Concurrent chemoradiotherapy combined with nimotuzumab in stage III–IVa nasopharyngeal carcinoma: a retrospective analysis

The efficacy and safety of nimotuzumab (NTZ) added to concurrent chemoradiotherapy (CCRT) were investigated in patients with stage III–IVa nasopharyngeal carcinoma (NPC). Patients with stage III–IVa NPC treated with CCRT, with or without NTZ, were screened between January 2015 and December 2017. We compared patients’ overall survival (OS), progression-free survival (PFS), locoregional recurrence-free survival (LRFS), and distant metastasis-free survival (DMFS) between different therapeutic regimens. Propensity score matching (PSM) was applied to reduce the selection bias. Nomogram models were developed to predict the survival of CCRT with or without NTZ. Four hundred and twenty-six patients were included after PSM, with 213 patients in each regimen. Compared with NPC patients receiving CCRT alone, patients who received NTZ plus CCRT treatment had significantly better OS (5 year OS, 76.1 vs. 72.3%, P = 0.004), PFS (5 year PFS, 73.2 vs. 69.0%, P = 0.002), and LRFS (5 year LRFS, 73.2 vs. 69.0%, P = 0.028). A multivariate Cox regression analysis demonstrated that, compared with receiving CCRT alone, NTZ plus CCRT was an independently positive factor for OS, PFS, and LRFS. No significant difference was observed in the major toxicities between the two treatments (all P > 0.05). In addition, the nomogram presented good accuracy for predicting the prognosis of NPC patients. CCRT combined with NTZ presented favorable clinical outcomes for stage III–IVa NPC patients with good tolerance and similar toxicity compared to CCRT alone. A prospective, randomized clinical trial is essential to validate the current findings.


Introduction
Nasopharyngeal carcinoma (NPC) commonly occurs in Southeast Asia and Southern China. It is characterized by a relatively high sensitivity to radiotherapy and chemotherapy . Radiotherapy with or without chemotherapy is the primary treatment modality for NPC patients . According to the 2021 Chinese Society of Clinical Oncology and American Society of Clinical Oncology guidelines for nasopharyngeal carcinoma, concurrent chemoradiotherapy (CCRT) is still recommended as the standard treatment for stage II-IVa NPC patients (Yu-Pei et al. 2021). Cisplatin-based chemotherapy with intensitymodulated radiotherapy (IMRT) has been routine care for NPC patients in recent years. Numerous studies have demonstrated that chemotherapy concurrent to radiation therapy provides better tumor control and more favorable outcomes (Meng-Xia et al. 2015). However, nearly 30% of NPC patients may still experience treatment failure after CCRT due to the relapse in situ and distant metastasis (Lee et al. 2005;Brigette et al. 2008). Therefore, identification of a more effective therapeutic regimen with acceptable toxicity is urgently needed for NPC patients.
The epidermal growth factor receptor (EGFR) is a member of the ErbB family of receptor tyrosine kinases. This receptor is over-expressed in about 90% of the head and neck squamous cell carcinomas, including NPC (Fortunato et al. 2001;Yewale et al. 2013). In addition, the EGFR is a promising new therapeutic target in malignant carcinoma (Fortunato and Giampaolo 2001). Several EGFR inhibitors, such as cetuximab (CTX), panitumumab, afatinib, and erlotinib, have elicited positive outcomes in clinical trials (Dorsey et al. 2013;Prenen et al. 2013;Yewale et al. 2013). Thus, anti-EGFR therapy may be a potential treatment combined with CCRT for NPC patients.
The most commonly used anti-EGFR monoclonal antibodies in NPC treatment are CTX and nimotuzumab (NTZ). Numerous previous studies have investigated the efficacy of anti-EGFR targeted therapy in NPC, though the results are controversial (Xia et al. 2013;Yang et al. 2017;Wei-Xiong et al. 2017;Xu et al. 2015;Rui et al. 2017bRui et al. , 2017aJian-Feng et al. 2017). Xia et al. and You et al. found that the CTX or NTZ plus CCRT was more effective than CCRT alone (Wei-Xiong et al. 2017;Xu et al. 2015). In contrast, Li et al. found that adding CTX to the CCRT regimen did not improve the prognosis in stage II to IV NPC patients (Yang et al. 2017). Furthermore, the addition of CTX treatment significantly increased the toxicity rate of acute mucositis and acneiform rash (Yang et al. 2017;Xu et al. 2015). Compared with CTX, NTZ caused less severe dermatological toxicity and exhibited a longer half-life (Tania et al. 2003). Previous studies with a small sample size have demonstrated that NTZ plus CCRT showed positive results for locally advanced NPC patients (Zhi-Gang et al. 2016;Zhi-Qiang et al. 2019). Therefore, studies with a larger population are needed to validate the efficacy of adding NTZ to the therapeutic schedule.
Nomograms can assess the outcomes of individual patients by graphical depictions (Kattan et al. 2007), and they have been applied to various types of cancers (Yizhou et al. 2013;Dong-Seok et al. 2012;Karakiewicz et al. 2007). Furthermore, nomograms have been proposed as an alternative method to guide the therapeutic regimen for cancer patients because of their specificity and visualization (Karakiewicz et al. 2007;Mariani et al. 2005). Thus, the primary aim of the current study was to evaluate the efficacy of NTZ in stage III-IVa NPC patients. A secondary purpose was to develop a practical model by combining NTZ with prognostic biomarkers to predict the survival rate for NPC patients individually.

