It’s very interesting that many clinical trials that we’re evaluating are now using some type of checkpoint inhibitor up front with cytotoxic chemotherapy. If they are positive trials, this could change how patients are treated up front.
Dr. Bauml. There was some really interesting data that were presented at ASCO this year by Matthew Hellmann, MD, which evaluated the predictive nature of PD-L1 vs tumor mutation burden and other biomarkers, including gene expression profiling. In this particular abstract, the PD-L1 and tumor mutation burden really do function as orthogonal biomarkers such that a patient who has high PD-L1 and high tumor mutation burden is the most likely to respond. Patients who are really low for both are unlikely to respond. We really need better biomarkers for immunotherapy, though. PD-L1 has a lot of limitations, namely, it is dynamic, so over time it changes. So I can do a biopsy at one point, then treat the patient and the PD-L1 may change.
More importantly, it’s heterogeneous. There was this great paper by McLaughlin and colleagues in JAMA Oncology (2016) who described a patient who had a small tumor biopsy. They took a micrograph of the tumor and showed that one part of the micrograph was completely floridly PD-L1 positive. At another site of the same biopsy it was completely stone-cold negative, which is humbling when you think about the fact that we stick small needles into tumors and make clinical decisions on the basis of that.
The KEYNOTE-024 study evaluated pembrolizumab vs chemotherapy in high PD-L1 expressers. It’s a very exciting study, but at the end of the day even in this highly select patient population, the response rate to immunotherapy was only about 50%, which is not the sort of biomarker-driven response that we’re used to seeing with our EGFR inhibitors. That’s really what we want to get to. More important even than that is being able to say the negative predictive value. One of the reasons that we’re probably seeing more responses among veterans is that we know that patients who are veterans who have high tobacco exposure have a higher tumor mutation burden. I’m surprised to hear about the immune-related AEs, actually, because one of the things that was reported this year at ASCO was some data that showed that patients who have immune-related AEs are more likely to have a better outcome, which is an interesting biomarker of response.
Dr. Das. I heard that as well, and I found that to be really interesting. The patients that I’ve had on nivolumab for over a year are doing very well. These are stage IV patients who have essentially had complete responses to treatment and have not had any or have had very minor immune-related AEs to date.
Overall, these are a small numbers of patients, but I have been curious to see why that might be the case. Anecdotally, my colleagues and I who treat patients at Stanford have seen significantly higher rates of grades 3 and 4 pneumonitis and other autoimmune toxicities, such as myocarditis and enterocolitis, in those lung cancer patients who are light or never-smokers treated with immune checkpoint inhibitors.
Dr. Bauml. I really feel that PD-L1 as a biomarker has significant limitations. I certainly hope that in at least 2 or 3 years we’re not going to be talking about PD-L1 anymore. I’m hopeful that we’ll be able to use better predictive biomarkers, such as mutational burden and gene expression profiling. In the data in head and neck that was presented this year at ASCO, patients who were low for both gene expression profiling and mutational burden had a very low response (Haddad et al, ASCO 2017).
That’s really what you want to be. You want to be able to say, “Here’s a person who will not benefit from this therapy.” From there you can identify, based upon these biomarkers, the combination that is going to be best for this person. Is it chemoimmunotherapy or combination immunotherapy with CTLA4, or another checkpoint blockade? That is really the way that we’re going to be able to fine-tune this, because the toxicity is substantial for some treatments, like the nivolumab/ipilimumab combination. Using them in a biomarker-blind fashion is just scary to me, honestly.
Managing Adverse Reactions
Dr. Tammaro. The increasing amount of oral chemotherapy has posed a significant challenge. As a clinical oncology pharmacist, it was difficult to grasp the most effective way to follow all these patients and ensure adherence, adverse drug event reporting/significance and adequate follow up. When patients are receiving IV chemotherapy, we know we will see them, we are assured compliance and are able to assess side effects in a timely manner. When we give oral chemotherapy, the tables are turned, where the responsibility is now on the patient. We are now depending on the patient to ensure they are taking the medication correctly and we may not see AEs if the patient misses an appointment or feels as though they are bothering the provider by calling.
