- Email: [email protected]
Comorbidity predicts poor prognosis in nasopharyngeal carcinoma: Development and validation of a predictive score model
Radiotherapy and Oncology xxx (2015) xxx–xxx
Contents lists available at ScienceDirect
Radiotherapy and Oncology journal homepage: www.thegreenjournal.com
Original article
Comorbidity predicts poor prognosis in nasopharyngeal carcinoma: Development and validation of a predictive score model Rui Guo a,1, Xiao-Zhong Chen b,1, Lei Chen a, Feng Jiang b, Ling-Long Tang a, Yan-Ping Mao a, Guan-Qun Zhou a, Wen-Fei Li a, Li-Zhi Liu c, Li Tian c, Ai-Hua Lin d, Jun Ma a,⇑ a
Department of Radiation Oncology, Sun Yat-sen University Cancer Center, State Key Laboratory of Oncology in Southern China, Collaborative Innovation Center for Cancer Medicine; Department of Radiation Oncology, Zhejiang Cancer Hospital; c Imaging Diagnosis and Interventional Center, Sun Yat-sen University Cancer Center, State Key Laboratory of Oncology in Southern China, Collaborative Innovation Center for Cancer Medicine; and d Department of Medical Statistics and Epidemiology, School of Public Health, Sun Yat-sen University, People’s Republic of China
b
a r t i c l e
i n f o
Article history: Received 4 November 2014 Received in revised form 5 December 2014 Available online xxxx Keywords: Nasopharyngeal carcinoma Comorbidity Prognosis Survival Multi-cancer centre
a b s t r a c t Background and purpose: The impact of comorbidity on prognosis in nasopharyngeal carcinoma (NPC) is poorly characterized. Material and methods: Using the Adult Comorbidity Evaluation-27 (ACE-27) system, we assessed the prognostic value of comorbidity and developed, validated and confirmed a predictive score model in a training set (n = 658), internal validation set (n = 658) and independent set (n = 652) using area under the receiver operating curve analysis. Results: Comorbidity was present in 40.4% of 1968 patients (mild, 30.1%; moderate, 9.1%; severe, 1.2%). Compared to an ACE-27 score 61, patients with an ACE-27 score >1 in the training set had shorter overall survival (OS) and disease-free survival (DFS) (both P < 0.001), similar results were obtained in the other sets (P < 0.05). In multivariate analysis, ACE-27 score was a significant independent prognostic factor for OS and DFS. The combined risk score model including ACE-27 had superior prognostic value to TNM stage alone in the internal validation set (0.70 vs. 0.66; P = 0.02), independent set (0.73 vs. 0.67; P = 0.002) and all patients (0.71 vs. 0.67; P < 0.001). Conclusions: Comorbidity significantly affects prognosis, especially in stages II and III, and should be incorporated into the TNM staging system for NPC. Assessment of comorbidity may improve outcome prediction and help tailor individualized treatment. Ó 2014 Elsevier Ireland Ltd. All rights reserved. Radiotherapy and Oncology xxx (2015) xxx–xxx
Nasopharyngeal carcinoma (NPC) has a distinct epidemiology and geographic distribution. Southern China including Hong Kong has the highest incidence of NPC, with an annual incidence of 20– 50 cases per 100,000 [1]. Currently, the extent of disease, as represented by the tumour-node-metastasis (TNM) staging system, is the main tool for determining clinical treatment strategies and predicting treatment outcome [2]. However, recent findings suggest that the current AJCC staging system is not satisfactorily accurate in terms of survival prediction [3]. Therefore, the identification of novel prognostic factors is of great importance for improving outcome prediction and tailoring individualized treatments in patients with NPC.
⇑ Corresponding author at: State Key Laboratory of Oncology in South China, Department of Radiation Oncology, Sun Yat-sen University Cancer Center, 651 Dongfeng Road East, Guangzhou 510060, People’s Republic of China. E-mail address: [email protected] (J. Ma). 1 Rui Guo and Xiao-Zhong Chen contributed equally to this article.
