Consistent administration of cetuximab is associated with favorable outcomes in recurrent/metastatic head and neck squamous cell carcinoma in an endemic carcinogen exposure area: a retrospective observational study

Patient characteristics

Clinicopathological data of patients with HNSCC were confirmed by pathological examination of specimens from biopsy or surgery, and the positive samples were collected and analyzed. A total of 106 cases of RM HNSCC were identified with metastasis or recurrence and were deemed unsuitable for locoregional curative treatment at the Kaohsiung Medical University Hospital. The inclusion criteria included: age at diagnosis 20 years or older; tumor histology of squamous cell carcinoma (grade 1 to grade 3); ICD-9 site code-specific for the oral cavity (OC), hypopharynx (HPC), oropharynx (OPC), and larynx; and treatment with cetuximab from January 2016 to April 2019. The exclusion criteria included secondary malignancy; tumor histology of carcinoma in situ; and SCC of the nasopharynx or salivary glands.

Study design

This was an observational, retrospective, single-center, single-arm study, and the treatment schema is shown in Fig. 1. The collected medical and demographic data included age, gender, alcohol, betel nut usage, tobacco habits, and other clinical parameters obtained from the medical records or interviews with patients. The clinicopathological factors included types and grade of histology, size of tumor, lymph node status, surgical margin, perineural invasion, lymphovascular invasion, and extranodal extension. We defined CRT (chemoradiotherapy)-refractory patients as patients with disease progression during CRT or within three months of the end of CRT. The primary endpoints were median OS and PFS. Specifically, the median OS and PFS (defined as the time from registration to objective disease progression or death from any cause) were determined after the addition of cetuximab to chemotherapy. Other endpoints included the assessment of treatment response and disease control. This study was approved by the Institutional Review Board and Ethics Committee of Kaohsiung Medical University Hospital (KMUHIRB-E(II)-20190357). The data were analyzed anonymously, and therefore, no additional informed consent was required. All methods were performed in accordance with approved guidelines and regulations.

Treatment

All patients received cetuximab (250 mg/m²) weekly with a loading dose of 400 mg/m² until disease progression was noted. The regimen of chemotherapy included PF 75/1000 (cisplatin at 75 mg/m² or carboplatin at AUC = 5 every 3 weeks plus fluorouracil at 1,000 mg/m²/d for 4 days every 3 weeks), PF 60/800 (cisplatin at 60 mg/m² or carboplatin at AUC5 every 3 weeks plus fluorouracil at 800 mg/m²/d for 4 days every 3 weeks), taxane-based chemotherapy (docetaxel and cisplatin 75 mg/m² both at day 1 and every 3 weeks for four courses of paclitaxel 80 mg/m² weekly), and MTX (methotrexate 40 mg/m² weekly). The patients could receive chemotherapy or concurrent chemoradiotherapy with weekly cisplatin administration previously before recruitment.

Treatment response and safety assessment

All patients were followed regularly as outpatients at the medical oncology and department (OPD) of otorhinolaryngology. During the cetuximab treatment period, the patients visited the OPD of medical oncology weekly and that of otorhinolaryngology monthly. The evaluation of disease status included tumor site inspection, laboratory text, and imaging studies. Treatment response was assessed and determined using computed tomography (CT) or magnetic resonance imaging (MRI) at baseline (before cetuximab) and at 3-month intervals after treatment was started. Imaging within 4 weeks before cetuximab was acceptable, and imaging could be performed whenever clinical physicians suspected disease progression. RECIST version 1.1 was used to determine disease progression and tumor response.

The treatment response of patients was classified into four categories: complete response (CR, disappearance of all target lesions), partial response (PR, decrease in target lesion diameter sum >30%), progression disease (PD, increase in target lesion diameter sum >20%), and stable disease (SD, does not meet other criteria). The calculation of overall response rate (ORR), including patients classified as having complete and partial responses, was based on the best objective response achieved during cetuximab treatment. The calculation of disease control rate (DCR) included patients classified as having complete response, partial response, and stable disease. After disease progression, further treatments and survival status were documented every 3 months. Regarding safety assessment, treatment-related adverse events were monitored weekly throughout the study and were evaluated using Common Terminology Criteria for Adverse Events version 4.0.

