Combined prognostic value of pretreatment anemia and cervical node necrosis in patients with nasopharyngeal carcinoma receiving intensity‐modulated radiotherapy: A large‐scale retrospective study

Abstract This study investigated the combined prognostic value of pretreatment anemia and cervical node necrosis (CNN) in patients with nasopharyngeal carcinoma (NPC). Retrospective review of 1302 patients with newly diagnosed nonmetastatic NPC treated with intensity‐modulated radiotherapy (IMRT) ± chemotherapy. Patients were classified into four groups according to anemia and CNN status. Survival was compared using the log‐rank test. Independent prognostic factors were identified using the Cox proportional hazards model. The primary end‐point was overall survival (OS); secondary end‐points were disease‐free survival (DFS), locoregional relapse‐free survival (LRRFS), and distant metastasis‐free survival (DMFS). Pretreatment anemia was an independent, adverse prognostic factor for DMFS; pretreatment CNN was an independent adverse prognostic factor for all end‐points. Five‐year survival for non‐anemia and non‐CNN, anemia, CNN, and anemia and CNN groups were: OS (93.1%, 87.2%, 82.9%, 76.3%, P < 0.001), DFS (87.0%, 84.0%, 73.9%, 64.6%, P < 0.001), DMFS (94.1%, 92.1%, 82.4%, 72.5%, P < 0.001), and LRRFS (92.8%, 92.4%, 88.7%, 84.0%, P = 0.012). The non‐anemia and non‐CNN group had best survival outcomes; anemia and CNN group, the poorest. Multivariate analysis demonstrated combined anemia and CNN was an independent prognostic factor for OS, DFS, DMFS, and LRRFS (P < 0.05). The combination of anemia and CNN is an independent adverse prognostic factor in patients with NPC treated using IMRT ± chemotherapy. Assessment of pretreatment anemia and CNN improved risk stratification, especially for patients with anemia and CNN who have poorest prognosis. This study may aid the design of individualized treatment plans to improve treatment outcomes.


Abstract
This study investigated the combined prognostic value of pretreatment anemia and cervical node necrosis (CNN) in patients with nasopharyngeal carcinoma (NPC). Retrospective review of 1302 patients with newly diagnosed nonmetastatic NPC treated with intensity-modulated radiotherapy (IMRT) ± chemotherapy. Patients were classified into four groups according to anemia and CNN status. Survival was compared using the log-rank test. Independent prognostic factors were identified using the Cox proportional hazards model. The primary endpoint was overall survival (OS); secondary end-points were disease-free survival (DFS), locoregional relapse-free survival (LRRFS), and distant metastasis-free survival (DMFS). Pretreatment anemia was an independent, adverse prognostic factor for DMFS; pretreatment CNN was an independent adverse prognostic factor for all end-points. Five-year survival for non-anemia and non-CNN, anemia, CNN, and anemia and CNN groups were: OS (93.1%, 87.2%, 82.9%, 76.3%, P < 0.001), DFS (87.0%, 84.0%, 73.9%, 64.6%, P < 0.001), DMFS (94.1%, 92.1%, 82.4%, 72.5%, P < 0.001), and LRRFS (92.8%, 92.4%, 88.7%, 84.0%, P = 0.012). The non-anemia and non-CNN group had best survival outcomes; anemia and CNN group, the poorest. Multivariate analysis demonstrated combined anemia and CNN was an independent prognostic factor for OS, DFS, DMFS, and LRRFS (P < 0.05). The combination of anemia and CNN is an independent adverse prognostic factor in patients with NPC treated using IMRT ± chemotherapy. Assessment of pretreatment anemia and CNN improved risk stratification, especially for patients with anemia and CNN who have poorest prognosis. This study may aid the design of individualized treatment plans to improve treatment outcomes.

