Clinical outcomes of residual or recurrent nasopharyngeal carcinoma treated with endoscopic nasopharyngectomy plus chemoradiotherapy or with chemoradiotherapy alone: a retrospective study

Study criteria and patient characteristics

Data for the patients with residual or recurrent NPC were retrieved from the database of The People’s Hospital of Guangxi Zhuang Autonomous Region (Nanning, China). All the patients recruited in this study provided written informed authorization consent, in which they agreed to use of their clinical and imaging data for non-commercial scientific research. We adhered to the bioethics principles of the Declaration of Helsinki. This retrospective study was approved by the ethics committee of The People’s Hospital of Guangxi Zhuang Autonomous Region (Ethical Application Ref: Keyan-Guangxi-Keji-2016-31).

The case inclusion criteria included (1) patients admitted to The People’s Hospital of Guangxi Zhuang Autonomous Region with residual or recurrent NPC. The diagnoses of residual or recurrent NPC for every case included in our study were strictly confirmed by pathological examination. The staging for each patient was based upon the T classification of residual or recurrent tumors. A residual tumor was defined as a persistent lesion in the nasopharynx within six months after full-dose radiotherapy. A recurrent tumor was defined as a local recurrence after complete remission within six months after full-dose radiotherapy and the subsequent emergence of new lesions in the nasopharynx after six months. Additional inclusion criteria were (2) a measurable lesion with evaluable efficacy, (3) a Karnofsky score ≥70, and (4) provision of written informed consent to undergo surgery and/or chemoradiotherapy. The case exclusion criteria were (1) patients with cervical lymph node metastasis or distant metastasis, (2) patients with severe damage to liver and kidney function, and (3) patients with severe immune deficiency.

A total of 62 patients with residual or recurrent NPC who were treated at our hospital between June 2011 and April 2013 were enrolled in the study (five cases of residual tumors and 57 cases of recurrent tumors). The patients included 47 men and 15 women, aged 22 to 70 years with a median age of 50 years. Of these 62 patients, 58 patients were pathologically identified as having undifferentiated non-keratinizing carcinoma (WHO type III), three patients had keratinizing squamous cell carcinoma (WHO type I), and only one patient had differentiated non-keratinizing carcinoma (WHO type II).

In general, surgery was recommended and preferred for all the regional residual or recurrent cases, except when the following exclusion criteria were encountered: (1) internal carotid artery encasement, massive intracranial intradural involvement, or orbital space invasion; (2) uncontrolled nasopharyngeal skull infection; (3) general anesthesia surgery contraindications; or (4) patient refusal to undergo surgery. To retrospectively analyze the clinical outcome of the patients who underwent endoscopic nasopharyngectomy, the patients were divided into two groups according to the treatment selection. In the surgical group, 36 patients underwent complete endoscopic nasopharyngectomy of the residual or recurrent tumor and subsequent post-operative chemoradiotherapy (CRT). In the CRT alone group, 26 patients received concurrent radiotherapy and chemotherapy. The patients’ sex, age, local staging, and EBV-DNA positive rate showed no statistically significant differences between the two groups (all P > 0.05) (Table 1).

Treatment methods

The surgical group received endoscopic nasopharyngectomy plus concurrent radiotherapy and a chemotherapy treatment regimen. Surgery was performed under general anesthesia. The patients were placed in a supine position, and a cotton swab soaked with epinephrine was used to shrink the nasal mucosa 2–3 times. Under the guidance of nasal endoscopy, an electric knife was then used to resect the nasopharyngeal tumor via a nasal approach. The resection margin was defined according to the results of nasopharyngeal skull-base magnetic resonance imaging (MRI) and intraoperative observations. Typical imaging data of the patients are shown in Fig. 1. According to post-operative pathological examinations, all the resections were confirmed to be residual or recurrent tumors and three patients in the surgery group (8.3%, 3/36) were found to have positive margins.

The chemotherapy regimen and doses for the surgical + CRT and CRT alone group were the same and consisted of 30 mg/m² cisplatin via intravenous infusion beginning on days 1–3 and 1,000 mg/m² gemcitabine via intravenous infusion on day 1 and day 8; each cycle lasted 21 days, and 2–4 cycles were administered consecutively. For the intensity-modulated radiotherapy (IMRT) for both groups, the target area was delineated on the basis of the tumor boundary shown in MRI, where in the gross tumor volume (GTV) was irradiated with 2.0 Gy/fraction 5days/week; the total dose of GTV was 60–66 Gy. The median radiotherapy dose for primary NPC was 71.2 (52–81) Gy. The median time span between the primary treatment and treatment of the recurrence was 20 (3–178) months.

Collection of peripheral blood samples and EBV-DNA load assay

A 2 mL blood sample was collected from each patient’s antecubital vein pre- and post-treatment in an anticoagulation tube containing EDTA, centrifuged to separate the blood plasma into the upper phase for DNA extraction, and stored at −20 °C for later use. A Light-Cycler 480 quantitative fluorescence PCR platform (Roche Diagnostics, Basel, Switzerland) was used to quantify EBV-DNA levels. The test kit was provided by the Da’an Gene Diagnostic Center of Sun Yat-Sen University (Cat. # DA-D065). The target gene for amplification was derived from the BamHI-W fragment of EBV. All procedures were performed according to the manufacturer’s protocol. EBV-DNA >0 copies/mL was considered EBV-positive. EBV-DNA(+) refers to the pretreatment EBV-DNA-positive level unless otherwise stated.

