MUNICH-Patients with recurrent or metastatic squamous head and neck cancer lived significantly longer when treated with the anti-PD-1 immune checkpoint inhibitor pembrolizumab-either as single agent or in combination with chemotherapy-as their first systemic therapy. These findings were from the phase III KEYNOTE-048 study reported at the 2018 ESMO Annual Congress.
The longest extensions of life were in patients who expressed the PD-L1 biomarker for the PD-1 molecule and had monotherapy with pembrolizumab-without any chemotherapy. In these patients, all grades of toxicity were significantly lower than in those who had regimens including chemotherapy.
"We can anticipate seeing pembrolizumab enter standard of care for the first-line management of recurrent and metastatic head and neck cancer," said lead author, Barbara Burtness, MD, Professor of Medicine and Co-Director of the Developmental Therapeutics Research Program at Yale Cancer Center, in New Haven, Conn. "As of today, we can recommend pembrolizumab alone for patients who are PD-L1 expressing. And we can recommend pembrolizumab plus chemotherapy for patients who are not PD-L1 expressing."
Study Details
KEYNOTE-048 randomized 882 patients with no prior systemic therapy (who were not curable by local therapy) into three treatment arms. One group (301 patients) had monotherapy with pembrolizumab. A total of 281 patients in the second treatment arm had pembrolizumab added to chemotherapy consisting of cisplatin or carboplatin combined with 5- fluorouracil (5-FU). And 300 patients in the third (control) arm of the study were treated with the standard-of-care EXTREME regimen using the same platinum-5-FU chemotherapy but with the EGFR inhibitor cetuximab added to it.
Primary endpoints were progression-free survival (PFS) and overall survival (OS). The KEYNOTE-048 data were analyzed for OS and PFS in three different populations: Those with PD-L1 combined positive score (CPS) of at least 1.0 (with 1% of tumor cells staining positive for PD-L1), those with CPS of at least 20 (20% positive PD-L1 staining), and the total study population irrespective of PD-L1 staining.
In the 255 patients who had the highest PD-L1 levels (CPS of at least 20), those treated with pembrolizumab alone had a median OS of 14.9 months compared with 10.7 months for patients who had the EXTREME regimen. This was equivalent to a significantly superior hazard ratio (HR) for the immunotherapy of 0.61 with a p value of 0.0007. "That was a highly statistically significant survival benefit for a therapy that proved to be much less toxic," said Burtness.
In the 512 patients who were PD-L1 positive but with lower levels of staining (CPS of at least 1.0), the median OS was 12.3 months with pembrolizumab monotherapy compared with 10.3 months in patients on the EXTREME regimen without any immunotherapy. The HR was 0.78 with a significant p value of 0.0086.
Pembrolizumab monotherapy remained non-inferior to EXTREME for OS even in the total population of 601 patients irrespective of PD-L1 staining. But it did not prolong PFS-even in patients with the highest PD-L1 expression (CPS of at least 20). Patients treated with pembrolizumab alone also had a lower objective response rate of 23 percent compared with 36 percent for EXTREME. But responses to pembrolizumab lasted about five times as long as those in patients treated with chemotherapy.
"For those patients, the responses were very durable," Burtness said. "Pembrolizumab appears to prolong life even when the cancer continues to grow."
Pembrolizumab alone had been less toxic compared to the EXTREME regimen, she notd. "There were fewer reported toxicities of all grades and fewer grade 3-5 toxicities." The quality-of-life analysis had not been completed, but in comparing the pembrolizumab-chemotherapy versus the EXTREME regimen the rate of all-grade toxicity had been identical between the two groups.
Research Outcomes
When asked about the clinical lessons emerging from KEYNOTE-048, Burtness said it showed head and neck cancer to have been amenable to immune checkpoint inhibition. "Early use of immune checkpoint inhibition in the metastatic setting has profound effects on survival," she noted. "And management of these patients may ultimately be partially driven by biomarkers."
The researchers concluded that pembrolizumab monotherapy significantly improved OS in comparison with the chemotherapy containing regimens in patients who were PD-L1 positive and was non-inferior in the total population. They also concluded that safety was favorable.
Combining pembrolizumab with chemotherapy also significantly improved OS in the total population in whom safety had been comparable to the other chemotherapy-containing treatments. "Whether pembrolizumab is given alone or with chemotherapy may depend on PD-L1 expression, and we are conducting analyses to answer this question," explained Burtness.
"We have never seen overall survival results like these in head and neck cancer. And I hope that pembrolizumab enters the standard of care for first-line treatment as soon as possible."
Tanguy Seiwert, MD, Head and Neck Cancer Program Director and Assistant Professor of Medicine at the University of Chicago, said KEYNOTE-48 had been the first study to show superior OS over the "decade-old" standard of care consisting of platinum-based chemotherapy plus cetuximab. "[It] establishes PD-L1 CPS as a valid marker for head and neck cancer that should be routinely measured in these patients."
But Seiwert noted that treatment benefit was not equally distributed but depends on a biomarker. "PD-L1 CPS expression will likely inform our choice between the two new options: Pembrolizumab alone-with a favorable side effect profile, and pembrolizumab combined with chemotherapy-which may be used in a larger group of patients," he said.
Higher PD-L1 expression had been associated with more benefit, but he said the exact cut points still needed to be determined. "Individual patient characteristics will play an important role as well. Separate analyses are needed in patients who have tumors with low or absent PD-L1 expression where there is potentially less benefit," noted Seiwert. And he believes further research is needed. "The usefulness of other biomarkers to select patients for treatment-such as tumor mutational burden-should also be examined."
Peter M. Goodwin is a contributing writer.