Introduction
Nasopharyngeal carcinoma (NPC) is a malignancy of the nasopharyngeal epithelium. Although NPC is rare globally, it is highly prevalent in southern China where the incidence is 20 to 50 cases per 100,000 males [1]. Radiotherapy (RT) is the primary therapy for nonmetastatic NPC [2,3]. The tumor-node-metastasis (TNM) staging system for NPC is the most powerful tool for guiding the selection of treatment strategies and predicting the prognosis of patients with NPC [4]. Although application of superior RT technique intensity-modulated radiation therapy (IMRT) and addition of chemotherapy and the precise imaging technology magnetic resonance imaging (MRI) have improved locoregional control, the survival outcomes among patients with advanced NPC remain unsatisfactory [5][6][7]. Therefore, the identification of clinically relevant prognostic factors to recognize patients at high-risk of failure is necessary.
Radioresistance and chemoresistance are the main factors leading to distant metastasis and tumor progression [8]. Tumor hypoxia is key factor in development of radioand chemo-resistance [9,10]. Numerous efforts have been made to identify tumor hypoxia-related prognostic factors for NPC in recent years [11][12][13][14][15][16][17][18][19][20]; hemoglobin (Hb) levels and cervical node necrosis (CNN) have attracted significant attention. Hb is the principal carrier of oxygen in red blood cells, which carry oxygen from the respiratory organs to the rest of the body. A low Hb level causes a reduction in blood oxygen and can lead to tumor hypoxia. CNN may indicate intratumoral hypoxia [21]. Anemia or/and CNN may in turn affect the efficacy of RT and chemotherapy. Accumulating data suggests there are positive relationships between pretreatment hemoglobin (Hb) levels and CNN and the survival of patients with NPC treated with IMRT ± chemotherapy [11][12][13][14][15][16][17]. As both anemia and CNN may serve as markers of low tumor oxygenation status, it is reasonable to hypothesize the pretreatment hemoglobin level may correlate with and complement CNN to enable improved prediction of the survival of patients with NPC. However, this relationship has not been confirmed experimentally.
We performed this retrospective study to evaluate the impacts of pretreatment anemia and CNN on survival outcomes and then investigated the combined prognostic value of these factors, with the aim of improving outcome

Patient selection and staging evaluation
The inclusion criteria were: (1) had newly diagnosed, pathologically proven, previously untreated NPC; (2) with cervical node metastasis; (3) no evidence of distant metastases; (4) treated using IMRT ± chemotherapy at our cancer center from November 2009 to February 2012; (5) no other tumors or serious illnesses; (6) available pretreatment MRI scans of the nasopharynx and neck; (7) pretreatment Hb level available (measured ≤1 week before treatment); (8) and no other tumors or serious illnesses. The selection criteria were met by 1302 consecutive patients. Participants completed pretreatment baseline evaluations including a complete history, hematological and biochemical profiles, physical and neurological examination, chest radiography, MRI of the neck and nasopharynx, abdominal ultrasonography, and a whole-body bone scan using single photon emission computed tomography. All participants were restaged based on the seventh edition of the International Union against Cancer/American Joint Committee on Cancer system [22]. Patient records/ information were collected by trained data reviewers and were analyzed anonymously. As this study was a retrospective assessment of routine data, the ethics committee of our Cancer Center waived the need for individual informed consent.

Treatment strategies
Treatment strategies were designed using standard protocols depending on clinical TNM stage and the general patient condition. In the relevant study period, institutional guidelines recommended concurrent chemoradiotherapy ± neoadjuvant/adjuvant chemotherapy for stage II to IVB as described previously [23]. Every patient was treated with IMRT at one fraction/day, 5 days per week. IMRT was performed as described previously [23].
Concurrent chemotherapy consisted of cisplatin weekly or every 3 weeks during RT. Neoadjuvant/adjuvant chemotherapy was 1-4 cycles of cisplatin with docetaxel, cisplatin with 5-fluorouracil or cisplatin with 5-fluorouracil and docetaxel every 2 to 3 weeks.

