Pretreatment quality of life as a predictor of survival for patients with nasopharyngeal carcinoma treated with IMRT

Background To evaluate the prognostic significance of pretreatment quality of life for patients with nasopharyngeal carcinoma treated with intensity-modulated radiotherapy. Methods We performed a prospective, longitudinal study on 554 newly diagnosed patients with NPC from April 2011 to January 2015. A total of 501 consecutive NPC patients were included. Patients were asked to complete the EORTC QLQ-C30 (version 3.0) and QLQ-H&N35 questionnaires before treatment. Results Global health status among QLQ-C30 correlates with EBV DNA(P = 0.019). In addition, pretreatment appetite loss was significantly correlated with EBV DNA(P = 0.02). Pretreatment teeth, opening mouth, feeding tube was significantly correlated with EBV DNA, with P value of 0.003, < 0.0001, and 0.031, respectively. In multivariate analysis, pretreatment cognitive functioning of QLQ-C30 was significantly associated with LRFS, with HR of 0.971(95%CI 0.951–0.990), P = 0.004. Among scales of QLQ-H&N35 for multivariate analysis, pretreatment teeth (P = 0.026) and felt ill (P = 0.012) was significantly associated with PFS, with HR of 0.984 (95%CI 0.971–.998) and 1.004 (95%CI 1.001–1.007), respectively. Felt ill of QLQ-H&N35 was significantly associated with DMFS, with HR of 1.004(95%CI 1.000–1.007), P = 0.043. There is no QoL scale significantly associated with OS after multivariate analysis. Conclusions In conclusion, our analysis confirms that pretreatment teeth and felt ill was significantly associated with PFS in NPC patients treated with IMRT. In addition, the posttreatment EBV DNA was significantly associated with OS.


Background
Nasopharyngeal carcinoma (NPC) is prevalent in Southern China and Southeast Asia, but rare in the Western world. The annual incidence of NPC is 15-50 cases per 100,000 [1]. NPC differs from other head and neck cancers in its epidemiology, association with Epstein-Barr virus (EBV), and high risk of distant metastasis [2]. Radiotherapy (RT) is the primary treatment for nonmetastatic disease [3,4]. Intensity modulated radiation therapy (IMRT) is the most frequently recommended radiation method, if conditions permit, because of excellent local control. Concurrent chemoradiotherapy (CCRT) is recommended as a first line therapy for locally advanced NPC [5,6]. Induction chemotherapy has been combined in several studies to improve clinical outcomes, but it remains controversial [7][8][9]. Distant metastasis is the major cause of mortality in NPC patients.
Quality of life (QoL) has been considered to be a prognostic factor for cancer patients, such as for head and neck cancer [10,11], hepatocellular carcinoma and cholangiocarcinoma [12], colorectal cancer [13], liver cancer [14] and lung cancer [15]. Few studies have explored the prognostic significance of pretreatment QoL in NPC [16,17]. Therefore, we conducted a prospective study using two self-administered questionnaires, the European Oganization for Research and Treatment of Cancer (EORTC) Quality of Life Questionnaire C30 (QLQ-C30) and the EORTC QLQ Head and Neck Cancer-Specific Module (H&N35), to assess the pretreatment QoL scores [18]. We assumed that felt ill among the H&N35 questionaire was significantly associated with PFS.

Patients
We performed a prospective, longitudinal study on 554 newly diagnosed patients with NPC in the Sun Yat-Sen University Cancer Center from April 2011 to January 2015. A total of 501 consecutive NPC patients were included in this study. This study was approved by the clinical research ethics committee of the Sun Yat-Sen University Cancer Center, and the participants provided written informed consent. Patients with the following characteristics were excluded: those with distant metastasis at initial diagnosis (n = 10), those lost to follow-up posttreatment (n = 2), those whose treatment was interrupted (n = 1), those who were unable to complete the questionnaire pretreatment (n = 1), those who were unable to complete the questionnaire posttreatment (n = 3), those who were unable to complete the questionnaire three months posttreatment (n = 3), those who did not test for EAIgA and VCAIgA before treatment (n = 10), those who did not test for EBV DNA before treatment (n = 17), and those who did not test for EBV DNA value posttreatment (n = 7). All patients were given a complete physical examination, a fiber-optic nasopharyngoscopy, magnetic resonance imaging (MRI) of the head and neck, chest radiography, abdominal sonography, electrocardiography, bone scan or PET/CT, complete blood count with a differential count, biochemical profile, and Epstein-Barr virus serology.

