In a phase I trial, non–small-cell lung cancer (NSCLC) patients with tumors that expressed PD-L1 had significantly better outcomes with MK-3475 therapy compared with patients with PD-L1–negative tumors.
At 6 months after initiation of treatment, 41% of PD-L1–positive patients had no disease progression compared with 17% of patients whose tumors had low PD-L1 expression. Of the PD-L1–positive NSCLC patients, 72% were alive at the time of analysis compared with 53% of patients with low PD-L1 expression.
Leena Gandhi, MD, PhD, assistant professor of medicine at Harvard Medical School in Boston, presented the results at an AACR press briefing.
Nine patients had high PD-L1 expression and 22 had low-threshold PD-L1 expression. The progression-free survival rate at 6 months was 67% for PD-L1–positive patients compared with 11% for patients with PD-L1 low expression (P = .009). The median progression-free survival has not yet been reached.
“The data from this study suggest that high levels of tumor PD-L1 may be a good biomarker for determining which patients with NSCLC are most likely to benefit from treatment with MK-3475,” Gandhi told Cancer Network.
But the data also suggest that PD-L1 expression is not the whole story, said Gandhi.
“The remarkable durability of responses seen with MK-3475 emphasize that it is critical to define as accurately as possible the right patient population for treatment with this and other similar drugs.”
Levels of PD-L1 of 50% or greater were used as the cutoff for high PD-L1 tumor expression. According to Gandhi, this threshold was defined based on the relationship between expression levels of PD-L1 and outcomes, which may be very different for different cancers. “For NSCLC, patients with tumors that had expression levels between 1% and 49% had overall similar outcomes to those that were completely negative,” Gandhi told Cancer Network. “In this particular case, the 50% cutoff range resulted in a higher positive predictive value of response for expression while retaining a high negative predictive value of response for those with weakly positive or negative expression.”
At the press briefing, Gandhi noted that the response rates of NSCLC and melanoma patients to MK-3475 are different, and that the results should not be compared between different tumor types.
The 50% cutoff point defined by this study will now be used to prospectively stratify patients and evaluate outcomes in a currently ongoing phase II/III randomized study comparing two different doses of MK-3475 to chemotherapy in previously treated NSCLC patients.
Preliminary results from the NSCLC cohort of this trial were reported in October 2013 at the 15th World Conference on Lung Cancer in Sydney. Of the 38 previously treated NSCLC patients given MK-3475 (10 mg/kg every 3 weeks), 24% responded to the therapy, most by week 9 of treatment, and the responses were durable. The median overall survival was 51 weeks.
During the press briefing, Daud acknowledged that there are currently no standards for assaying PD-L1 expression in any tumor type, nor is there an agreed-upon threshold that defines a PD-L1–positive tumor. Daud also noted that it is currently not possible to compare the PD-L1 assays used in different studies and among different tumor types.
1. Gandhi L, Balmanoukian A, Hui R, et al. MK-3475 (anti-PD-1 monoclonal antibody) for non-small cell lung cancer (NSCLC): Antitumor activity and association with tumor PD-L1 expression. American Association for Cancer Research Annual Meeting 2014; April 5–9, 2014; San Diego. Abstr CT105.