Statistical analysis

The primary goal of the study was to analyze the outcome of cetuximab-based chemotherapy in recurrent or metastatic settings, including a comparison between median PFS and OS among patients receiving various cycles of cetuximab and regimens of chemotherapy. The location of primary sites (OC, OPC, or HPC), histological grade (Grades 1, 2, 3), tumor size and status (T1, T2, T3, T4), lymph node status (N0, N1, N2, N3), stage at initial diagnosis (I, II, III, or IV), surgery status (with or without previous surgery), CRT (with or without previous CRT), and chemotherapy before cetuximab therapy (with or without prior chemotherapy) were all included for analysis. Between-group comparisons were analyzed using Fisher’s exact test and Pearson’s chi-square test for various categorical variables. We calculated median PFS and OS using Kaplan–Meier analysis, and we analyzed differences between the curves using the log-rank test. We defined the median PFS as the time between the start of disease progression and treatment, including disease progression or death. Patients alive and without disease progression by the final follow-up visit were considered potential right censoring subjects, and the follow-up interval was truncated at the end of study. Univariate and multivariable analyses using the Cox proportional hazard model were preformed to analyze prognostic factors associated with cetuximab treatment. The factors for this analysis included age at initial diagnosis, location of primary sites, histological grade, pathological features (margin, lymphovascular invasion, perineural invasion, and extranodal extension), tumor size, lymph node status, stage at initial diagnosis, previous treatment before cetuximab (surgery, chemotherapy, or CRT), combined regimen and dosage of chemotherapy. All p-values were considered significant if p < 0.05 and were two-sided. Statistical analyses were performed using STATA version 11 (STATA Corp., TX, USA).

Baseline characteristics of patients

Clinical data from 106 patients (99 males and 7 females) with a median age of 55.1 years were collected. Among these patients, 65 patients (61.3%) had metastatic disease and 41 patients (38.4%) had recurrent disease with initiation of cetuximab. Almost all patients had addictions to alcohol or betel nuts or history of smoking, including 61 patients (57.5%) with exposure to all three carcinogens. Only 5 patients (4.7%) had no previous exposure to these risk factors. Regarding the tumor site, most of the primary sites had origins in the oral cavity (64.1%), followed by the hypopharynx (19.8%), and oropharynx (16.0%). The majority of patients had advanced disease, including T3-4, N2-3, or clinical stage 4. The details of basic information of the study population are listed in Table 1.

Treatment modality

With respect to prior treatment before cetuximab treatment, most patients had undergone various HNSCC treatments, including surgery (78.3%), chemotherapy (81.1%) and CRT (80.2%). In addition, there were 34 CRT-refractory patients who suffered from disease progression during CRT or within three months of the end of CRT.

The major reason for cetuximab treatment was recurrent disease with metastatic tumors. The median number of cycles of cetuximab was 11 (2–24), with 60 patients receiving ≥11 cycles of cetuximab, and 46 patients receiving <11 cycles of cetuximab. Among these patients, 76 patients received chemotherapy with the EXTREME regimen (cisplatin and fluorouracil) and 17 patients received taxane-based chemotherapy. The median number of cetuximab administration cycles in these 76 patients with a PF regimen was 11 (range: 2–24) while the median number of cetuximab cycles in 17 patients using taxane-based regimen was 12 (range: 4–23). There was no significant difference in the number of cetuximab cycles between the two groups (p = 0.427). The details of the treatment modalities are shown in Table 2. The demographic data of various cetuximab cycles (≥11 and <11) are shown in Tables S1 and S2. Interestingly, there was no difference in terms of previous treatments, including surgery, chemotherapy, and CRT, between patients who received <11 cycles of cetuximab and those who received ≥11 cycles of cetuximab.

Treatment outcomes

After cetuximab treatment, clinical responses were observed in 30 patients including 1 complete response and 29 partial responses, with ORR of 28.3%. When the patients with stable disease (n = 21, 19.8%) were included in the analysis, the disease control rate was 48.1%. The median PFS and OS were 5 months and 9.23 months, respectively. As of the cut-off date, only one patient did not progress, and 38 patients survived. The median PFS was 5 months (95% CI [3.0–6.0] months) and the median OS was 9.23 months (95% CI [7.03–13.84] months). The treatment responses according to various stages are shown in Table S3.

The median PFS in various subgroups stratified by treatment modalities is shown in Fig. 2. Notably, the patients who received more cetuximab treatment (≥11 cycles) had a better median PFS than did patients who received less cetuximab (7 months vs. 3 months, p < 0.001). The median PFS was longer in patients without prior CRT (6 months vs. 4 months, p = 0.046). Other factors including chemotherapy regimen (PF or taxane-based), chemotherapy dose (PF dose), or CRT refraction status did not lead to significant effect on PFS. In regard to analysis of OS, the patients who received more cetuximab treatment (≥11 cycles) had a better median OS than those who received less cetuximab (12.43 months vs. 4.46 months, p < 0.001). Other factors, including chemotherapy regimen and dose, did not lead to significant effects on PFS. The OS curves are shown in Fig. 3.

