Objective: This meta-analysis aims to evaluate the efficacy and safety of antiprogressive disease (PD)-(L)1-based neoadjuvant therapy in head and neck squamous cell carcinoma (HNSCC) patients and identify potential prognostic biomarkers.
Data sources: Databases were systematically searched for prospective clinical trials evaluating the efficacy and safety of anti-PD-(L)1-based neoadjuvant therapy for HNSCC before January 12, 2024.
Review methods: We estimated the efficacy and safety of neoadjuvant immune checkpoint inhibitors. Subgroup and sensitivity analyses were further performed.
Results: A total of 570 patients from 20 studies were included. The pooled major pathological response (MPR), pathological complete response (pCR), and partial pathological response (PPR) rates were 30.7%, 15.3%, and 68.2%, respectively. Surgical complications, surgical delayed rate, all grade treatment-related adverse effects (TRAEs) and ≥Grade 3 TRAEs were 0.6%, 0.3%, 82.6%, and 9.7%, respectively. Best MPR or pCR rate was detected in patients receiving neoadjuvant anti-PD-(L)1 therapy + radiotherapy (with MPR rate of 75.5% and pCR rate of 51.1%) and neoadjuvant anti-PD-(L)1 therapy + chemotherapy groups (with MPR rate of 57.5% and pCR rate of 26.7%). No differences were detected in subgroups stratified by neoadjuvant treatment cycles, human papillomavirus (HPV) status, and tumor location. Patients with baseline Combined Positive Score (CPS) ≥ 20 have higher MPR and pCR rates compared to patients with CPS < 20. High Tumor Cell Proportion Score was also associated with MPR and pCR. Objective response rate is a strong predictor of MPR (odds ratio [OR] = 7.78, 95% confidence interval [CI] = 3.20%-18.91%) and pCR (OR = 3.24, 95% CI = 1.40%-7.48%).
Conclusion: Anti-PD-(L)1-based neoadjuvant therapy was effective and safe for HNSCC patients.
Keywords: CPS; TPS; head and neck squamous cell carcinoma; meta‐analysis; neoadjuvant anti‐PD‐(L)1 therapy; pathological response.
© 2024 American Academy of Otolaryngology–Head and Neck Surgery Foundation.