QoL assessments
The self-administered EORTC QLQ-C30 (version 3.0) and the QLQ-H&N35 questionnaires were prospectively given to the enrolled patients [18][19][20]. The questionnaires are used by a large number of research groups in cancer clinical trials and have also been used in various other, non-trial studies. The Taiwan Chinese version was available and easily completed by our patients. Patients were asked to complete the Chinese version of the EORTC QLQ-C30 (version 3.0) and QLQ-H&N35 questionnaires before treatment. The QLQ-C30 contains 15 scales: five functional scales (physical, role, emotional, cognitive, and social functioning), three symptom scales (fatigue, nausea and vomiting, pain), six single-item symptom scales (dyspnea, insomnia, appetite loss, constipation, diarrhea, financial difficulties), and one global health status/QoL scale. The QLQ-H&N35 is meant for use among head and neck cancer patients with varying disease stages and treatment modalities. The QLQ-H&N35 is composed of seven multi-item symptom scales (pain, swallowing, sensation, speech, eating from a social perspective, social interactions, and sexuality) and 11 single-item symptom scales (teeth, opening mouth, dry mouth, sticky saliva, coughing, felt ill, pain medication use, nutritional supplementation, feeding tube requirement, weight loss, and weight gain). All of the scales and items ranged in score from 0 to 100. A high score for a functional or global QoL scale represents a relatively high/healthy level of functional or global QoL, whereas a high score for a symptom scale or item represents a high number of symptoms or problems.

Study treatments RT techniques
All patients (501 patients) were treated with IMRT. The dose fractionation and total dose of IMRT for NPC patients followed the guidelines of our institute [21,22], which are in accordance with the International Commission on Radiation Units and Measurements reports 50 and 62. All the target volumes were depicted slice-byslice on the treatment planning computed tomography scan. The primary nasopharyngeal gross tumor volume (GTVnx) and the involved cervical lymph nodes were determined based on imaging, clinical, and endoscopic findings. The enlarged retropharyngeal nodes together with primary gross tumor volume (GTV) were outlined as the GTVnx on the IMRT plans. The first clinical tumor volume (CTV1) was defined as the area from 0.5-1.0 cm outside the GTV, a site that involves potential sites of local infiltration. The clinical target volume 2 (CTV2) was defined as the margin from 0.5-1.0 cm around CTV1 and the lymph node draining area (Levels II, III, and IV). For stage N1-3 patients, the lower neck area received conventional anterior cervical field radiation with a midline shield to 50 Gy in daily fractions of 2 Gy. For patients with stage N0 disease, RT was not delivered to the lower neck area. The prescribed 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 30 to 33 fractions. In total, 30-33 fractions were administered at 1 fraction per day, 5 days per week.

Chemotherapy
Patients with clinical stage I were treated with RT alone. Patients with stage II-IVa were treated with CCRT or induction chemotherapy+CCRT. A total of 249 (49.7%) patients received induction chemotherapy followed by CCRT, the regimen of induction chemotherapy regimens were various regimens of based on cisplatin. Overall, 214 (42.7%) patients received concomitant chemotherapy with cisplatin. Of the 214 patients treated with concomitant chemotherapy of cisplatin regimen, a total of 37 patients received cumulative cisplatin dose of < 100 mg/m 2 , 123 patients received cumulative cisplatin dose of 101-200 mg/m 2 and 54 patients received cumulative cisplatin dose of 200-300 mg/m 2 . A total of 38 patients (7.6%) were treated with RT alone.

