Nondriver Adenocarcinoma: Determining Therapy

Stephen Liu, MD: We now have several FDA-approved immunotherapy-based regimens for patients with metastatic non–small cell lung cancer that lacks an actionable driver. We have checkpoint inhibitor monotherapy. We have dual-checkpoint blockade and multiple chemotherapy-immunotherapy strategies. We know that all these are superior to chemotherapy, but they haven’t been compared to one another. Our charge as oncologists is choosing among the available options. I’ve seen some very complex flow charts, but I find it challenging to devise a straightforward algorithm. We need to employ a very individualized patient approach. We have the most options for the PD-L1–positive subset. Although there are some advantages to chemotherapy-immunotherapy in terms of response rate and progression-free survival, particularly in those first 3 months of treatment, not all the patients we see are candidates for chemotherapy. Especially during the coronavirus pandemic, there was some value in avoiding myelosuppressive chemotherapy. That said, if there’s a clinical need, a sense that the disease is advancing quickly or there’s a large burden of disease, these combined chemotherapy-immunotherapy strategies do seem more appropriate. In the PD-L1–negative subset, chemotherapy-immunotherapy remains our standard, but the data for dual-checkpoint blockade are intriguing. Although nivolumab and ipilimumab are not technically approved for the PD-L1–negative subset, that’s where its use may be most valued.

There are many factors to consider in choosing an initial immunotherapy-based strategy. With regard to the tumor, I do factor in PD-L1 expression. At our institution we use the Agilent Technologies, Inc Dako 22C3 clone. Many lab-derived tests are utilized appropriately. Generally, we consider 3 groups: 0, low, and high. There is emerging evidence that the ultrahigh PD-L1 expression carry additional predictive power showing that PD-L1 expression is a continuous variable but is not perfect. It is dynamic. It is inconsistent.

We look at other factors among the tumor histology, particularly when you’re selecting chemotherapy backbone for combinations, as well as overall prognosis. Tumor mutational burden initially had some interest, but it’s been difficult to interpret, and there are a lot of differences among those panels. It does not factor into my decision-making. Increasingly, we also have to look at the host, looking at the patient. Patient characteristics are critical. Performance status [PS], comorbidities, these are going to influence suitability for specific regimens, particularly those that involve chemotherapy. We have to consider patient values, and I anticipate a lot of nonclinical factors as well that will come into the fold—schedule, cost, formulary restrictions—and those may end up tipping the scales when all other factors are equal. Each case really needs to be considered in its own context.

Performance status does factor into my treatment decision. We know patients with a poor performance status will have worse outcomes. In addition, they’re going to tolerate therapy less. When I look at performance status, I try to get a sense of how quickly the performance status is changing and how much of that performance status is related to the cancer. How much of that performance status decline is, in my opinion, reversible? Unfortunately, we don’t live in a world where all our patients are PS 0 or even PS 1, so we have to make some extrapolation. Our comfort with that extrapolation changes but also varies with the specific regimens that we’re choosing. If we feel a regimen is relatively well tolerated and safe, even in the context of a borderline performance status, we can go forward carefully. We all take certain measures to watch those patients a little more closely.

Transcript Edited for Clarity