Case 4: MSI Testing and Tumor Mutation Burden in mCRC

EXPERT PERSPECTIVE VIRTUAL TUMOR BOARDShubham Pant, MD:Coming back to our discussion, this is a male with metastatic colorectal cancer with microsatellite instability [MSI]. Greg, we’ve discussed this through the different cases but tell me about MSI testing. What are the different ways that MSI can be tested for in colorectal cancer?

Gregory Riedlinger, MD, PhD:Traditionally, there are 2 ways. First there is typically what people would call the MMR [mismatch repair] testing, which typically refers to the immunohistochemistry [IHC] looking for the 4 mismatch repair proteins and seeing if any of those were lost. The alternative would be so-called MSI testing, which is kind of a readout for deficiency in 1 of the mismatch repair proteins. Traditionally, that was done with an NCI [National Cancer Institute]-recommended panel of 5 or 6 genes. If 2 or more of those were altered they were considered MSI-high.

In the alteration, the panel is essentially looking at 3 mononucleotide repeats and 2 dinucleotide repeats. That’s where you see the defect: when you lose the mismatch repair proteins. With some of these large NGS [next-generation sequencing]—based tests, they’re also able to look at similar microsatellite regions and do their own calculations. Actually, I think Foundation Medicine does 90 actual slides to test for MSI-high status.

Shubham Pant, MD:Richard, in your clinical experience what do you send it out for? Do you send it out for IHC? Do you look at the NGS? What do you use to determine the MSI status?

Richard Kim, MD:I think the turnaround time for MMR is the quickest. That is IHC. It takes 2 or 3 days. NGS testing takes longer—anywhere from 2 to 3 weeks. So ideally, if I wanted to get the answer back very quickly to determine if the patient is a candidate for immunotherapy or not, I would do MMR testing, which is very quick and can be done internally.

Shubham Pant, MD:What do you test along with that? Do you test for PD-L1 [programmed death-ligand 1] in these colorectal cancer patients?

Michael Morse, MD:In both of the trials, particularly in CheckMate-142, this has actually been looked at. There’s no impact of PD-L1—positivity on response rate, so there’s no reason to check for it.

Shubham Pant, MD:Which one do you check in your clinical practice? Do you do an MMR test? Do you do the IHC testing? Or do you do more with NGS?

Michael Morse, MD:We typically start with IHC. As Richard pointed out, it’s the quickest way to achieve results.

Richard Kim, MD:Greg, as a pathologist, should you do both? Should you do MMR and MSI? Or, if you have a case that is MMR-deficient, is that it?

Gregory Riedlinger, MD, PhD:I think most places, as you have said, actually do the immunohistochemistry upfront for 2 reasons. One reason is because of its faster turnaround time. It’s also less expensive. The sensitivities are generally reported to be similar—around 92% or 93%—for detecting microsatellite instability. They both have reasons for why they could potentially fail. Most institutions I’ve been at typically do the mismatch repair protein immunohistochemistry upfront. And then, if there’s some equivocal result, they might think about doing MSI testing through a secondary mechanism.

Shubham Pant, MD:Richard, if it shows up, do you then send them for germline testing?

Richard Kim, MD:Yes, if there’s a strong family history and the patient is young. We do have a genetic counselor available to send the patient to. So in those cases, to determine whether or not it’s Lynch syndrome, that would be the next step.

Gregory Riedlinger, MD, PhD:I think most institutions are typically doing this, especially if there is loss of MLH1. The sporadic mismatch repair-deficient colon cancers typically have hypermethylation of theMLH1gene as opposed to a germline mutation. It’s been found that about 75% of those cases have aBRAFV600E mutation. So when MLH1 is lost, a lot of people are actually testing forBRAFV600E. If that’s mutated, you know it’s sporadic. If it’s not mutated, then it could go either way.

Shubham Pant, MD:That’s correct. SoBRAFis done automatically.

Michael Morse, MD:I was going to point out Richard’s comment about sending somebody to a genetic counselor. Even with the information about this case, it is important because you have a family history of breast cancer and pancreatic cancer. That does not usually make you think about Lynch syndrome. Perhaps it makes you think about aBRCAmutation. It’s possible that there’s more than 1 mutation in the family. So I don’t think we can just stop at the fact that this person has Lynch syndrome. There may be other reasons for other family members. They may have other mutations. And so, it’s complex enough, beyond what we, as medical oncologists, can really handle now.

Shubham Pant, MD:Right. I think the important takeaway from this case is that you should really be asking about family history. That’s a very important factor to talk to patients about. That can help us pick up clues as to when we can test them for additional things and send them to genetic testing.

Greg, what about tumor mutation burden? We are reading more and more and more about it, mostly in the lung literature. What is tumor mutation burden? How does it differ from MSI, or from PD-L1 testing?

Gregory Riedlinger, MD, PhD:Obviously, as you said, this is another emerging marker for lung cancer, melanoma, and, potentially, colon cancer. Essentially, it’s just looking at the number of mutations, typically per megabase. A lot of this was initially done on The Cancer Genome Atlas data. The way that testing was done, a lot of companies were trying to then translate that into—as opposed to doing whole exome sequencing—more targeted panels where you may only have a megabase of coverage. But it essentially looks at the number of mutations per megabase. With certain tumors, there’s a correlation with response to immunotherapy. With the MSI, there tends to be high concordance between tumors that are mismatch repair-deficient and tumors that have high tumor mutation burden.

Shubham Pant, MD:Mike, are you using tumor mutation burden in your practice?

Michael Morse, MD:No, because it’s not been validated for colorectal cancer. So far the checkpoint molecules only appear to have activity in microsatellite instability-high patients, not microsatellite stable patients. So until that data are available, we’re not using it.

Transcript edited for clarity.