Comorbidity refers to the presence of one or more co-existing conditions that are not caused by the primary tumour [4]. A relationship between the severity of comorbidity and treatment outcome has been observed in several other types of cancer, including head and neck, breast, prostate and colon cancers [5– 7]. The existence of comorbidity in a patient can affect treatment decisions in daily clinical practice, increase the chances of complications after treatment and, consequently, affect their final outcome [8]. Therefore, the evaluation of comorbidity as a component of prognostic risk assessment may drive improvements in the treatment and survival of patients with NPC. NPC has distinct epidemiology, pathology and clinical attributes compared to other head and neck cancers. To date, only one study has investigated the effect of comorbidity in NPC, and reported that comorbidity did not affect prognosis independently of the TNM staging system in 59 patients with NPC [9]. The precise role of comorbidity in NPC needs to be satisfactorily investigated in order to enable comprehensive risk evaluation before treatment. Additionally, there is no information available on whether comorbidity
http://dx.doi.org/10.1016/j.radonc.2014.12.002 0167-8140/Ó 2014 Elsevier Ireland Ltd. All rights reserved.
Please cite this article in press as: Guo R et al. Comorbidity predicts poor prognosis in nasopharyngeal carcinoma: Development and validation of a predictive score model. Radiother Oncol (2015), http://dx.doi.org/10.1016/j.radonc.2014.12.002
2
Impact of comorbidity on prognosis in NPC
could be reliably incorporated into the TNM staging system for patients with NPC. On this basis, we conducted a retrospective study to obtain insights into the effect of comorbidity on the survival of patients with NPC and investigated whether combining information on comorbidity with prognostic staging systems could improve our ability to assess the prognosis of patients with NPC. Patients and methods Patient characteristics and study design A total of 2055 patients with newly-diagnosed, non-metastatic, histologically-proven NPC who were treated at Sun Yat-sen University Cancer Center (Guangzhou, People’s Republic of China) and Zhejiang Cancer Hospital between February 2003 and January 2007 were retrospectively reviewed. Of these, 87 cases were excluded due to a lack of substantial information regarding comorbidity; the remaining 1968 cases were included in the current study. This retrospective study was approved by the Institutional Review Boards of Sun Yat-sen University Cancer Center and Zhejiang Cancer Hospital. All patients underwent pretreatment evaluations, including a complete medical history, physical examination, hematology and biochemistry profiles, coagulation test, electrocardiogram (ECG), infectious disease examination (Hepatitis, HIV, syphilis), magnetic resonance imaging (MRI) of the neck and nasopharynx, chest radiography, abdominal sonography and whole-body bone scan using single-photon emission computed tomography (ECT); 205 (10.4%) patients also underwent positron emission tomography CT (PETCT). All patients were restaged according to the 7th edition of the International Union against Cancer/American Joint Committee on Cancer (UICC/AJCC) system [10].
Follow-up and statistical analysis Patients were followed up every 3 months during the first 2 years, and every 6 months thereafter until death. The primary end point was overall survival (OS), defined from the first day of therapy to death due to any cause. The secondary end points included disease-free survival (DFS), defined as the time from the first day of therapy to progression or death, whichever occurred first. We randomly divided the 1316 patients from Sun Yat-sen University into a training set of 658 cases and a validation set of 658 cases using a computer-generated allocation sequence. To assess whether the comorbidities identified in these sets had similar prognostic values in different populations, we used the 652 cases from Zhejiang cancer hospital as an independent set to externally validate