Biomarker Chase Continues With Checkpoint Agents in Lung Cancer

Roy Herbst, MD

The benefit of PD-1/PD-L1 inhibitors in patients with pretreated, advanced non—small cell lung cancer (NSCLC) has been clearly established. However, achieving true precision medicine with these agents in NSCLC will only come from biomarker-driven frontline regimens delivered to the right patients, according to Roy Herbst, MD, chief of Medical Oncology, Yale Cancer Center and Smilow Cancer Hospital at Yale-New Haven.

“As someone who treats lung cancer and speaks with patients, we have therapies that are working, but it’s going to be up to us to find and use biomarkers to find that group that benefits most so that they can get these therapies early on, and identify the group that isn’t benefiting so that they can get combination therapies and other approaches,” said Herbst in a presentation at the 2015 International Lung Cancer Congress.

PD-L1 is the primary biomarker that has emerged with these checkpoint agents; however, “The biomarker is okay at best,” according to Herbst.

Results with PD-L1 have been mixed in the key trials of PD-1/PD-L1 inhibitors in NSCLC.

In March 2015, the FDA approved the anti—PD-1 agent nivolumab (Opdivo) for patients with advanced squamous NSCLC who have progressed on or after platinum-based chemotherapy. The approval was based on data from the phase III CheckMate-017 trial, in which treatment with nivolumab improved overall survival (OS) by 41% versus docetaxel (9.2 vs 6.0 months; HR, 0.59; 95% CI, 0.44-0.79; P = .00025) in previously treated patients with advanced or metastatic squamous cell NSCLC. In this trial, PD-L1 status was not shown to be a biomarker for treatment efficacy, with similar outcomes regardless of patients’ PD-L1 levels.

Positive PD-L1 status was, however, associated with improved survival in NSCLC patients with nonsquamous histology in the phase III CheckMate-057 trial, which randomized patients to nivolumab or docetaxel after the failure of platinum-based doublet chemotherapy.

Across the overall CheckMate-057 population, the median OS was improved by 27% with nivolumab versus chemotherapy (12.2 versus 9.4 months; HR, 0.73; 96% CI, 0.59-0.89; P = .00155). In PD-L1—positive patients (PD-L1 expression on ≥1% of tumor cells), median OS was improved by 41% (HR, 0.59). The OS benefit continued to rise as PD-L1 levels increased. The reduction in the risk of death was 57% and 60% for patients expressing PD-L1 on ≥5% and ≥10% of their tumor cells, respectively.

PD-L1 status was also linked to efficacy with the PD-1 inhibitor pembrolizumab (Keytruda) in the pivotal phase I KEYNOTE-001 trial. The study examined pembrolizumab in previously treated and treatment-naïve patients with advanced or metastatic NSCLC. Regardless of PD-L1 expression, the agent was shown to be safe and effective, with an ORR of nearly 20% in the overall study population.

However, in a validation cohort, higher PD-L1 expression was clearly linked to improved response. In the validation group, patients were divided into three groups, based on whether they had PD-L1 expression in their tumor cells of ≥50%, 1%-49%, or <1%. ORR in the three groups was 45.2%, 16.5%, and 10.7%, respectively.

Based on the KEYNOTE-001 data, the FDA granted a priority review to pembrolizumab as a treatment for patients with advanced NSCLC across all histologies whose disease has progressed on or after platinum-containing chemotherapy, as well as a targeted agent in EGFR- or ALK-positive patients. A final approval decision is scheduled by October 2, 2015, with the potential that PD-L1 status could be included as an approval criterion.

Herbst listed several reasons why the value of PD-L1 as a biomarker has varied among trials. One of the issues that arises is that PD-L1 expression is heterogeneous and varies with the specfiic antibody used. Other factors that play a role with using PD-L1 as a biomarker include the interval between biopsy and treatment, use in the primary versus metastatic disease setting, staining conditions, and the definition of positive PD-L1 status.

Factors involved in the varying definitions of PD-L1 positivity, according to Herbst, include the cell type expressing PD-L1 (immune cell vs tumor vs both), the location of expression (cell surface vs intracellular vs stromal), intensity (percent of cells “positive”), and distribution (patchy vs diffuse, intratumoral vs peripheral).

To underscore how certain factors can affect the value of PD-L1, Herbst discussed the significance of the location of PD-L1 expression that was demonstrated in the POPLAR trial. The study randomized previously treated patients with squamous and nonsquamous NSCLC to the anti—PD-L1 agent atezolizumab (MPDL3280A) or docetaxel. Importantly, Herbst noted, the biomarker analysis in POPLAR included measuring PD-L1 in not just the tumor cells, but also in the tumor-infiltrating immune cells.

Herbst said that in the overall study population, the median OS was “modestly” improved with atezolizumab at 11.4 months versus 9.5 months with docetaxel (HR, 0.77; P = .11). Only when factoring in PD-L1 status, as measured on both tumor and immune cells, was the population deriving a real benefit from the drug identified.

In patients with the highest level of PD-L1 expression (tumor cells [TC]/immune cells [IC] 3), the median OS with atezolizumab was not yet reached compared with 11.1 months for docetaxel (HR, 0.46). Further, in patients without PD-L1 expression (TC/IC 0), a difference was not observed between the two groups (HR, 1.12).

“We’ve learned that the biomarker doesn’t actually have to be on the tumor,” said Herbst. “If the T cells aren’t there, it doesn’t matter how much PD-L1 you have on the tumor.”

As seen with the POPLAR data, PD-L1 can be effective, but optimizing its use remains a significant challenge. According to Herbst, it is only through refining use of PD-L1, discovering other biomarkers, and identifying patient characteristics linked to improved response that true precision medicine with immunotherapy in NSCLC can be reached.

These mechanisms will lead to “better responses and better activity early on by selecting the right patients or combining the right drugs and thereby increasing the number of patients who make it to long-term survival,” Herbst said.