Traveling Through the Lung Cancer Treatment Paradigm: Biomarker Guidance in NSCLC

In this second episode of OncChats: Traveling Through the Lung Cancer Treatment Paradigm, Aaron Franke, MD, of the University of Florida Health, discusses the need for additional circulating biomarkers beyond PD-L1 in non–small cell lung cancer (NSCLC).

"Another topic that’s gaining a lot of momentum, but still has a lot of unanswered questions and an unmet need is the biomarker guidance in non–small cell lung cancer treatment. This is outside of the exciting oncogene explosion where we now have, I believe, 11 actionable, targetable mutations, as well as a smattering, a handful of prognostic mutations that are helping [move the needle] as they gain more and more data, [such as] STK11, KEAP1, KRAS, and [others]. However,...outside of a research initiative...you must remember that PD-L1, despite all its caveats and downfalls, really stands to be the most studied and only validated recommended biomarker to guide immune checkpoint inhibitor clinical decision making in NSCLC.

I won't get into tumor mutational burden in this segment, but really, PD-L1 is the one where guidelines really have the dichotomy and the algorithm kind of grouped together based on this, outside of an oncogene expression. However, while PD-L1 tumor cell expression remains a unique and uncontested immune biomarker and clinical tool, its prognostic significance remains more unclear than most people would think. There are considerable gaps in our knowledge that remain, and a relatively a lack of consensus on how to best harmonize PD-L1 testing and mitigate or at least control for the multitude of clinical pitfalls we see with PD-L1 testing. More importantly, the biologic and clinical implications of such an imperfect immunotherapy biomarker [are not fully understood].

Notwithstanding the practical issues, I do want to emphasize that PD-L1 is not in the same league as other much more informative biomarkers like EGFR and ALK. [The latter] biomarkers, when present or absent, have an inherent value; they reliably tell the clinician to predict a therapeutic benefit in an overwhelming majority of patients or to predict no benefit in the absence of such a mutation or a fusion. Conversely, when you look at PD-L1, this is a continuous variable; it’s more of an enriching biomarker. If you raise the threshold of the PD-L1 cutoff, you get a higher response rate in terms of the patients who are treated; if you scale it down, [with] the defined cutoff on the lower end, you're going to decrease the proportion of patients who are receiving the interventional treatment...

Although these are both valuable objectives to consider, let’s say when designing trials or even thinking about your clinic patients as you’re going to expand the eligibility pool and ostensibly maximize the magnitude of therapeutic benefit, they're diametrically opposed intentions. As such, attempting to achieve a kind of bidirectional satisfaction here can further complicate an already clinically complex situation, and one could argue maybe challenge some bioethical, philosophical principles. We [should] remember that the reliability, the reproducibility, is very limited by the spatial and temporal heterogeneity. Treatment obviously has a dynamic regulation and susceptibility to PD-L1 expression where you are biopsying the primary vs the metastatic site, the nodal vs the primary mass, pre [or] post treatment. We even see huge discrepancies between preoperative biopsies and post–surgical resection specimens when we’re talking about the percentage of PD-L1 high and intermediate when there can be anywhere from a 20% to 45% discrepancy in, let's say, a PD-L1–high expression. This has treatment implications.

We now have neoadjuvant and adjuvant immunotherapies that come into play based on, again, this PD-L1 biomarker. I think the overall goal here is not to [take away from] all the good that PD-L1 has done or take back all of the trials that are obviously rooted and predicated on...splitting up groups into certain treatment subsets, most of them in a post-hoc fashion, but I think it's more to open the door and try to say we need to move forward and try to evolve with newer, smarter biomarkers, circulating biomarkers. [We need to] really try to elevate this field of where we’re going, so we can better enrich for the patient populations benefiting [from certain therapies] and limit the patients who aren’t benefiting from therapies that are not without significant adverse effects in some of these patients.

That’s something we need to keep in mind moving forward. That’s an unmet need that we’re all striving to get involved with, from a research standpoint, because I think it’s very, very translatable to the clinical realm of lung cancer treatment."

Check back next Wednesday for the next episode in this series.