the results. We assessed the relationships between clinical characteristics and ACE-27 score using the v2 test or Fisher’s exact test. Actuarial rates were calculated using the Kaplan–Meier method and differences were compared using the log-rank test. Univariate and multivariate analyses using a Cox proportional hazards model were used to test the independent significance of different factors by backward elimination. To develop a more sensitive predictive tool, we constructed a prognostic score model by combining the independent prognostic factors identified in the training set (see appendix for details). The regression coefficient of each independent prognostic factor in the proportional hazards model was divided by the regression coefficient of the ACE-27, and rounded into an integer value to generate the risk score.[20]. We compared the prognostic validity of this model with TNM staging alone using receiver operating curve (ROC) analysis in the validation set and independent set. Stata Statistical Package (STATA 12; StataCorp LP, College Station, Texas, USA) was used for all analyses. Results
Adult Comorbidity Evaluation-27 system Comorbidity was assessed using the Adult Comorbidity Evaluation-27 (ACE-27) system, which was developed for patients with cancer, specifically head and neck cancer [11–15]. The ACE-27 system is a validated 27-item comorbidity index [16], by which specific conditions are graded into three degrees (grade 0 = none, grade 1 = minimal, grade 2 = moderate, grade 3 = severe) based on organ system decompensation. The overall comorbidity score is assigned based on the highest-ranking single ailment. When two or more moderate ailments occurred in different organ systems, the overall comorbidity score was graded as severe. The ACE-27 scores were manually calculated by one physician who reviewed detailed information on the patients’ baseline medical condition and comorbidities. Treatment All patients were treated with definitive radiation therapy. Details of the radiotherapy techniques applied at Sun Yat-sen University Cancer Center have been reported previously [17–19]. Among the 1968 patients, 1238 (62.9%) underwent two-dimensional radiotherapy (2D-CRT) and 730 (37.1%) underwent threedimensional conformal radiotherapy (3-DCRT) or intensity-modulated radiotherapy (IMRT). During the study period, as per institutional guidelines, the majority (964/1312; 73.5%) of patients with stage III or IV NPC (classified as T3-T4 and/or N2-N3 disease) underwent chemotherapy. Inductive chemotherapy (IC) or adjuvant chemotherapy (AC) consisted of cisplatin with 5-fluorouracil or cisplatin with taxoids every 3 weeks for two to three cycles. Concomitant chemotherapy consisted of cisplatin at weeks 1, 4 and 7 of radiotherapy or weekly cisplatin.
The median duration of follow-up for the entire patient group was 64.4 months (range, 2.7–126.5 months). A total of 222 (11.3%) patients developed local failure or regional failure, 271 (13.8%) developed distant metastases, and 429 (21.8%) patients died. The 5-year DFS and OS rates for the entire cohort were 71.2% and 79.1%, respectively. The overall distribution of patients with stages I, II, III, and IVAB disease was 7.1%, 26.2%, 39.8%, and 26.9%, respectively. The clinical characteristics of the patients in the training, validation and independent sets are shown in Table 1. Prevalence, type and severity of comorbidity Review of the patients’ medical records showed that a total of 795 patients had comorbidities (40.4%). Of the 795 patients, 592 (30.1%) had an ACE-27 score of 1; 179 (9.1%), an ACE-27 score of 2; and 24 (1.2%), an ACE-27 score of 3. Gastrointestinal disease, substance abuse and cardiovascular disease were most common, with an incidence of 19.4%, 16.1% and 7.1%, respectively. No patients were diagnosed with acquired immune deficiency syndrome or human immunodeficiency virus infection. Data on the prevalence and severity of the comorbidities in the training, validation and independent sets are presented in Table 2. There was no correlation between disease stage or histology and comorbidity burden in the training, validation and independent sets. Not surprisingly, comorbidity varied significantly by age and gender (P < 0.05; Table 1). Prognostic value of comorbidity In the training set (n = 658), the 5-year OS and DFS rates for patients with an ACE-27 score 61 vs. > 1 were 81.1% vs. 59.3%