Patient selection
Between January 2015 and December 2017, a total of 11,382 patients who were diagnosed with nasopharyngeal carcinoma at Sun Yat-sen University Cancer Center were retrospectively enrolled. The pre-treatment clinical information of the patients was recorded, including blood tests, imaging examination, and histological diagnosis. Continuous variables were transformed into categorical variables. In this study, the median plasma EBV-DNA level was 2500 copies/ ml (range, 0-2.74*10 6 copies/ml) and chosen as the cutoff level. The methodology for detecting plasma EBV DNA has been described previously (Jian-Yong et al. 2004). All patients had histologic confirmation of NPC, and enhanced computed tomography (CT)/MRI verified the metastatic disease and clinical stage. The disease was restaged according to the International Union Against Cancer/American Joint Committee on Cancer (UICC/AJCC) TNM classification (8th edition, 2017) based on clinical and radiography data (Jian Ji et al. 2016;Ling-Long et al. 2017;William et al. 2017). The exclusion criteria are presented in Fig. 1. Briefly, patients were ruled out if they had clinical stage I-II disease; distant metastasis; no/unclear therapy; no receive concurrent chemotherapy; other targeted drugs; nimotuzumab less than six cycles; or lost to follow-up. Finally, 3340 patients were included in the study cohort; 3127 patients received CCRT alone, and 213 patients received CCRT combined with NTZ. In the 1:1 propensity score matching (PSM) analysis, patients who accepted nimotuzumab plus CCRT were individually matched to patients receiving CCRT alone according to body mass index (BMI), the levels of C reactive protein (CRP), and lactate dehydrogenase (LDH), which were significantly different between the two treatment regimens on the patients' baseline characteristic. The ethics committee of SYSUCC approved the current study (B2021-366-01). Written informed consent for therapy was obtained from all patients.

