Potential Biomarkers Identified for Pembrolizumab in Head and Neck Cancer

PD-L2 and the IFN-γ 6-gene signature may offer options beyond PD-L1 to predict response to pembrolizumab (Keytruda) in patients with recurrent or metastatic head and neck squamous cell carcinoma.

Laura Quan Man Chow, MD

PD-L2 and the IFN-γ 6-gene signature may offer options beyond PD-L1 to predict response to pembrolizumab (Keytruda) in patients with recurrent or metastatic head and neck squamous cell carcinoma (HNSCC), according to an exploratory analysis of the phase Ib KEYNOTE-012 trial presented at the 2016 ASCO Annual Meeting.1

“Biomarkers beyond PD-L1 may offer additional insights to identify patients most likely to benefit from treatment with pembrolizumab,” lead author author Laura Quan Man Chow, MD, said when presenting the data at the ASCO meeting.

“The very exciting biomarker data from this study offer important insights into the underlying mechanisms associated with response to pembrolizumab in head and neck squamous cell carcinoma,” added Chow, who is an associate professor of Medical Oncology at the University of Washington School of Medicine. In the HNSCC cohorts of the nonrandomized KEYNOTE-012 trial,2 192 patients with recurrent/metastatic HNSCC received pembrolizumab. In the first arm, patients with PD-L1—positive HNSCC received the PD-1 inhibitor at 10 mg/kg every 2 weeks (n = 60). In the second group, patients received pembrolizumab at 200 mg every 3 weeks, regardless of PD-L1 status (n = 132). Patients received treatment for a maximum of 24 months or until progressive disease or unacceptable toxicity.

The median age of patients was 60 years (range, 20-84), and the majority were males (83%). Overall, the median number of prior therapies was 2, with 45% having received ≥3 lines of systemic therapy. Fifty-seven percent of patients had received prior platinum therapy and cetuximab. The ECOG performance status was primarily 1 (70%) and most patients had M1 disease (88%). In both cohorts, pretreatment samples were collected for biomarker analyses.

In the overall patient population, the median progression-free survival (PFS) was 2.1 months and the 6-month PFS rate was 24%. The median overall survival (OS) was 8 months (95% CI, 8-11), with a 6-month OS rate of 65%. The overall response rate (ORR) was 18%.

In their analysis, Chow et al assessed the correlation between PD-L1/PD-L2 expression in pretreatment samples with clinical outcomes in patients who received at least 1 dose of pembrolizumab. PD-L1 status was assessed in 188 samples and PD-L2 status was assessed in 172 samples.

Positive PD-L1 and PD-L2 status were defined as expression levels of at least 1%. PD-L1 expression was measured by both tumor proportion score (TPS), which included tumor cells only, and combined positive score (CPS), which included both tumor and inflammatory cells. PD-L2 status was only measured using CPS.

Using TPS scoring, 123 samples were scored as PD-L1—positive; however, with the expanded CPS, 152 patients were scored as PD-L1–positive. When assessing overall response by PD-L1 status in tumor cells alone (TPS), there did not seem to be a correlation between outcomes and PD-L1 status. The ORR was 18% (95% CI, 12-26) in PD-L1+ patients versus 19% (95% CI, 10-30) in PD-L1–negative patients (P = .461).

When using the CPS score, however, the ORR was 21% (95% CI, 15-28) in PD-L1+ patients and 6% (95% CI, 1-19) in PD-L1—negative patients (P = .023).

“CPS improves the ability to predict response compared to tumor cells alone (TPS). CPS may be an important consideration in future studies,” said Chow.

This trend was continued when observing PFS by PD-L1 status. Using the TPS measurement, the median PFS was 63 days (95% CI, 58-98) in the PD-L1+ group and 62 days (95% CI, 59-67) in the PD-L1—negative cohort (P = .378). Using CPS, the median PFS was 64 days (95% CI, 59-98) in the PD-L1+ group and 60 days (95% CI, 51-66) in the PD-L1—negative group (P = .026).

Using TPS, the median OS was 290 days (95% CI, 241-377) versus 246 days (95% CI, 174-626) in PD-L+ versus PD-L1—negative patients, respectively (P = .478). With CPS, the median OS was 303 days (95% CI, 259-385) versus 151 days (95% CI, 84-247), respectively.

“Similarly, the trend is also seen with overall survival. Incorporation of inflammatory cells into the assay for CPS also improved the ability to determine overall survivors,” said Chow.

When assessing PD-L2 status and overall response using the CPS assay, the ORR was 23% (95% CI, 15-31) in PD-L2—positive patients and 10% in PD-L2–negative patients (P = .022).

“We believe that PD-L2 expression on both tumor and inflammatory cells was predicative of response to pembrolizumab,” said Chow.

Chow also noted that there was a significant association (P <.001) between PD-L1 and PD-L2 expression. Further, the ORR among the 108 patients who were both PD-L1— and PD-L2–positive was 23% (95% CI, 16-32).

The researchers also assessed the 6-gene IFN-γ signature (IDO1, CXCL10, CXCL9, HLA-DRA, STAT1, IFNG) with the Nanostring nCounter platform using RNA extracted from FFPE tumor tissue collected prior to treatment. The IFN-γ 6-gene signature score was significantly (P <.001) associated with response in the 150 available samples. The signature was also associated with improved PFS and OS.

In April 2016, the FDA granted granted a priority review designation to pembrolizumab as a treatment for patients with recurrent or metastatic HNSCC following a platinum-based chemotherapy. Under the Prescription Drug User Fee Act, the FDA is scheduled to make an approval decision by August 9, 2016.


  1. Chow L QM, Mehra R, Haddad RI. Biomarkers and response to pembrolizumab (pembro) in recurrent/metastatic head and neck squamous cell carcinoma (R/M HNSCC). J Clin Oncol 34, 2016 (suppl; abstr 6010).
  2. Mehra R, Seiwert TY, Mahipal A, et al. Efficacy and safety of pembrolizumab in recurrent/metastatic head and neck squamous cell carcinoma (R/M HNSCC): Pooled analyses after long-term follow-up in KEYNOTE-012. J Clin Oncol 34, 2016 (suppl; abstr 6012).