Neoadjuvant immunotherapy in combination with chemotherapy in resectable locally advanced head and neck squamous cell carcinoma: A randomized, open label, phase II clinical trial.

6043 Background: The efficacy and safety of neoadjuvant immunotherapy (NAI) for resectable locally advanced head and neck squamous cell carcinoma (LAHNSCC) remain unclear, requiring further exploration. While PD-1 inhibitors combined with chemotherapy have shown promise, most regimens focus on single-target inhibitors. Dual-target inhibitors, such as PD-1/CTLA4 or PD-1/VEGF combinations, demonstrated superior efficacy in recurrent/metastatic HNSCC. This study aims to compare the efficacy and safety of single- and dual-target NAI combined with chemotherapy for resectable LAHNSCC to identify the optimal strategy. Methods: This phase II randomized trial will enroll resectable LAHNSCC patients eligible for surgery. Patients will be randomized into three cohorts: Cohort 1 will receive ivonescimab (PD-1/VEGF antibody, 10 mg/kg), Cohort 2 will receive cadonilimab (PD-1/CTLA-4antibody, 6 mg/kg), and Cohort 3 will receive penpulimab (PD-1 antibody, 200 mg), all in combination with cisplatin and nab-paclitaxel. Dose adjustments are allowed based on toxicity, and surgery will be performed within 2-4 weeks after 3 cycles of neoadjuvant treatment. Patients achieving pCR will receive 16 cycles of adjuvant immunotherapy. Those without pCR will undergo adjuvant radiotherapy or chemoradiotherapy, followed by 16 cycles of adjuvant immunotherapy. The primary endpoints were pCR and safety. Secondary endpoints include MPR, ORR, EFS and OS. Results: A total of 24 patients were enrolled, with 13 evaluable for analysis. The median age was 58 (range 34-70). The primary site included oral cavity (n=7), oropharynx (n=2, including 1 HPV-positive), and larynx (n=4). The clinical stages were as follows: T2 (n = 3), T3 (n = 6), T4a (n = 5), and cN0/1 (n = 3), cN2 (n = 9), cN3 (n = 1). The pCR rates were 80% (4/5), 0% (0/2), and 66.7% (4/6) in Cohort 1, 2, and 3, respectively. The major pathologic responses (MPR) rates were 80% (4/5), 100% (2/2), and 66.7% (4/6) in Cohort 1, 2, and 3, respectively. Pathological non-response (pNR) occurred more frequently in Cohort 3 (2/6) and absent in Cohorts 1 (0/5) and 2 (0/2). The ORR was 100% in Cohort 1 (CR: 3/5, PR: 2/5) and Cohort 2 (100%, PR: 2/2), while Cohort 3 showed a lower ORR of 83.3% (CR: 3/6, PR: 2/6, SD: 1/6). There were no Grade ≥3 TRAEs or unexpected surgical delays/complications. Conclusions: Neoadjuvant single- or dual-target immunotherapy combined with chemotherapy showed promising pathological responses in resectable LAHNSCC. While dual-target therapy showed potential benefits, the small sample size limits definitive conclusions. The treatment was well-tolerated, with no serious TRAEs. Further analyses will be conducted as patient enrollment continues in larger cohorts. Clinical trial information: NCT06444009 . RECIST and pathologic response. RECIST Pathologic Response CR PR SD PD MPR pPR pNR pCR Cohort 1 3 2 4 1 Cohort 2 2 2 Cohort 3 3 2 1 4 2