Radiotherapy
All patients received IMRT in Sun Yat-sen University Cancer Center. Briefly, gross tumor volume (GTVnx) included the primary tumor, and GTVnd was the volume of clinically positive cervical lymph nodes. GTVnx and GTVnd were verified by MRI/CT or PET-CT imaging and the nasopharyngoscopy findings. The first clinical tumor volume (CTV1) was defined as the gross tumor volume within 0.5-1.0 cm margin (0.2 to 0.3 cm posterior margin), including the high-risk sites of microscopic extension and the nasopharynx. The second clinical tumor volume (CTV2) was derived as a geometric expansion of CTV1 plus 0.5-1.0 cm margin (0.2-0.3 cm posterior margin) to cover the low-risk areas of microscopic extension, the level of lymph node, and the potentially metastatic neck area. The radiotherapy dose in this retrospective study ranged from 2·00 Gy to 2·33 Gy per fraction with five daily fractions per week for 6-7 weeks. The prescribed total radiation dose was 66-70 Gy to the planning target volume (PTV), 60 Gy to PTV1, 54 Gy to PTV2, and 60-66 Gy to the PTV of the involved cervical lymph nodes in 28-33 fractions. All patients received treatment once daily, five fractions weekly. The dose was applied within the tolerance according to the RTOG 0225 protocol.

EGFR-antibody therapy and chemotherapy
All patients received cisplatin-based concurrent chemotherapy, and the treatment regimen of concurrent chemotherapy was 100 mg/m 2 cisplatin on day 1 every three weeks, or 30 mg/m 2 cisplatin weekly. Induction chemotherapy (IC) consisted of cisplatin combined with taxanes or 5-fluorouracil or both every three weeks for 2 or 3 cycles (Ying et al. 2016). In our study, a total of 213 patients accepted full doses of nimotuzumab therapy (200 mg weekly) after PSM.

Toxicity assessment and follow-up
Chemotherapy toxicities were evaluated by the National Cancer Institute Common Toxicity Criteria (NCI CTCAE, version 4.0), and radiotherapy-related toxicities were assessed according to the Acute and Late Radiation Morbidity Scoring Criteria of RTOG. Dosage was adjusted based on the grade of toxicities individually. Patients were monitored weekly during treatment. The post-treatment clinical followup was generally at intervals of 3 months after treatment for the first 2 years, half-year for the next 3 years, and annually after that. Assessments included patient history, physical and imaging examination of the nasopharynx. The last follow-up date was 31 December 2020.

Study endpoints and statistical analysis
Overall survival (OS) was defined as the time from the date of enrollment to the date of death due to any cause or the date of last follow-up. Progression-free survival (PFS), locoregional recurrence-free survival (LRFS), and distant metastasis-free survival (DMFS) were calculated from the day after the completion of therapy to the first diagnosis of any cause of treatment failure, locoregional nasopharynx/ Fig. 1 Diagram of the analytic cohort for survival analysis. Abbreviations: NPC nasopharyngeal carcinoma; CCRT concurrent chemoradiotherapy; SYSUCC Sun Yat-Sen University Cancer Center; NTZ nimotuzumab cervical lymph nodes relapse, and the distant recurrence/ metastasis, respectively, or the date of last investigation.
The baseline characteristics of patients are described in the CCRT plus NTZ group and CCRT alone group. The receiver operating characteristics (ROC) curve analysis was applied to determine the cut-off point for the GTVnx and GTVnd. In brief, the points of GTVnx were 25 cm 3 for OS, 22 cm 3 for PFS, 22 cm 3 for LRFS, and 25 cm 3 for DMFS. The points of GTVnd were 15cm 3 for OS, 12cm 3 for PFS and LRFS, and 15cm 3 for DMFS. Therefore, a uniform cutoff point of 25 cm 3 and 15 cm 3 were selected to stratify patients into high and low GTVnx and GTVnd groups for analysis, respectively. The χ 2 test was used to compare categorical variables, and the t test was applied to compare continuous variables. Propensity score matching analysis was applied. Using a caliper width of 0.2, 1:1 matching was performed in two groups based on the propensity scores. OS, PFS, LRFS, and DMFS were calculated using the Kaplan-Meier method. Cox analysis was performed to identify the significant prognostic factors. Variables as potential prognostic factors were included to construct a nomogram model. Calibration curves of the nomogram were plotted to compare predicted survival with actual survival. The optimal threshold of the linear prediction score derived from the nomogram was chosen using X-tile version 3.6.1, which categorized patients into low-and high-risk groups (Robert et al. 2004). Furthermore, a Kaplan-Meier curve was generated for each risk group to assess the model's predictive power and clinical application. The statistical analyses were executed using the SPSS 20.0 (Chicago, IL, USA) and R (version 3.2.3). All tests were two-sided, and P < 0.05 was considered statistically significant.

