Best Practices Surrounding IO Treatment for mCRC

Tanios S. Bekaii-Saab, MD: Raghav, where does that put us with IO [intraosseous infusion], specifically pembrolizumab as a single-agent PD-1 [programmed cell death protein 1] inhibitor beyond first-line therapy? Is there still a role? What Cathy presented to us, with KEYNOTE-177, is that now everyone should get it in the first line. Are there any instances where you think we can and should still use it beyond the first line?

Kanwal Raghav, MD, MBBS: Hopefully not, but if for some reason you have missed out on MSI [microsatellite instability] in the first-line setting, then yes there is a clearer role and the efficacy seems compatible to when you use it in first-line. The response rates are around that 40% to 50% margin. You can use pembrolizumab. There is also data to use nivolumab [nivo] single-agent. There is also data with nivo/ipi [ipilimumab]. Coming to that question of nivo/ipi, in MSI-high tumors, it prolongs response rates in patients and does add that small percentage of patients who can respond above and over a single-agent activity. Hopefully, we'll see whether the first-line trial shows that dual-checkpoint inhibition is better than single-checkpoint inhibition.

Tanios S. Bekaii-Saab, MD: We have an approval from the FDA with TMB [tumor mutational burden] high, defined as ten and above, in patients with various cancers and pembrolizumab. When should we consider this option for our patients? At what line of therapy?

Kanwal Raghav, MD, MBBS: From what we've seen, colon is different from most other tumor types from various aspects, but especially considering the TMB status. If you look at large-scale data analyses, you find that MSI high status and high TMB are closely linked together. In one of the studies of over 6000 patients, where they had at least a standard way of assessing TMB, there have been multiple studies. One of the limitations of those studies is the NGS [next-generation sequencing] is not uniform or optimized. The other is that the MSI has been tested either by IFC [integrated fluidic circuits] or by PCR [polymerase chain reaction] and we know that there are fallacies in both of those areas.

When you look at one of the larger studies, 99.9% of all patients with MSI have a high tumor mutation burden defined by greater than 10 mutations per mega-base. On the other hand, only about 2% to 3% of MSI-stable patients will have this kind of a high TMB. When you divide those 2, MSI captures 2 things. Number one, what is your mutational burden? Number two, how is your body actually responding to that mutation burden? You miss out if you just look at TMB. For example, a high TMB in an MSI stable patient does not mean you've lodged a response because they could have other mechanisms by which they're suppressing that.

This was clear from some of the analyses that was done in the TAPUR Study. I know John said that the study met its primary end point of disease control, but if you look at the partial response, there was one patient who responded. That patient had a TMB of 10 and was an MSS-stable [microsatellite] patient, which is not unlike what we have seen for immunotherapy even in standard MSS patients where you can find these response rates of 4% to 5% with IO agents. We don't know why that is specifically. In the MSI-high cohort, an analysis of MSI-high patients showed us that if your TMB is more than 40, then your chances of response is almost as high as 90%.

If you're MSI-high, but your TMB is lower than that, your response rate falls down to 30%. If you're MSI-high, you're getting pembrolizumab in the first line. That's not that useful, although it would be nice to see the future studies try to select a population based on that TMB. Unless you have some good options for them, it seems hard to keep pembrolizumab away, based on an invalidated cutoff. In the MSI-stable population, I would like to see a few more studies of combinations, not single agents, in this MSI-stable but TMB-high population.

Tanios S. Bekaii-Saab, MD: TMB-high seems to be a better predictor than MSI-high, real TMB-high, above 30 or 40?

Kanwal Raghav, MD, MBBS: In MSI-high population.

Tanios S. Bekaii-Saab, MD: Understood. But if you're microsatellite-stable and had a TMB of 30 to 40, you're going to have an MSI-high-like response.

Kanwal Raghav, MD, MBBS: We don’t have that data, because the TAPUR Study wasn't published. I don't know their distribution. Their distribution was anywhere between 6 TMB to 54 TMB. I don't know what the cutoff was, but they only had one response. It wouldn’t matter. Whoever was about 30 didn't still respond.

John H. Strickler, MD: There were three partial responses reported in that TAPUR Study and the TMBs for the three responders were 10, 54, and 17.

Kanwal Raghav, MD, MBBS: Weren't the 2 that responded, one which was MSI-high and the other one that was actually MSI-intermediate? It's talking specifically about the MSI-stable population.

Tanios S. Bekaii-Saab, MD: If you have an MSS with a very high tumor mutational burden, 2 things come to mind. One of them is a POLE and possibly a POLD1 in others. Those patients have TMB super high. I've had one with a TMB of 190 and the tumor melted away. It sounds that it's a relatively good predictor, as long as it's high enough to some point and more work needs to be done.

Transcript Edited for Clarity