Abstract
BACKGROUND: In recent years, the addition of cetuximab to chemotherapy has improved treatment outcomes for patients with recurrent/metastatic head and neck squamous cell carcinoma (R/M HNSCC). In this study, we present the real-world survival data of R/M HNSCC patients who received cetuximab-containing regimens from thirteen medical centers in Taiwan, as well as a three-level risk stratification model for this patient population. METHODS: This study enrolled R/M HNSCC patients from thirteen medical centers in Taiwan who received cetuximab-containing regimens from January 1, 2017 to June 6, 2022. The cases were divided into a training cohort and a validation cohort based on the start of treatment. Overall survival (OS) was evaluated in both cohorts and exploratory analysis was performed to identify associated adverse clinical and laboratory factors. The results of the exploratory analysis were used to construct a three-level risk stratification prediction model for OS. RESULTS: A total of 1434 patients with R/M HNSCC were enrolled in this study and received cetuximab-containing regimens. The overall population had a median OS of 8.57 months (95% CI: 8.07 - 9.08). Multivariate analysis of the training cohort identified poor ECOG performance status, heavy alcohol consumption, and prior adjuvant CCRT or lack of prior RT as adverse prognostic factors. Comparison of laboratory data between patients with OS≦6 and OS > 6 also revealed unfavorable factors, including increased white blood cell count, decreased hemoglobin level, increased platelet count, increased absolute neutrophil count, decreased absolute lymphocyte count, and increased neutrophil-to-lymphocyte ratio. Using forward prediction, a three-level risk stratification prediction model was constructed using the variables of ECOG performance status, alcohol consumption, skin metastasis, modality of radiation therapy, hemoglobin level, and neutrophil-to-lymphocyte ratio. The median OS in the low-risk, intermediate-risk, and high-risk groups were 12.02 months (95% CI 10.44-13.61), 7.5 months (95% CI 7.33-8.17), and 4.01 months (95% CI 3.94-4.08), respectively, with a log-rank test p-value < 0.001. CONCLUSION: This study presents a three-level risk stratification model with strong prediction ability for OS in R/M HNSCC patients who received cetuximab-containing regimens. The results are based on real-world data and may provide valuable information for clinicians in treatment planning and future drug development.