Patient characteristics and follow-up
From January 2015 to December 2017, 3340 NPC patients were enrolled in our retrospective study. Patient characteristics, grouped by the treatment regimen (with or without NTZ), are detailed in Supplemental Table 1. The distributions of BMI, and the serum levels of CRP and LDH, were significantly different between the two groups (P < 0.05) (Supplemental Table 1). To eliminate potential confounding factors, we included a well-balanced cohort via PSM. After PSM, 426 eligible patients were identified for the matched analysis. No significant difference was observed in the clinical characteristics between the two regimens in the PSM cohort (Table 1). For the PSM cohort, the median followup times were 60.0 months (ranging from 3 to 60 months). The follow-up observation indicated that 17 and 4 patients  Fig. 2 showed that patients in the NTZ plus CCRT group had significantly better OS (log rank = 8.394, P = 0.004), PFS (log rank = 9.486, P = 0.002), and LRFS (log rank = 4.810, P = 0.028) than NPC patients underwent CCRT alone (Fig. 2a-c). In addition, patients who received the NTZ plus CCRT treatment had more favorable DMFS than those who received CCRT alone, at the very edge of significance (log rank = 3.741, P = 0.053) (Fig. 2d).
The incidence of acute toxicity in the PSM cohort was evaluated between CCRT with and without NTZ. There were no significant differences between the two regimens in major toxicities such as anemia, thrombocytopenia, leukocytopenia, skin eruption, mucositis, and organ toxicity (all P > 0.05) ( Table 2). In addition, we evaluated the interaction effects between the two treatments and other prognostic factors after  Table 2).

Construction of the nomogram and risk stratification
Potential prognostic factors were utilized to construct a nomogram and specifically predict each patient's prognosis. Prognostic models were developed for OS, PFS, LRFS, and DMFS, respectively (Fig. 5). The four models illustrated good accuracy for predicting the survival rate of NPC patients. When the nomogram models were authenticated internally, with the 1000 bootstrap resampling technique, the concordance indexes (c-index) of 0.874 (95% CI 0.799-0.948), 0.814 (95% CI 0.731-0.897), 0.856 (95% CI 0.746-0.966), and 0.858 (95% CI 0.708-1.000) for OS, PFS, LRFS, and DMFS were attained, respectively. The calibration plots demonstrated a favorable relationship between the predicted and observed survival (Fig. 6). Smaller points in the nomogram indicated a better prognosis. The models indicated that NTZ plus CCRT had a positive impact on the survival of NPC patients.
Each patient's risk score derived from the nomogram was calculated by R (Supplemental Table 3). Patients were divided into high-and low-risk groups by the optimal cut-off value determined by X-tile software: a low-risk group (< 33 points, n = 319) and a high-risk group (≥ 33 points, n = 107) for OS; a low-risk group (< 35 points, n = 400) and a highrisk group (≥ 35 points, n = 26) for PFS; and low-risk groups (< 24 points, n = 277) and high-risk groups (≥ 24 points, n = 149) for both LRFS and DMFS. The Kaplan-Meier curve was generated to evaluate the survival rate in different risk groups. Patients in the low-risk group had better 5 year overall survival than those in the high-risk group (98.7% [95% CI 98.6-98.8] vs. 82.4% [95% CI 79.3-85.6]) (Fig. 7a).
The 5 year PFS was better in the low-risk group than in highrisk group ( (Fig. 7c, d). These results suggested that higher total points of nomogram were related to worse survival rate, and conversely, lower total points of the model were associated with a better outcome.