Please cite this article in press as: Guo R et al. Comorbidity predicts poor prognosis in nasopharyngeal carcinoma: Development and validation of a predictive score model. Radiother Oncol (2015), http://dx.doi.org/10.1016/j.radonc.2014.12.002
3
R. Guo et al. / Radiotherapy and Oncology xxx (2015) xxx–xxx Table 1 Characteristics of the 1968 patients with nasopharyngeal carcinoma. Characteristic
Training set (n = 658) ACE-27 6 1
Age (y) 650 >50 Gender Male Female Histology WHO type I WHO type II/III T classification* T1 T2 T3 T4 N classification* N0 N1 N2 N3 Clinical stage* I II III IVA-B OS 5-year survival DFS 5-year survival
Validation set (n = 658)
ACE-27 > 1
P
ACE-27 6 1
ACE-27 > 1
0.011 426 (70.2) 181 (29.8)
27 (52.9) 24 (47.1)
454 (74.8) 153 (25.2)
48 (94.1) 3 (5.9)
4 (0.7) 603 (99.3)
0 (0) 51 (100.0)
135 142 197 133
(22.2) (23.4) (32.5) (21.9)
10 (19.6) 7 (13.7) 19 (37.3) 15 (29.4)
179 (29.5) 274 (45.1) 105 (17.3) 49 (8.1)
10 (19.6) 26 (51.0) 8 (15.7) 7 (13.7)
53 (8.7) 156 (25.7) 228 (37.6) 170 (28.0)
3 (5.9) 10 (19.6) 16 (31.4) 22 (43.1)
81.1
59.3
74.1
43.9
Independent set (n = 652) P
ACE-27 6 1
ACE-27 > 1
320 (56.9) 242 (43.1)
38 (42.2) 52 (57.8)
383 (68.1) 179 (31.9)
82 (91.1) 8 (8.9)
3 (0.5) 559 (99.5)
0 (0) 90 (100)
88 (15.7) 267 (47.5) 125 (22.2) 82 (14.6)
17 44 18 11
(18.9) (48.9) (20.0) (12.2)
116 (20.6) 206 (36.7) 192 (34.2) 48 (8.5)
20 28 28 14
(22.2) (31.1) (31.1) (15.6)
21 (3.7) 159 (28.3) 259 (46.1) 123 (21.9)
4 (4.4) 23 (25.6) 38 (42.2) 25 (27.8)
78.1
62.1
70.5
49.8
0.008 432 (72.5) 164 (27.5)
35 (56.5) 27 (43.5)
428 (71.8) 168 (28.2)
58 (93.5) 4 (6.5)
3 (0.5) 593 (99.5)
0 (0) 62 (100)
140 135 193 128
10 15 21 16
<0.001
0.009
<0.001
0.561
<0.001
0.576
0.308
0.669
0.591 (23.5) (22.7) (32.4) (21.5)
(16.1) (24.2) (33.9) (25.8)
0.503
0.797
0.765 185 (31.0) 252 (42.3) 111 (18.6) 48 (8.1)
16 (25.8) 26 (41.9) 14 (22.6) 6 (9.7)
55 (9.2) 155 (26.0) 219 (36.7) 167 (28.0)
4 (6.5) 13 (21.0) 23 (37.1) 22 (35.3)
82.6
66.4
74.6
55.3
0.150
0.173
0.545
<0.001
0.626
0.001
<0.001
P
<0.001
0.001
<0.001
RT, Radiotherapy; 2-DRT, two-dimensional radiotherapy; 3-DCRT, three-dimensional conformal radiotherapy; IMRT, intensity-modulated radiotherapy; ACE-27, Adult Comorbidity Evaluation-27. * According to the 7th AJCC/UICC staging system.
Table 2 Presence and severity of comorbidities in the study population training set, validation set and independent set. Disease classification
Training set n = 658
ACE-27 score
Grade 1
Grade 2
Grade 3
Validation set n = 658 Grade 1
Grade 2
Grade 3
Grade 1
Grade 2
Grade 3
Gastrointestinal system Substance abuse Cardiovascular system Respiratory system Renal system Endocrine system Neurological system Psychiatric Rheumatologic Immunological system Malignancy Body weight Overall ACE-27 score
164 (24.9) 48 (7.3) 31 (4.7) 14 (2.1) 18 (2.7) 11 (1.7) 2 (0.3) 1 (0.2) 0 (0.0) 0 (0.0) 2 (0.3) 0 (0.0) 237 (36.0)
5 (0.8) 39 (5.9) 1 (0.2) 2 (0.3) 0 (0.0) 1 (0.2) 0 (0.0) 0 (0.0) 0 (0.0) 0 (0.0) 3 (0.5) 1 (0.2) 48 (7.3)
0 0 0 0 0 0 0 0 0 0 1 0 3
150 (22.8) 41 (6.2) 35 (5.3) 11 (1.7) 16 (2.4) 9 (1.4) 4 (0.6) 0 (0.0) 2 (0.3) 0 (0.0) 2 (0.3) 1 (0.2) 204 (31.0)
6 (0.9) 41 (6.2) 6 (0.9) 2 (0.3) 0 (0.0) 1 (0.2) 0 (0.0) 0 (0.0) 0 (0.0) 0 (0.0) 2 (0.3) 0 (0.0) 58 (8.8)
0 0 0 0 0 0 0 0 0 0 4 0 4
41 (6.3) 103 (15.8) 48 (7.4) 4 (0.6) 1 (0.2) 7 (1.1) 0 (0.0) 0 (0.0) 2 (0.3) 0 (0.0) 0 (0.0) 0 (0.0) 150 (23.0)
15 (2.3) 43 (6.6) 13 (2.0) 4 (0.6) 0 (0.0) 6 (0.9) 1 (0.2) 1 (0.2) 1 (0.2) 0 (0.0) 3 (0.5) 0 (0.0) 73 (11.2)
2 (0.3) 0 (0.0) 5 (0.8) 1 (0.2) 1 (0.2) 0 (0.0) 1 (0.2) 0 (0.0) 0 (0.0) 0 (0.0) 2 (0.3) 0 (0.0) 17 (2.6)
(0.0) (0.0) (0.0) (0.1) (0.0) (0.0) (0.0) (0.0) (0.0) (0.0) (0.2) (0.0) (0.5)
Independent set n = 652
(0.0) (0.0) (0.0) (0.0) (0.0) (0.0) (0.0) (0.0) (0.0) (0.0) (0.6) (0.0) (0.6)
ACE-27, Adult Comorbidity Evaluation-27.
