Ezra Cohen, MD
Pembrolizumab (Keytruda) reduced the risk of death compared with standard of care therapy in patients with relapsed/metastatic head and neck squamous cell carcinoma (R/M HNSCC), but the difference fell just shy of statistical significance.
Pembrolizumab was associated with a 19% improvement in overall survival (OS) in the phase III KEYNOTE-040 trial. Responses were also more frequent and durable for patients assigned to the PD-1 inhibitor, according to Ezra Cohen, MD, Cohen, associate director, Translational Science, University of California-San Diego Moore Cancer Center. He presented the findings at the 2107 ESMO Congress in Madrid.
Cohen said that subsequent immunotherapy in the standard of care arm may have confounded the OS analysis. As 3 times as many patients in the standard of care arm received immune checkpoint inhibitor therapy than in the pembrolizumab arm.
KEYNOTE-040 enrolled 495 patients with SCC of the oral cavity, oropharynx, hypopharynx, or larynx who had progressive disease after a platinum-containing regimen, or recurrence or progressive disease within 6 months of multimodal platinum-based therapy. They were randomly assigned 1:1 to 200 mg pembrolizumab every 3 weeks for 24 months (n = 246) or investigator’s choice of 40 mg/m2
weekly methotrexate, 75 mg/m2
docetaxel every 3 weeks, or 250 mg/m2
weekly cetuximab (n = 234).
Randomization was stratified by Eastern Cooperative Oncology Group performance status (0 vs 1), known p16 status for cancer of the oropharynx, and PD-L1 tumor proportion score (TPS) (≥50% vs <50%). TPS was calculated as the percentage of tumor cells with membranous PD-L1 expression. The results were also analyzed according to a PD-L1 positive score (CPS), which is the number of PD-L1-positive cells divided by the total number of tumor cells X 100.
Treatment continued until confirmed disease progression or intolerable toxicity. The primary endpoint was OS in the intent-to-treat population. The one-sided prespecified efficacy boundary was an alpha of 0.0175, corresponding to a hazard ratio of 0.80.
After a median follow-up of 7.3 months, 22 patients in the pembrolizumab arm continue on pembrolizumab and 2 continue on standard of care therapy. About one fourth of the population in each treatment arm were p16-positive, one fourth in each arm had PD-L1 TPS ≥50%, and close to 80% in each arm had PD-L1 CPS ≥1.
There were 179 deaths in the pembrolizumab arm and 201 in the standard of care arm, for a hazard ratio (HR) of 0.81 (95% CI, 0.66-0.99) favoring pembrolizumab (P
= .0204). Median OS was 8.4 months and 7.1 months, respectively.
Cohen said the differences are further exaggerated in favor of pembrolizumab in the biomarker-specific cohorts. In the population with PD-L1 CPS ≥1, the 1-year survival was 40.1% in the pembrolizumab arm versus 26.7% in the control arm (HR, 0.75; P
= .0078). Median PFS was 8.7 months and 7.1 months, respectively.
In the population with PD-L1 TPS ≥50%, 1-year survival was again significantly improved from 25.8% in the control arm to 46.6% in the pembrolizumab arm (HR, 0.54; P
= .0017). Median OS was 11.6 months in patients pembrolizumab and 7.9 months for those assigned to standard of care.
The objective response rate (ORR) was also higher in pembrolizumab-treated patients. ORR was 14.6% with pembrolizumab compared with 10.1% with standard of care, respectively were (P
= .0610) in the ITT analysis. Similarly, ORR favored pembrolizumab in both the PD-L1 CPS ≥1 subgroup (17.3% vs 9.9%; P
= .0171), and the PD-L1 TPS ≥50% subgroup (26.6% vs 9.2; P
Median progression-free survival (PFS) was not significantly different between the 2 treatment groups in the ITT analysis and in the subgroup with PD-L1 CPS ≥1, but in the subset with PD-L1 TPS ≥50%, median PFS was significantly superior at 3.5 months in the pembrolizumab group compared with 2.2 months in the standard of care arm (HR, 0.58; P
Best overall response followed a similar pattern, with 4 complete responses vs. 1 in the pembrolizumab and standard of care arms, respectively.
“What I’ll highlight here, interestingly, is that all the complete responses and almost all the partial responses were seen in patients whose tumors expressed some degree of PD-L1, as it is defined by CPS ≥1,” said Cohen.
The median time to response was twice as long in patients assigned to pembrolizumab (4.5 vs. 2.2 months), and median duration of response was almost 4 times greater (18.4 vs. 5.0 months). Nearly three-quarters of patients (71.5%) in the pembrolizumab group had a response duration ≥6 months vs. only 47.1% of patients in the standard of care group.