Discussion
The current study evaluated the therapeutic effect of nimotuzumab combined with concurrent chemoradiotherapy in stage III-IVa nasopharyngeal carcinoma patients. It demonstrated that this combination achieved a better prognosis than CCRT alone, with similar toxicity rates between the two regimens. Furthermore, we constructed nomogram models to assess the survival probability according to the clinical characteristics of individual patients.
Although nasopharyngeal tumors are sensitive to radiotherapy, some patients still develop recurrence and metastases after treatment. EGFR is over-expressed in most NPC patients, and its high expression level indicates substantial invasion and drug resistance (Fortunato and Giampaolo 2001). Several reports have examined the efficacy of NTZ treatment for NPC patients, but the conclusions are discrepant (Zhi-Gang et al. 2016;Jian-Feng et al. 2017;Fangzheng et al. 2018aFangzheng et al. , 2018bJi-Jin et al. 2018;Zhi-Qiang et al. 2019). For example, Li et al. conducted a retrospective study and they found that NPC patients receiving cisplatin and radiotherapy achieved better survival than those receiving NTZ combined with radiotherapy, while NTZ plus radiotherapy caused less toxicity. They suggested that NTZ plus radiotherapy is an optional treatment for stage II or older NPC patients . However, NPC patients in stage II rarely receive anti-EGFR drug treatment in our cancer center. Therefore, only stage III-IVa NPC patients are discussed in the current study, and the therapeutic effect of the NTZ combination for stage II NPC patients remains elusive.  year survival based on a patient's combination of covariates. For example, the patient's treatment regimen is identified, and a line is drawn upward to the "Points" axis to determine the score associated with that nimotuzumab. The process is repeated for each variable, and the scores are summed and decided on the "Total Points" axis. A line is drawn straight down to determine the likelihood of 1-, 3-and 5 year survival. EBV DNA Epstein-Barr virus DNA; OS overall survival; PFS progression-free survival; LRFS locoregional recurrence-free survival; DMFS distant metastasis-free survival concurrently with IMRT compared to those in the standard cisplatin-IMRT combination. Furthermore, the subgroup analysis revealed frustrating locoregional control in N3 stage patients treated with CTX/NTZ. Additionally, G3-G4 hematologic toxicities and an increased rate of CTX-related mucositis were observed in the CTX/NTZ group (Rui et al. 2017a). You et al. then compared the efficacy of CTX or NTZ and CCRT, and CCRT treatment alone, in stage II-IVa NPC patients. They suggested that the combination of CTX/ NTZ with CCRT yielded better OS in patients with stage II-IVa NPC compared with CCRT alone. In addition, the positive effect of the combined CTX/NTZ treatment was also evident in DFS and DMFS (Rui et al. 2017b). Our data are partially consistent with these previous findings. In the current study, the combination of NTZ and CCRT improved OS, PFS, and LRFS for stage III-IVa NPC patients without increasing the incidence of acute toxicity. However, DMFS did not benefit from the addition of NTZ treatment (Figs. 2 and 3). We hypothesize that the inconsistent results might be due to a discrepancy in patient enrollment. Specifically, the analysis by You et al. included patients between January 2009 and December 2013, with 189 patients in the CTX/ NTZ plus CCRT arm. In addition, the molecularly targeted drug was CTX or NTZ, not specifically NTZ, and eligible patients received only one treatment cycle. In the current study, we enrolled NPC patients from 2015 to 2017 and focused on the efficacy of NTZ specifically. Patients receiving at least six doses of NTZ were included, and 213 eligible patients were assigned to each cohort using the PSM method. In addition, Sun et al. conducted a study to investigate the efficacy of anti-EGFR drugs added to palliative chemotherapy in patients with de novo metastatic NPC. They reported that de novo metastatic NPC patients might not benefit from anti-EGFR drug treatment, and adding CTX to CCRT might exacerbate the acute mucositis and skin reactions (Xue-Song et al. 2019). Collectively, the efficacy of anti-EGFR drugs in suppressing distant metastasis remains to be determined.