Next, we applied a landmark method for further validation. Because responses could be observed within the first 3 months following cetuximab exposure, a 3-month landmark was used. After excluding patients who progressed or died within the three months, the patients with more cycles of cetuximab (≥ 11 cycles) still showed better median PFS (8 months vs. 2 months, p = 0.057) and OS (13.9 months vs. 5.07 months, p = 0.0002) than the patients treated with fewer cycles of cetuximab.

To clarify the effects of CRT-refraction on survival, we evaluated median PFS and OS in patients with or without CRT-refraction. In the non-CRT-refractory cohort (n = 72), the median PFS and OS were 5.00 months (95% CI [3.00–7.00]) and 10.43 months (95% CI [7.03–14.64]), respectively. The 3-year OS was 28.72% (95% CI [17.25–41.24]). On further evaluation of these 72 subjects, 27 patients with <11 cetuximab cycles obtained a 3-year PFS rate of 3.70% (95% CI [0.27–15.90), and a 3-year OS rate of 2.22% (95% CI [0.18–10.15]). Additionally, 45 patients with ≥ 11 cetuximab cycles obtained a 3-year PFS rate of 11.57% (95% CI [1.04–36.08]), and a 3-year OS rate of 37.07% (95% CI [21.60–52.59]). The patients treated with more cetuximab cycles also showed a better median PFS and OS then did the patients treated with fewer cetuximab cycles, shown in Fig. 4.

In the CRT-refractory patients, the median PFS and OS were 3.00 months (95% CI [3.00–6.00]) and 7.8 months, respectively. The 3-year OS rate was 25.30% (95% CI [10.32–43.53]). Six CRT-refractory patients who used taxane-based regimens obtained a median PFS and OS of 3.00 months (95% CI [2.00–8.00]) and 5.62 months (95% CI [2.03–NA]), respectively. The 3-year OS was 16.67% (95% CI [0.77–51.68]).

Risk factor investigation for disease progression

Risks of disease progression were analyzed using univariate regression consisting of parameters as age, alcohol, betel nuts, tobacco consumption, tumor site, margin positivity, histologic features (including lymphovascular invasion, perineural invasion, and extranodal extension), tumor size, lymph node status, stage, previous treatment modality (including surgery, chemotherapy, and CRT), treatment status, cetuximab cycles, dose, and regimens of chemotherapy. In addition, a subsequent multivariable regression analysis was performed to evaluate the significant progression factors in univariate analysis.

As shown in Table 3, positive perineural invasion was the independent factor related with shorter median PFS. N3 disease showed a trend toward poorer PFS (p = 0.055, univariate analysis). After adjustment for other different variables in the multivariable analysis, this difference became significant (HR = 2.57; p = 0.043). Significantly, treatment with more cetuximab cycles (≥11 cycles) was a favorable factor associated with better median PFS (HR = 0.19; p < 0.001, and HR = 0.18; p < 0.001 in univariate and multivariable analysis, respectively).

Determining the risk factor for poorer overall survival

Similar clinicopathological factors were analyzed for overall survival. N2 disease had a significantly negative impact on OS (HR = 2.09; P = 0.022 and HR = 4.79; p = 0.006 in univariate and multivariable analyses, respectively). Treatment with more cetuximab cycles showed a significant, positive effect on OS (HR = 0.46; p = 0.002 and HR = 0.48; p = 0.010 in both univariate and multivariable analyses, respectively). Other factors with trends toward shorter OS included N3 disease (p = 0.170). After adjustment for other variables, this difference became significant in the multivariable analysis (HR = 7.34; p = 0.005). These results are shown in Table 4.

Although endemic habits showed no significant impact on PFS and OS, multiple endemic habits might increase risk in PFS and OS compared to single or double endemic habits. The impact of multiple endemic habits on PFS and OS are summarized in Table S4.

Safety and tolerability

All grades and the worst grade 3 and grade 4 treatment-related adverse events (AEs) in patients receiving cetuximab therapy are listed in Table 5. Among the patients treated with the platinum/5FU and cetuximab regimen, the most common AEs were skin rash (2.6%), anemia (2.6%), neutropenia (1.3%), vomiting (1.3%) and fever (1.3%). Among patients treated with taxane-based regimens, only one patient suffered from grade 3 fever (5.9%). There were no grade 3 or grade 4 AEs in other groups. In general, skin rash was the most frequent cetuximab-related AE; however, most of patients tolerated it. There was no interstitial lung disease observed in our patients.