In the phase III KEYNOTE-689 trial, perioperative use of the PD-1 inhibitor pembrolizumab given with standard therapy significantly improved event-free survival in newly diagnosed, previously untreated patients with locally advanced head and neck squamous cell carcinoma, researchers reported at the 2025 American Association for Cancer Research (AACR) Annual Meeting.1 These patients were also significantly more likely to achieve a major pathologic response.
Among all patients, the reduction in recurrence and death was 27% (P = .00411), and for patients with a combined positive score (CPS) ≥ 10, the risk was reduced by 34% (P = .00217), according to Ravindra Uppaluri, MD, PhD, the Brigham and Women’s Hospital Endowed Chair in Otolaryngology, Director of Head and Neck Surgical Oncology at the Dana-Farber Brigham Cancer Center, and Associate Professor of Otolaryngology/Head and Neck Surgery at Harvard Medical School.
“Neoadjuvant pembrolizumab followed by surgery and adjuvant pembrolizumab, concurrent with and after postoperative (chemo)radiotherapy, represents a new standard of care in the treatment of patients with resectable, locally advanced head and neck cancer,” said Dr. Uppaluri, who indicated this is the first major change in treatment for this particular patient subset in about 20 years.
As Dr. Uppaluri explained, the front-line approach to newly diagnosed head and neck squamous cell carcinoma is typically surgery followed by radiation with or without chemotherapy, though outcomes for many patients remain unsatisfactory. The premise of perioperative immunotherapy is that the immunotherapeutic agent—in this case, pembrolizumab—enhances the immune response before surgery (when the antigen load is high), and with continued administration, it helps eradicate any residual disease.
The primary endpoint, event-free survival, was met in KEYNOTE-689, with these findings presented:
- Among all patients, median event-free survival was 51.8 months in the pembrolizumab arm vs 30.5 months in the control arm (hazard ratio [HR] = 0.73; P = .00411).
- In the CPS ≥ 10 population (the primary endpoint), median event-free survival was 59.7 months vs 26.9 months, respectively (HR = 0.66; P = .0022). This was the same median event-free survival for the CPS ≥ 1 population (HR = 0.70; P = .0014).
- In the analysis at 3 years, 59.8% of the pembrolizumab arm vs 45.9% of the standard therapy arm remained event-free in the CPS ≥ 10 population, as did 58.2% and 44.9% of the CPS ≥ 1 population.
About KEYNOTE-689
KEYNOTE-689 trial enrolled 714 patients with stage III or IVA newly diagnosed, resectable head and neck squamous cell carcinoma (including in the larynx, hypopharynx, and oral cavity). Patients were randomly assigned to two cycles of neoadjuvant and three cycles of concurrent (during postoperative radiation) pembrolizumab followed by 12 cycles of adjuvant pembrolizumab at 200 mg intravenously every 3 weeks. All patients received standard therapy, which included surgery plus postoperative radiotherapy at doses depending on risk: 60 Gy in 30 fractions for low-risk patients, 66 Gy in 33 fractions plus three cycles of concurrent cisplatin for high-risk patients, and 70 Gy in 35 fractions plus cisplatin for those with gross residual disease.
All patients underwent surgery and then received pathology-directed adjuvant therapy; 363 of these patients were randomly assigned to receive pembrolizumab before and after surgery. The statistical design involved a sequential testing strategy that included event-free survival as the primary endpoint, first tested in the CPS ≥ 10 population by blinded independent review. The population of 714 included 682 patients determined to be CPS ≥ 1 and 465 who tested CPS ≥ 10. Median follow-up was 38.3 months.
Major pathologic responses were achieved by none of the patients in the control arm but by 13.7% of the CPS ≥ 10 group receiving pembrolizumab, by 9.8% of the CPS ≥ 1 pembrolizumab arm, and by 9.4% of all participants receiving pembrolizumab; these differences were all highly significant (P < .00001). Acknowledging that the achievement of major pathologic responses in this malignancy is exciting but there is room for improvement, Dr. Uppaluri said this study “opens up the idea of considering additional neoadjuvant immunotherapy approaches, whether that be doublets or other combinations, which may lead to better pathologic responses that may eliminate surgery in select patients.”
Overall survival in the CPS ≥ 10 population was a key secondary endpoint. Median overall survival was not reached in the pembrolizumab arm and was 61.8 months with standard therapy; 68.2% of the pembrolizumab arm were alive at 3 years vs 59.2% of the control arm (HR = 0.72; P = .02; 95% confidence interval = 0.52–0.98). The significance boundary was not met at this first interim analysis, but further analysis is forthcoming.
In a press conference, the question of PD-L1 status for patient selection was raised. Dr. Uppaluri noted that the CPS ≥ 10 population constituted about 65% of the study population, and the CPS ≥ 1 population encompassed 95% of patients. “It’s hard, therefore, to make conclusions about PD-L1–negative patients,” he said. “As a whole, I would say this regimen is for all patients.”
Adverse Events
Treatment-related adverse events grade ≥ 3 occurred at similar rates in both arms of the study: 45% with pembrolizumab plus standard therapy vs 43% with standard therapy alone. Four deaths were reported in the pembrolizumab arm, and one death was reported in the control arm. As expected, more immune-mediated adverse events occurred in the patients receiving pembrolizumab (43%), most commonly hypothyroidism, but more stomatitis and radiation skin injury were observed with standard therapy.
DISCLOSURE: Dr. Uppaluri has received consulting fees from MSD, Daiichi Sankyo, and Regeneron.
REFERENCE
1. Uppaluri R, et al: 2025 AACR Annual Meeting. Abstract CT001. Presented April 27, 2025.
