Dr. Mark Awad is a medical oncologist who treats patients with all types of lung cancer at Dana-Farber’s Lowe Center for Thoracic Oncology, where he joined the faculty in 2014 and is now the clinical director. He is also an Assistant Professor of Medicine at Harvard Medical School in Boston.
Dr. Awad’s research focuses on understanding mechanisms of response and resistance to immunotherapy in lung cancer, and he is involved in developing novel therapies for lung cancers with genomic mutations that include MET and ALK.
OBR spoke with Dr. Awad about his thoughts on how Tumor Mutational Burden (TMB)—an emerging genomic biomarker—is being used in precision medicine today, and its nascent role in lung cancer and immunotherapy treatment.
OBR: As we peer into the future of precision medicine as it relates to cancer treatment, please explain what tumor mutational burden (TMB) is and how it is different from other cancer biomarkers.
MA: Cancers can have any number of somatic mutations that arise during their evolution. TMB is an attempt to quantify the number of mutations within a cancer. It is often reported as a fraction or number of nonsynonymous mutations per megabase of sequenced DNA.
Some cancer biomarkers are binary (i.e. they are either present or absent). For example, lung cancers either have an EGFR gene mutation or lack this mutation, and we typically only prescribe EGFR inhibitors in cancer that are positive for an EGFR mutation.
By contrast, TMB is a continuous biomarker. No particular cut point or threshold defines high or low TMB. Some cancers are genomically quiet, or have low TMB. Other tumors have high TMB, and still others are in the middle.
TMB appears to be an important predictor of response to immune checkpoint inhibitors. The immune system better recognizes more heavily mutated cancers (those with a high TMB), since those cancers appear more different to the immune system compared to a normal healthy cell in the body. However, no one agreed-upon TMB assay is available yet for treatment selection.
OBR: TMB is important enough to be part of clinical trial design and FDA submissions by drug companies. Please summarize the evidence supporting TMB and compare it to programmed death-ligand 1 (PD-L1).
MA: Several large, targeted next-generation sequencing panels to assess TMB have been tested in prospective clinical trials. Initial studies of melanoma and non-small cell lung cancer (NSCLC), and now multiple other tumor types, show that more highly mutated tumors tend to do better after treatment with checkpoint inhibitors. TMB may be at play for some patients to determine primary response or resistance to chemotherapy.
A retrospective analysis of NSCLC patients who went onto immunotherapy in the CheckMate 227 trial used TMB as a biomarker, with a cutoff of 10 mutations per megabase. This subset analysis found that patients with higher TMB achieved longer progression-free survival with immunotherapy than with chemotherapy.
Another biomarker checked through immunohistochemistry is the PD-L1 tumor proportion score, which represents the percent of a cancer that expresses PD-L1. For NSCLC patients with PD-L1 greater or equal to 50%, we have two treatment options – pembrolizumab monotherapy or pembrolizumab plus platinum doublet chemotherapy. If PD-L1 is low, pembrolizumab plus platinum doublet chemotherapy is the preferred treatment.
We are still learning about the interplay between TMB and genomic subtypes of lung cancer and PD-L1. Which biomarker is most important for initial treatment selection or subsequent lines of therapy? The ability to devise a composite score with predictive value has become more complex with the growing number of biomarkers.
OBR: PD-L1 is part of FDA labels. What do you do with PD-L1 information when treating NSCLC patients?
MA: PD-L1 is a good predictive, continuous biomarker. The higher the PD-L1 level, the more likely lung cancer responds to immunotherapy. In some trials, we see a benefit with combination immunotherapy and chemotherapy regardless of PD-L1 levels, but that doesn’t mean PD-L1 does not predict benefit.
Similar to TMB, we need to harmonize assessment of PD-L1. Many companies have their own companion antibody and platform for expressing PD-L1, but each company has looked at different cutoffs for PD-L1. The concordance rate between the different antibodies does appear to be quite good, however.
OBR: Tell us about the goal of the cancer researchers organized by the Friends of Cancer Research and what they hope to accomplish regarding TMB.
MA: Since no uniform TMB calculation exists, the goal of this group is to harmonize the results from one testing platform to another. Each tumor type has its own TMB distribution. Ideally, this group will identify platform values of high vs. low TMB within each tumor type.
OBR: How do we apply TMB to lung cancer today, and what do you hope to see in 1 to 5 years?
MA: TMB is not yet an approved biomarker for treatment selection in lung cancer. Some clinical trials have presented early data on how TMB affects efficacy, but no overall survival data have been published. We don’t use TMB for assigning therapies, but I look at TMB to explain why a patient may or may not have responded to immunotherapy.
In the next few years, we will need to pool data across many institutions and partner with industry to come up with composite scores to select the best therapies for patients. That means integrating genomic mutations, TMB, PD-L1 and other factors to explain why patients respond to immunotherapy.
Today, we look for targetable genomic alterations and assign targeted therapy depending on DNA analysis. We hope to have more immunotherapy treatment options in the near future. Biomarkers may predict lack of response to chemotherapy, but we may not have immunotherapy options yet.
OBR: When a newly diagnosed NSCLC patient comes in, is TMB part of the panel with other lung cancer biomarkers?
MA: Yes, we look at TMB on a research basis. Commercial clinical reports show whether a cancer has high or low TMB, even though none of the approved agents in lung cancer or other tumor types indicate that immunotherapy should be prescribed or withheld based on a TMB result. TMB is an emerging biomarker with some potential role for understanding immunotherapy response and resistance, but is not quite ready for use in clinical decision making.
OBR: How would you rate the usefulness of genetic tests for ALK, EGFR, ROS1, and BRAF as well as the biomarkers PD-L1 and TMB?
MA: Genetic tests for ALK, EGFR, ROS1, and BRAF (and RET, MET, NTRK, HER2, etc) are extremely useful. These binary biomarkers clearly predict long-term, durable responses and in some cases survival benefit with targeted therapies. Patients with those alterations should receive targeted therapy first.
PD-L1 is a highly useful biomarker. For PD-L1 high patients, I would offer pembrolizumab monotherapy to spare potential side effects of platinum doublet chemotherapy.
TMB is an evolving biomarker. It has relevance, but not necessarily to the point to base treatment selection on.