Follow-up and study endpoints
Patients were followed up every 3 months throughout the first 3 years, every 6 months for the next 2 years and annually thereafter. Physical examinations, nasopharyngoscopic examinations, MRIs, chest X-rays, abdominal ultrasounds and EBV DNA tests were performed at each follow-up visit. The follow-up duration was calculated from the first day of treatment to either the day of death or the day of the last examination. The median followup duration was 32 months (6-57 months). The primary end point of this study was progression free survival (PFS), and the secondary end points were overall survival (OS), local recurrence-free survival (LRFS) and distant free survival (DMFS). PFS was defined as the time from treatment of NPC to events that included death or disease progression at local, regional, or distant sites or until the date of the last follow-up. OS was defined as the time from treatment of NPC to the date of death or until the date of the last follow-up. LRFS was defined as the time from treatment of NPC to the absence of a primary site or neck lymph node relapse or until the date of the last follow-up. DMFS was defined as the time from treatment of NPC to the date of the first observation of distant metastases or until the date of the last follow-up. The last follow-up date was February 6, 2016.

Statistical methods
All analyses were performed using SPSS version 18.0 (version 18.0; SPSS Inc., Chicago, III). All tests were 2tailed. The correlation between EBV DNA and QoL scale was analyzed by Spearman's correlation . Univariate analysis measured by the Cox proportional hazards regression model was used to calculate the P value of each QoL scale from QLQ-C30 and H&N35. When the P value of the QoL scale in univariate analysis was less than 0.05, the scale was separately calculated by multivariate analysis adjusted for age (< 45 vs. ≥ 45), gender (male vs. female), marriage (yes vs. no), education (<high school vs. ≥high school), smoking history (yes vs. no), alcohol history (yes vs. no), T stage (T1,2 vs. T3,4), N stage (N1,2 vs. N3,4), pre-treatment EBV DNA (< 4000 vs. ≥4000) and post-treatment EBV DNA (negative vs. positive).

Patient characteristics
In this study population, there were 380 male patients and 121 female patients, with a male: female ratio of 3.14:1. The median age was 44 years (range, 11-72 years). There were 498 (99.4%) of the 501 patients had World Health Organization (WHO) type II or III disease, and 3 (0.6%) had WHO type I disease. There were 9 (1.8%) patients with American Joint of Cancer Committee (AJCC) stage I; 50(10.0%) patients with stage II, 281 (56.1%) patients with stage III, 161 (32.1%) patients with stage IV. A total of 496 (99.0%) patients had an Eastern Cooperative Oncology Group (ECOG) score of 1. More than half of the patients (337, 67.3%) had a history of smoking, and the use of alcohol was not common (53, 10.6%). We represented the characteristics divided by sex in Table 1.

Correlation between EBV DNA and QoL
We analyzed correlation between each scale among the QLQ-C30 questionnaire and pretreatment EBV DNA, found that global health status correlates with pretreatment EBV DNA(P = 0.019). In addition, pretreatment appetite loss was significantly correlated with pretreatment EBV DNA(P = 0.02). We also analyzed the correlation between each scale among the QLQ-H&N35 questionnaire and pretreatment EBV DNA. We found that pretreatment teeth, opening mouth, feeding tube was significantly correlated with pretreatment EBV DNA, with P value of 0.003, < 0.0001, and 0.031, respectively. Appendix: Tables 7  and 8 represented the correlation between EBV DNA and QLQ-C30 or QLQ-H&N35.

Univariate analysis pretreatment
In QLQ-C30, there was no functional scale or symptom scale that was significantly associated with OS, PFS and DMFS in QLQ-C30 pretreatment. Only pretreatment cognitive functioning was significantly associated with LRFS in QLQ-C30 (Fig. 1).
In QLQ-H&N35, were pain and swallowing significantly associated with OS. There were three scales significantly associated with PFS: pain, teeth (Fig. 2) and felt ill (Fig. 3). There were six scales in QLQ-H&N35 that were significantly associated with LRFS: pain, swallowing, speech, social eating and teeth. There were two scales in QLQ-H&N35 that were significantly associated with DMFS: pain and felt ill (Fig. 4). (Appendix: Table 7).

