can we do about those 95 percent that aren't. We’ve found 5 percent and we have a new therapy form, but there is still a majority of patients that don't have this biomarker and need a new therapy. There are studies of combinations for this population such as a large phase 3 study that looked at the combination of Tecentriq (atezolizumab) and Cotellic (cobimetinib) within microsatellite stable mCRC, and those results could lead to changes of standard of care for that cohort. Though exciting, that data is likely not coming until the end 2017 or 2018.
The other interesting area is targeted therapy, which is being presented at this meeting. There will be a study presented that looked at BRAF-mutated CRCs, so CRC with a BRAF V-600E mutation, which is a subset of about 8 percent to 10 percent of all CRC patients, who, unfortunately, tend to have a worse outcome in comparison to the other subgroups in this disease. The study presented here will look Zelboraf (vemurafenib)/ Sylvant (siltuximab)/irinotecan, versus the standard Sylvant /irinotecan. The idea is that by adding the BRAF V600E inhibitor Zelboraf into the standard combination, the study will show that there will a benefit with the novel combination.
The good news is that we are making headway and finding subsets of CRC, focusing on them and having success. Moving forward, we just need to find more of those subsets to continue that effort.