NTZ is a recombinant humanized IgG1 monoclonal antibody that binds to EGFR directly (Cristina et al. 1997). Therefore, NTZ can inhibit the binding of EGF and transforming growth factor alpha to the EGFR and thereby block the EGFR signaling pathway. In addition, NTZ is reported to induce apoptosis and inhibit the cell cycle and angiogenesis in tumors (Crombet-Ramos et al. 2002). Furthermore, NTZ has high safety and low toxicity with a relatively low risk of skin and mucosa toxicities commonly caused by other EGFR-targeting antibodies (Mei et al. 2018;Zhi-Qiang et al. 2019). These findings suggest that NTZ has advantages over other EGFR inhibitors, such as CTX, in less toxicity (Mei et al. 2018). In addition, the affinity constant of NTZ is low, which leads to selective tumor uptake and low normal tissue absorption.
Previous clinical trials that studied the efficacy of NTZ combined with radiotherapy in patients with locally advanced head, and neck squamous cell carcinoma reported that the addition of NTZ caused mild toxicity and may improve the radiation sensitivity of unresectable head and neck tumors (Talavera et al. 2009). Randomized studies that enrolled Indian patients with advanced head and neck squamous cell carcinoma concluded that concurrent use of NTZ with radiotherapy or chemoradiotherapy was safe and provided long-term survival benefits (Reddy et al. 2014;Patil et al. 2019). In NPC, CCRT significantly increased radiotherapy-and chemotherapy-related adverse events such as mucositis, skin reactions, and organ toxicity (Lei et al. 2012). Therefore, it is essential to clarify the negative side effects of combination treatments with NTZ. Wang et al. concluded that long-term usage of NTZ did not increase the incidence of radiation-related toxicities; no skin rashes or infusion reactions were observed in the treated NPC patients (Fangzheng et al. 2018a). Consistent with this finding, Wang et al. conducted a 1:3 PSM analysis in III-IVa stage NPC patients who received IMRT with or without NTZ. No significant differences were observed between the two regimens in acute toxicities (Zhi-Qiang et al. 2019). We made similar observations in the current analysis (Table 2), demonstrating that NTZ plus CCRT may be a safe therapy regimen for locally advanced NPC patients.
The nomogram models showed good predictive accuracy and discriminative ability. Some studies have established a nomogram model for predicting NPC survival (Lin-Quan et al. 2015;Wang-Zhong et al. 2021;Jianpei et al. 2018). In the present study, we included potential prognostic biomarkers in the nomogram models to predict the survival rate. The models suggested that the combination of NTZ and CCRT treatment predicted a better prognosis for stage III-IVa NPC patients. We divided the patients into low-and high-risk groups based on the total points, and the differences between the two groups were statistically significant for OS, PFS, LRFS, and DMFS. Therefore, the above results indicated that our nomogram model was a reliable and precise tool to predict prognosis in NPC patients.
Although our analysis verified the positive efficacy of the NTZ plus CCRT regimen for local advanced NPC patients, and the nomogram models provided valuable tools for clinical decision-making, the current study still had several limitations. Firstly, as a retrospective study, potential selection bias was unavoidable, even though the bias was minimized by recruiting consecutive NPC patients and using the PSM analysis. Secondly, information on the EGFR expression level is missing; therefore, the relationship between the EGFR expression level and the efficacy of NTZ is unclear. Finally, the study population was enrolled from a single center. Thus, external validation, including a prospective and randomized study, is needed to verify our findings.

Conclusion
The present study suggested that the combination of NTZ with CCRT resulted in a favorable clinical OS outcome for stage III-IVa NPC patients with good tolerance and similar toxicities compared to CCRT alone. In addition, we established applicable nomogram models to evaluate the benefit of CCRT in combination with NTZ for individual NPC patients, which could help personalized treatment.