Follow-up and outcome measures

The patients began regular follow-up examinations after the treatment ended. During the first year after treatment, the patients returned to the hospital once every three months for examination as inpatients or outpatients. After 1 year, they received follow-up examinations once every six months, and after three years, they received follow-up examinations once every 12 months. The date of the final follow-up visit was 30 July 2016; the median follow-up time was 31 months, and the follow-up rate was 100%.

Statistical analysis

The SPSS 18.0 statistical software package (SPSS Inc., Chicago, IL, USA) was used to establish the database. Between-group comparisons of the count data were performed using the χ² test. Pre- and post-treatment comparisons of EBV-DNA levels were performed using the Wilcoxon rank-sum test. Comparisons of EBV-DNA levels among different T-stages were performed using the Kruskal–Wallis H test. Kaplan–Meier survival curves were used for univariate survival analysis, and the log-rank test was performed for the between-group comparison of survival curves. Multivariate analysis was performed using Cox regression analysis. P < 0.05 indicated that the differences were statistically significant.

Association of EBV-DNA with the clinical stages and changes in EBV-DNA pre- and post-treatment

The EBV-DNA positive rate in the serum of all patients was 43.55%. The mean EBV-DNA copy numbers of patients with stages rT1, rT2, rT3, and rT4 disease were 510, 583, 956, and 3,909 copies/mL, respectively. The EBV-DNA levels increased with NPC stages (χ² value=11.674, P = 0.009) (Fig. 2A). The EBV-DNA levels decreased significantly in the surgery + post-operative CRT group (Z =  − 3.484, P < 0.001); compared with the pretreatment levels, the post-treatment EBV-DNA levels in the CRT group did not decrease significantly (Z =  − 1.956, P = 0.051) (Fig. 2B).

Prognosis analysis

The 3-year overall survival (OS) for all patients was 51.40%, and the 3-year disease-free survival (DFS) was 46.86%. The univariate analysis indicated that local staging, treatment method, and pre- and post-treatment EBV-DNA levels correlated with the OS and DFS (all P < 0.05) (Table 2). Kaplan–Meier survival analysis indicated that the prognosis of patients with local early-stage NPC (T1 + T2) was significantly better than that of patients with local late-stage NPC (T3 + T4) (Figs. 3A and 4A). The prognosis of the surgery + CRT group was superior to that of the CRT alone group (Figs. 3B and 4B). Pre- and post-treatment EBV-DNA levels in the peripheral blood were associated with the clinical prognosis (Figs. 3C, 3D, 4C and 4D). Notably, the patients with post-treatment EBV-positive disease were found to had obviously worse outcomes (Figs. 3D and 4D), which indicated that post-treatment EBV-positive strongly suggests a poor prognosis in patients with residual or recurrent NPC. Multivariate analysis indicated that local staging, pretreatment EBV-DNA levels, and the treatment method were independent risk factors for OS, whereas local staging and the treatment method were independent risk factors for DFS (Table 3). By the end of our follow-up, a total of 29 patients had died, of whom17 died of local recurrence, seven died of distant metastasis, three died of internal carotid artery rupture caused by skull necrosis, and two died of cervical recurrence.

Subgroup analysis

The patients were divided into subgroups on the basis of local staging and the treatment method. The four subgroups were the local early-stage (T1 + T2) and surgery + CRT subgroup (17 cases), the local early-stage (T1 + T2) and CRT alone subgroup (12 cases), the local late-stage (T3 + T4) and surgery + CRT subgroup (19 cases), and the local late-stage (T3 + T4) and CRT alone subgroup (14 cases). The results indicated that for the local early-stage subgroups, the OS and DFS did not show significant differences among the different treatment methods (all P > 0.05) (Figs. 5A and 5B). For the local late-stage subgroups, surgery + CRT showed a better DFS than CRT alone (P = 0.032), whereas OS did not show a significant difference (P > 0.05) (Figs. 5C and 5D).

The patients were also divided into subgroups on the basis of the presence of EBV-DNA and the treatment method. The four subgroups were the EBV-DNA negative and surgery + CRT subgroup (19 cases), the EBV-DNA negative and CRT alone subgroup (16 cases), the EBV-DNA positive and surgery + CRT subgroup (17 cases), and the EBV-DNA positive and CRT alone subgroup (10 cases). The results indicated that the EBV-DNA negative and surgery + CRT subgroup had better OS and DFS than the CRT alone subgroup (OS: P = 0.009, DFS: P = 0.003) (Figs. 6A and 6B). Among the patients who tested positive for EBV-DNA, the OS and DFS did not show significant differences between the two treatment regimens (all P > 0.05) (Figs. 6C and 6D).

Complications from surgery and chemoradiotherapy

In the surgical group, the most common complications from endoscopic surgery and/or radiotherapy were secretory otitis media (47.2%, 17/36), temporal lobe necrosis (25.0%, 9/36), skull base necrosis (19.4%, 7/36), cranial nerve palsy (13.9%, 5/36) and nasopharyngeal hemorrhage (5.6%, 2/36). In the non-surgical (CRT alone) group, the common complications from radiotherapy were temporal lobe necrosis (53.9%, 14/26), secretory otitis media (46.2%, 12/26), cranial nerve palsy (23.1%, 6/26), and skull base necrosis (15.4%, 4/26). The most common side-effects from chemotherapy in the both group were hematological toxicities, including leukopenia, neutropenia, and thrombocytopenia. Other non-hematological toxicities included mild transaminase elevation and tolerable nausea and vomiting. In general, the chemotherapy-related side effects were tolerable and could be well-controlled.