Incorporating Immunotherapy-Related Biomarkers Into Clinical Practice

June 9, 2017

Dr. Karen Reckamp talks about ways to incorporate biomarkers for the treatment of lung cancer with immunotherapy into clinical practice at the 2017 American Society of Clinical Oncology Annual Meeting, held June 2–6 in Chicago.

As part of our coverage of the American Society of Clinical Oncology annual meeting, held June 2nd to June 6th in Chicago, Illinois, we are discussing the ways to incorporate biomarkers for the treatment of lung cancer with immunotherapy into clinical practice with Karen L. Reckamp, MD, a medical oncologist who specializes in the treatment of lung cancer and is the co-director of the lung cancer and thoracic oncology program at the City of Hope Comprehensive Cancer Center in California. Dr. Reckamp discussed this issue during an education session dissecting the current and potential roles of biomarkers for immunotherapy in cancer.

-Interviewed by Anna Azvolinsky, PhD

OncoTherapy Network: First, could you tell us if there are currently biomarkers that have been validated for use with certain immunotherapy agents and how these are used in clinical practice? And then maybe also what the limitations are?

Dr. Reckamp: That is a big question and I would say that first off, the role of biomarkers and testing for immunotherapy is a very dynamic area at this time, especially in lung cancer. At this time, the best validated biomarker for immunotherapy is PD-L1 which is found on tumor cells and immune cells. The assays that have been validated are different for each drug that has been developed.

The first biomarker that showed validation and correlation with clinical outcomes was the 22C3 assay for pembrolizumab. In the phase I study with that drug, the researchers looked at PD-L1 expression and found a cut point of about 50% at which most patients had responses and based on that, the researchers performed the assay and assessed at 50% and lower and 50% and above, and found throughout all lung trials with pembrolizumab that there were better outcomes and responses in patients who had PD-L1 expression of 50% or more. So this has become the first biomarker to demonstrate significant benefit. Based on this, the 22C3 assay for PD-L1, when it is over 50%, patients are eligible to receive pembrolizumab as first-line therapy rather than chemotherapy.

The other PD-1 and PD-L1 inhibitors also have their own assays, each of which have shown correlations with response and outcome. None of them have been approved as predicting outcome for any of the PD-1 or PD-L1 immunotherapies. For example, for nivolumab the PD-L1 assay does not specifically correlate with overall survival or progression-free survival-although the trend was towards that higher PD-L1 expression levels did better and this was the same for atezolizumab which is also approved for non–small-cell lung cancer. All patients benefited from the PD-L1 inhibitor and in that company’s assay, they look at tumor cells and immune cells and come up with a score, and those with a higher score had better responses, but those with low scores or zero staining also had responses. So, this goes into our limitations of these assays.

These assays are variable for each agent and although there is some overlap and some concordance they are not interchangeable. Each assay has a different cut point and the ideal cut point is not known, and a high level in general shows more benefit to these drugs, but does not guarantee a benefit from these drugs. A low level generally says there will be less of a benefit, but doesn't exclude a patient from having a benefit from these drugs. These are not perfect markers for choosing a drug for the patient, not at this time. 

OncoTherapy Network: You just went over some of the challenges and caveats of making treatment decisions based on the PD-L1 staining assay of a tumor sample. Do you see potential for these PD-L1 assays to be made better as we learn more about the biology?

Dr. Reckamp: At this time, the PD-L1 assays are the best that we have. There are several groups that are working on looking at the correlations between these assays and in larger and larger cohorts to try to see where these assays will correlate. The ideal state would be to have one assay so that there can be a specific clinical decision. The most important assay that predicts clinical benefit right now is the 22C3 assay that allows a patient to receive pembrolizumab alone versus chemotherapy in the first-line setting. Outside of that, the challenges are great.

Number one are the cut points, number two the variability of the assay, but there are others such as using archival versus fresh tissue. We know that PD-L1 expression can change over time-it can be upregulated, it can be downregulated at various times in the tumor in response to therapies. So having a tumor sample from a different time point may change the results of the test. And tumors are heterogeneous and biopsies only get a small portion of the tumor, and so we may be getting a portion of the tumor that is not representative of the whole tumor. Then we move towards biomarkers in the blood and everyone is interested in having more blood-based biomarkers rather than having to get tissue for each biomarker that needs to be assayed. And in the blood, these tests are still experimental and not ready for clinical use at this time.

OncoTherapy Network: What other biomarkers are being studied that may help make decisions about whether a patient should receive a specific immunotherapy?

Dr. Reckamp: There are many biomarkers that are being evaluated and probably the most prominent is the level of tumor mutational burden, which looks at the non-synonymous tumor mutations, and the higher that level, generally, the more responsive the tumor is to immunotherapy. This has been shown in tumors like melanoma and non–small-cell lung cancer with squamous non–-small-cell lung cancer generally having a higher tumor mutational burden and generally more responsive to inhibitors. But these analyses have been done on small sample sizes and haven’t been validated as a predictive marker at this time; these have been mainly retrospective studies. So that is an area that is being highly investigated. New antigens both in blood and within the tumor are being evaluated and again all of these biomarkers, looking at them in the blood to see if we could develop assays that don’t require additional tumor tissue. I think this is an exciting area and a very dynamic area where what we think is the ideal biomarker may change over the next several years.

OncoTherapy Network: Are there other biomarkers that can help identify patients who are experiencing resistance to their current immunotherapy that are clinic-ready? And if not, where are we in identifying such biomarkers?

Dr. Reckamp: At this time, biomarkers that help identify resistance are not ready for the clinic. We are at the point where, again, we are doing retrospective studies about mutations that can occur that can cause changes in the immune milieu that decrease the immune activity against a tumor. But the immune system is complex and there are multiple mechanisms and areas where the immune system can be modulated to turn off the antitumor effect of immunotherapy by the cancer. At this time, what is moving forward in the clinic are combinations of therapies especially for patients who are PD-L1-negative who have less chance of benefit, to try to activate the immune system in these patients with combinations and overcome the intrinsic resistance that is there. 

OncoTherapy Network: Thank you so much for joining us today, Dr. Reckamp.

Dr. Reckamp: Thank you.