Assessment of anemia and CNN
Anemia was defined based on the Chinese Society of Clinical Oncology (CSCO) clinical practice guidelines and National Cancer Institute criteria on cancer-related anemia: pretreatment hemoglobin <120 g/L in males and <110 g/L in females [21,24,25].
All patients underwent MRI using a 1.5-T MR system (Signa CV/i; General Electric Healthcare, Chalfont St. Giles, United Kingdom) using a head and neck combined coil to examine the region from the suprasellar cistern to the inferior margin at the sternal end of the clavicle. Two radiologists specializing in HN cancer reviewed all MRI scans separately and any disagreements were resolved by discussion every 2 weeks. The definition of necrosis on MRI was T1-weighted contrast-enhanced images showing focal low-signal intensity with or without a surrounding rim of enhancement or T2-weighted images showing focal high-signal intensity.
Patients were classified into four groups: non-anemia and non-CNN group, anemia group (anemia and non-CNN group), CNN group (non-anemia and CNN group), and anemia and CNN group.

Follow-up and end-points
Follow-up time was measured from day 1 of treatment to the time of last visit or death. To detect possible locoregional relapse or distant metastasis, all patients underwent conventional examinations (similar to the pretreatment assessment) every 3 to 6 months during the first 24 months, and 6-monthly thereafter until death. End-points (from first day of treatment to first defining event) were overall survival (OS), disease-free survival (DFS), distant metastasis-free survival (DMFS), and locoregional relapse-free survival (LRRFS).

Statistical analysis
Data analysis used Statistical Package for the Social Sciences version 17.0 software (SPSS, Chicago, IL, USA); Survival rates were calculated using the Kaplan-Meier method, and survival curves were compared using log-rank tests. Clinical characteristics and failure patterns between groups were compared using the Chi-square test or Fisher's exact test. If more than 20% of the expected count in cells is less than 5, use the Fisher's exact probability tests, otherwise use the Pearson chi-square test. Univariate analysis was conducted using the log-rank test. Multivariate analysis used the Cox proportional hazards model including the following covariates: gender, age, T category, N category, WHO histological type, presence or absence of pretreatment anemia, presence or absence of CNN, and chemotherapy (yes/no). Two-tailed P-values < 0.05 were considered significant.
There was no significant difference in the distribution of age (P = 0.459), histological type (P = 0.627), chemotherapy strategies (P = 0.461) or T category (P = 0.197) when the patients were stratified by anemia; however, significant differences were observed in terms of sex (P = 0.004) and N category (P = 0.026). There was no significant difference in terms of age (P = 0.826), sex (P = 0.498) and histological type (P = 0.867) when the patients were stratified by CNN. However, significant differences were observed in terms of T category (P = 0.048), N category (P < 0.001), and chemotherapy strategies (P = 0.001). The pretreatment clinicopathological characteristics of the 1302 patients are listed in Table 1.

Treatment outcomes for the entire cohort
The median follow-up duration was 47.  (87.5%) of these deaths were attributed to NPC, and 19 (12.5%) had died of other causes.
The anemia and CNN group had the poorest survival outcomes, and the non-anemia and non-CNN group had the best survival outcomes. The differences in survival between the four groups are presented in Figure 3. Compared with the non-anemia and non-CNN group, the anemia group, CNN group, anemia and CNN group experienced poorer OS, DMFS, LRRFS, and DFS. In addition, the anemia and CNN group had poorer OS, DMFS, LRRFS and DFS than the CNN group; and the anemia and CNN group had poorer DMFS and DFS than the anemia group. There was no significant difference in any end-point between the anemia group and CNN group.
Multivariate analysis demonstrated that the combined classification was an independent prognostic factor for OS (P = 0.034), DMFS (P = 0.001), LRRFS (P < 0.001), and DFS (P = 0.001). WHO pathology, T category, N category, and age were significant prognostic factor for OS and DFS; N caterogy was an independent prognostic factor for OS; and WHO pathogory was an independent prognostic factor for LRRFS. In addition, the anemia and CNN group had much higher HRs for OS, DMFS, LRRFS, and DFS than the other three groups, whereas the nonanemia and non-CNN group had the lowest HRs for all endpoints (Table 4).

