We identified 498 patients with R/M HNSCC between 2015 and 2022 from THNSCRD for outcomes comparison. Patients’ characteristics are presented in Table 1. Among eligible patients, 92% were male, nearly 60% of them younger than 60 years, and more that 80% of them had Eastern Cooperative Oncology Group Performance Status (ECOG PS) 2. Half of our patients had their primary tumor location in the oral cavity, followed by hypopharynx, oropharynx, and larynx. P16 testing was only performed on 30% of our patients with 3–4% positive rate. PD-L1 expression was only estimated on 40% of our patients with 10% positive rate. The main reason of missing PD-L1 data was that archive tissue is not suitable for PD-L1 staining. As for initial stage, 80% of our patients had stage 3–4 disease at diagnosis. Nearly 60% of the patients have ever received curative surgery, induction chemotherapy or chemoradiotherapy. Half of our patients had distant metastasis while the remainder had local recurrent disease. Approximately 55% of our patients had platinum-sensitive disease with platinum-free interval (PFI) of more than 6 months, and 45% of our patients had platinum-refractory disease with PFI of less than 6 months. Patients were stratified according to second-line treatment. In summary, 259 patients continued cetuximab-based regimen as second-line treatment (TBP group), while 239 patients did not continue cetuximab treatment (non-TBP group). Table 1 presents the distributions of characteristics between TBP and non-TBP groups. The clinical variables in these two groups were similar. There were no significant differences in age, gender, ECOG PS, primary tumor location, P16 testing, PD-L1 expression status, initial stage, previous history of treatment, disease status upon enrollment, and PFI.

Treatment disposition of 498 patients with R/M HNSCC are summarized in Table 2. The chemotherapy regimens were similar between these two groups. The most common first-line chemotherapy regimen was cetuximab, platinum, and 5-fluorouracil (EXTREME regimen), accounting for more than 60% in each group. Moreover, 13% of our patients in each group received cetuximab, platinum, and tegafur/uracil (UPEx regimen). A total of 10% of our patients were treated with cetuximab, platinum, and taxane (TPEx regimen). After progression on first-line chemotherapy, every patient underwent second-line chemotherapy as per our inclusion criteria. For patients in the TBP group, all patients received cetuximab-containing regimen as second-line treatment, accounting for 54% cetuximab plus taxane, 15% cetuximab plus immune checkpoint inhibitors (ICI), 13% EXTREME regimen, and 4% cetuximab plus methotrexate. For patients in the non-TBP group, every patient was treated with non-cetuximab regimen as second-line treatment, accounting for 48% taxane-based regimen, 18% ICI-based regimen, 12% platinum plus 5-fluorouracil, and 6% methotrexate.

The median follow-up period was 11.1 months. Among these patients, 77% died, and cancer was the main cause of their deaths. Survival curves with TTD, PFS2, and OS 1, 2 are plotted in Fig. 1, respectively. In Fig. 1A, B, TTD and PFS2 were significantly longer in the TBP group than those in the non-TBP group. Median TTD and PFS2 were 7.7 months versus 4.6 months (p < 0.001) and 11.7 months versus 7.7 months (p < 0.001) for TBP and non-TBP group, respectively. In Fig. 1C, D, median OS was significantly better in the TBP group than those in the non-TBP group. Median OS1 and OS2 were 14.1 months versus 10.9 months (p = 0.016) and 9.8 months versus 6.5 months (p = 0.002) for TBP and non-TBP group, respectively. Subgroup analysis suggested survival benefits remained significant regardless PD-L1 expression and P16 status, as shown in Fig. 2. Median OS1 was 22.8 months versus 9.2 months (p = 0.013) and 13.0 months versus 9.9 months (p = 0.010) for patients with PD-L1 positive and PD-L1 negative disease, respectively. Median OS1 was not reached (NR) versus 10.9 months (p = 0.046) and 14.1 months versus 10.9 months (p = 0.002) for patients with P16 positive and P16 negative disease, respectively. Table 3 presents the results of the Cox regression analyses of survival in the total population. After adjusting for potential confounding factors, ECOG PS and cetuximab TBP were independent prognosticators correlated with survival.

A Time to cetuximab discontinuation, B progression-free survival 2, C overall survival 1, and D overall survival 2 of 498 patients with R/M HNSCC receiving first-line cetuximab-based chemotherapy, stratified by TBP

Overall survival 1 of patients with R/M HNSCC receiving first-line cetuximab-based chemotherapy, stratified by TBP. A Patients with PD-L1 positive, B patients with PD-L1 negative, C patients with P16 positive, and D patients with P16 negative