Recurrent or metastatic head and neck squamous cell carcinoma (HNSCC) is frequently complicated by dysphagia and malnutrition due to tumor location and prior treatments. Reduced skeletal muscle mass is a phenotypic criterion of malnutrition and has been linked to oncologic outcomes. Standard measures of muscle mass often use abdominal CT at L3, but routine head and neck imaging for HNSCC follow-up does not include that level. Muscle measures at cervical levels may be affected by prior neck surgery or radiation.
Temporal muscle thickness (TMT) measured on routine head CT has been reported to correlate with whole-body skeletal muscle mass and to have prognostic utility in other cancers. Prior HNSCC reports focused on curative-intent cases; the association between TMT and outcomes after immune checkpoint inhibitors (ICIs) in recurrent or metastatic (R/M) HNSCC had not been reported. This study aimed to determine whether TMT on routine head CT is associated with objective response and survival in patients with R/M HNSCC treated with nivolumab or pembrolizumab.
This was a retrospective cohort study of patients treated at The University of Osaka Hospital between May 2017 and May 2023. Eligible patients had histologically or cytologically confirmed HNSCC, received nivolumab or pembrolizumab for R/M disease, and had axial CT images at the level of the Sylvian fissure and orbital roof within three months before ICI start. Exclusion criteria included prior ICI therapy for other malignancies, tumor involvement or deformation of the temporal muscle, and insufficient clinical data.
Clinical data collected from the medical record included sex, age, Eastern Cooperative Oncology Group performance status (PS), TMT measured within three months prior to ICI, primary tumor site, human papillomavirus (HPV) status, the ICI used (nivolumab or pembrolizumab), and chemotherapy regimens. Data access occurred between December 1, 2024 and March 31, 2025; identifiers were removed before analysis.
TMT was measured on axial CT images at the level where the Sylvian fissure was most clearly visualized. Measurements were taken perpendicular to the long axis of the temporalis muscle on the right and left sides; the mean of bilateral values was recorded as the patient’s TMT. Sex-specific cutoff values for TMT were established using time-dependent receiver operating characteristic analysis for 12-month overall survival.
Objective response was assessed using RECIST version 1.1. Response rate (RR) was the proportion of patients achieving complete response (CR) or partial response (PR). Disease control rate (DCR) included CR, PR, or stable disease (SD). Overall survival (OS) was time from ICI initiation to death from any cause. Progression-free survival (PFS) was time from ICI initiation to disease progression or death from any cause.
Categorical associations were tested with chi-square tests; continuous-to-categorical comparisons used Kruskal–Wallis tests. Logistic regression estimated odds ratios (ORs) for objective response. Survival estimates used Kaplan–Meier curves and log-rank tests; Cox proportional hazards models provided hazard ratios (HRs). A two-sided p value < 0.05 was considered significant. Sex-specific TMT cutoffs for 12-month OS were defined by time-dependent ROC analysis. Analyses were performed in JMP version 17 and R (timeROC and survival packages).
The study protocol was approved by the Institutional Review Board of The University of Osaka (16329). The requirement for written informed consent was waived because of the retrospective design, and the study complied with the Declaration of Helsinki and relevant Japanese ethical guidelines.
From an initial pool of 130 patients who met inclusion criteria, exclusions were applied: non-SCC histology (n = 18), lack of CT images at the desired level (n = 20), and temporal muscle involvement by tumor (n = 1). The final analysis included 109 patients. The cohort comprised 89 men and 20 women; 66% were aged 65 years or older at ICI start. The most common primary tumor site was the hypopharynx, followed by oropharynx and oral cavity. Cytotoxic drug–containing regimens were administered to 33% of patients.
Sex-specific TMT cutoffs selected by time-dependent ROC for 12-month OS were 5.53 mm for men and 3.92 mm for women. Objective response was evaluable in 105 patients. There were 12 complete responses, 30 partial responses, 18 stable disease, and 45 progressive disease, yielding a response rate of 40.0% and a disease control rate of 57.1%.
Higher TMT was associated with a greater likelihood of objective response. In univariable analysis the odds ratio for response in the high-TMT group was 2.64 (95% CI 1.18–5.90; p = 0.018). After adjustment for chemotherapy use, the association remained significant (adjusted OR 2.42, 95% CI 1.05–5.58; p = 0.038).
Survival outcomes favored the high-TMT group. One-year PFS was 36.5% in the high-TMT group versus 15.8% in the low-TMT group (absolute difference 20.7%; p = 0.010); the adjusted hazard ratio for PFS was 0.57 (95% CI 0.36–0.90; p = 0.017). One-year OS was 73.1% versus 47.4% (absolute difference 25.7%; p = 0.001); the adjusted hazard ratio for OS was 0.54 (95% CI 0.34–0.86; p = 0.010).
The robustness of the TMT associations was evaluated by treating mean TMT as a continuous variable in regression models, by adding body mass index (BMI) to Cox models to assess whether BMI explained the TMT–survival association, and by examining treatment-related variables including chemotherapy use and choice of ICI agent. Results reported in the source indicate these analyses were performed; specific numeric results for all sensitivity tests beyond those described above were reported in the manuscript tables and figures.
In this retrospective cohort of patients with R/M HNSCC treated with nivolumab or pembrolizumab, temporal muscle thickness measured on routine head CT was independently associated with both objective response and survival. Because head CT is often available in HNSCC follow-up and TMT is measurable at the Sylvian fissure level, TMT may offer a practical, host-related imaging biomarker for prognostic risk stratification in this population. The authors note that cervical-level muscle measures can be confounded by prior treatment effects in HNSCC, whereas temporal muscle may better reflect whole-body skeletal muscle mass.
TMT on routine head CT was independently associated with response and survival in patients receiving immune checkpoint inhibitors for R/M HNSCC. The metric may serve as a practical imaging marker for prognostic stratification; further validation in external cohorts would be needed.
The study was funded by Japan Society for the Promotion of Science KAKENHI grants awarded to Y.T. The funders had no role in study design, data collection, analysis, or decision to publish. The authors declared no competing interests. All relevant data are provided within the manuscript and supporting information.