Discussion
Radioresistance and chemoresistance are the major factors associated with poor locoregional control and distant failure in malignant tumors [8]. Tumor hypoxia plays a major role in radioresistance and chemoresistance [9,10]. Recent evidence indicates both anemia and CNN are associated with tumor hypoxia, which negatively impact treatment efficacy [11][12][13][14][15][16][17]. Therefore, a large populationbased study was performed to determine the combined prognostic value of anemia and CNN in current study. Based on this work, we propose a new prognostic model that enables accurate stratification of individual patients with NPC receiving IMRT ± chemotherapy into four risk groups.
This study revealed pretreatment anemia was a prognostic factor for poorer DMFS, and pretreatment CNN was a prognostic factor for poorer OS, DFS, DMFS, and LRRFS. Therefore, we sought to establish a prognostic model. The non-anemia and non-CNN group had better survival outcomes than the anemia group, CNN group, anemia and CNN group, and the anemia and CNN group experienced the poorest survival outcomes. The combined classification was an independent prognostic factor for OS, LRRFS, DMFS, and DFS. The anemia and CNN group had much higher HRs for all endpoints than the other three groups, indicating combined assessment of CNN and anemia improves risk stratification and prognostication and suggesting more intensive treatment modality may be more suitable for patients with pretreatment anemia and CNN.
The prognostic significance of low pretreatment anemia has previously been investigated in various cancers; low Hb is associated with poorer outcomes in many types of human cancer, particularly in HN cancers and gynecological tumors [26][27][28]. A positive relationship between survival outcomes and Hb levels has been demonstrated in patients with NPC receiving RT ± chemotherapy: low pretreatment Hb was related to increased risk of death or metastasis [11][12][13][14][15].
Moreover, CNN has also been demonstrated to have prognostic value in HN cancer after RT and/or chemotherapy [29][30][31]: the presence of hypodense lymph nodes, which indicate CNN [17], was associated with regional treatment failure. Lan et al. [14] were the first to report CNN was a negative prognostic factor for OS, DMFS, RRFS, and DFS in patients with NPC treated using IMRT, three-dimensional radiotherapy (3D-RT) or 2D-RT. Moreover, in a recent study by Zhang et al., CNN was demonstrated to be an independent, adverse prognostic factor for OS, DFS, LRRFS, and DMFS in patients with NPC treated with IMRT [17].
The major purpose of this research was to combine the prognostic value of anemia and CNN to improve risk stratification and prognostication in NPC. We found the combination of anemia and CNN improved prognostic ability. However, the precise mechanisms underlying the cumulative adverse effect of anemia and CNN remain unclear. As both anemia and CNN are associated with hypoxia, we speculate the presence of both may indicate more severe tumor hypoxia compared to one factor alone. As there is a potential relationship between hypoxia and radioresistance and chemoresistance and the anemia and CNN group had the poorest prognosis compared to the other three groups, we suggest more aggressive IMRT ± chemotherapy protocols should be designed for high-risk patients with anemia and CNN.
The principal limitation of this study was the fact treatment heterogeneity was inevitable as this was a retrospective analysis of patients with NPC from a single institution. Thus, a large, multicenter prospective study is required to further confirm the findings of current research.

Conclusion
Classification combining anemia and CNN is an independent prognostic factor for OS, DFS, DMFS, and LRRFS in patients with NPC treated with IMRT ± chemotherapy. Combined assessment of anemia and CNN improves risk stratification and prognostication: patients in the nonanemia and non-CNN group had the best survival outcomes; and the anemia and CNN group experienced the poorest survival outcomes. This data may aid the design of individualized treatment plans to improve the treatment outcomes of patients with NPC receiving IMRT ± chemotherapy.