Discussion
There have been previous studies regarding quality of life on NPC patients and head and neck cancer. Until now, only one study had explored the prognostic significance of QoL in QLQ-C30 questionnaires by assessing  [17]. To our knowledge, this is the first large scale study of NPC patients in the IMRT era that prospectively explored functional scales and symptom scales in both QLQ-30 and H&N35.
We found that global health status significantly correlates with EBV DNA. High pretreatment EBV DNA level always associates with large tumor or multiple lymph nodes which represents advanced stage. Patients with advanced stage represents poor quality of life scores. This may be the possible explanation for global health status significantly correlates with EBV DNA. In addition, pretreatment appetite loss was significantly correlated with EBV DNA. We found that pretreatment teeth, opening mouth, feeding tube was significantly correlated with EBV DNA. This is the first time that the correlation between quality of life and EBV DNA is reported. The exact mechanism remains unknown. More studies about the correlation between quality of life and EBV DNA is expected to do in the future.    In the present study, pretreatment teeth in QLQ-H&N35 predicted longer PFS. This result may be explained by a sensitivity to radiotherapy resulting in uncomfortable sensation in the teeth. The exact mechanism is unknown. Interestingly, felt ill pretreatment in QLQ-H&N35 predicted shorter DMFS in multivariate analysis. The possible explanation would be as follows. At the beginning of treatment, pain mostly comes from large tumor region, probably because of invasion along the cranial nerve. Large tumors of head and neck cancers or NPC are significantly associated with distant metastasis. A previous study found that pretreatment pain influences OS in 2340 newly diagnosed patients with head and neck squamous cancer [23]. We found that a high cognitive functioning score pretreatment in QLQ-C30 predicted longer LRFS. This finding is consistent with previous studies in head and neck cancer [24] and NPC [17]. The exact mechanism of why cognitive function correlates with survival is unknown. The causative relationship between cognitive functioning and survival is indeterminate. Cognitive functioning might be a surrogate for the QoL scales that were potentially prognostic, and we speculate that it may display as a physiological appearance for some undetected predictive factors.
In this study, post treatment EBV DNA predicted OS better than pretreatment EBV DNA. Using multivariate analysis, posttreatment EBV DNA significantly predicted OS for NPC patients in this study. Pretreatment EBV DNA did not show predict value of OS in this study in multivariate analysis, revealed that the prognostic value of pretreatment EBV DNA was covered up by posttreatment EBV DNA in this study. This finding is consistent with previous studies. A recent study explored EBV DNA loading of 273 NPC patients at different time points and found that post treatment EBV DNA was significantly associated with PFS, DMFS and OS [25]. Several studies in Taiwan concluded that post treatment EBV DNA was an important independent prognostic factor for clinical outcomes [26,27].
Our results revealed that QoL and post treatment EBV DNA can effectively predict survival for NPC patients. The results provide a promising way to guide treatment strategy for NPC patients. Our study has several strengths. First, the present study has the longest longitudinal collection of QoL data that has been used to examine prognostic value during the initial management of patients with NPC. Second, this is the first time that QoL scores in QLQ-H&N35 were found to predict survival for NPC patients. Our study evaluated the prognostic significance of QoL using both the QLQ-C30 questionnaire and QLQ-H&N35 questionnaire.
There were some limitations in the present study. First, this is a single center study in a high incidence area in Southern China. Future studies are needed to calculate the prognostic significance of QoL in NPC patients in other areas in the world. Second, the median follow-up time of this study was 32 months; a longer follow-up time is needed to further validate our results.

Conclusions
In conclusion, our analysis confirms that pretreatment teeth and felt ill was significantly associated with PFS in NPC patients treated with IMRT. In addition, the posttreatment EBV DNA was significantly associated with OS.