(P < 0.001) and 74.1% vs. 43.9% (P < 0.001), respectively (Fig. 1A, 1B). In the validation set (n = 658), the 5-year OS and DFS rates for patients with an ACE-27 score 61 vs. > 1 were 82.6% vs. 66.4% (P = 0.001) and 74.6% vs. 55.3% (P = 0.001) (Fig. 1C, 1D). In the independent set (n = 652), the 5-year OS and DFS rates for patients with an ACE-27 score 61 vs. >1 were 78.1.6% vs. 62.1% (P < 0.001) and 70.5% vs. 49.8% (P < 0.001) (Fig. 1E, 1F). Overall, patients with an ACE-27 score >1 had poorer OS and DFS than patients with an ACE-27 score 61. In the multivariate analysis, ACE-27 score was a significant, independent prognostic factor for OS (hazard ratio [HR] = 2.027) and DFS (HR = 2.365) in the training set, OS (HR = 1.660) and DFS (HR = 1.674) in the validation set, and OS (HR = 1.829) and DFS (HR = 1.826) in the independent set (all P < 0.05; Table 3).
Varying impact of comorbidity by stage In order to analyse the prognostic value of comorbidity in patients with the same stage of disease, all patients (n = 1968) with NPC were stratified according to TNM stage. In stage I, the 5-year OS and DFS rates for patients with an ACE-27 score 61 vs. >1 were 98.3% vs. 87.5% (P = 0.084) and 95.9% vs. 78.8% (P = 0.079), respectively. In stage II, the 5-year OS and DFS rates for patients with an ACE-27 score 61 vs. >1 were 90.1% vs. 77.1% (P = 0.002) and 84.0% vs. 63.8% (P < 0.001), respectively. In stage III, the 5-year OS and DFS rates for patients with an ACE-27 score 61 vs. >1 were 83.3% vs. 55.3% (P < 0.001) and 75.1% vs. 41.9% (P < 0.001), respectively. In stage IV, the 5-year OS and DFS rates for patients with an ACE-27 score 61 vs. >1 were 62.5% vs. 54.4% (P = 0.107) and
Please cite this article in press as: Guo R et al. Comorbidity predicts poor prognosis in nasopharyngeal carcinoma: Development and validation of a predictive score model. Radiother Oncol (2015), http://dx.doi.org/10.1016/j.radonc.2014.12.002
4
Impact of comorbidity on prognosis in NPC
Table 3 Multivariate analysis of the impact of all variables on survival in the training, validation and independent sets. Patients
Endpoint
Variable
HR
95% CI
P-Valueb
Training set
OS
ACE-27 T classificationa N classificationa Age
2.027 2.147 2.363 1.021
1.259–3.265 1.500–3.073 1.690–3.305 1.006–1.035
0.004 <0.001 <0.001 0.006
DFS
ACE-27 T classificationa N classificationa
2.365 1.760 2.161
1.574–3.554 1.307–2.370 1.610–2.900
<0.001 <0.001 <0.001
Age ACE-27 T classificationa N classificationa Age
1.012 1.660 2.155 1.920 1.031
0.999–1.024 1.043–2.643 1.481–3.136 1.355–2.719 1.016–1.047
0.069 0.033 <0.001 <0.001 <0.001
DFS
ACE-27 T classificationa N classificationa Age
1.674 1.996 1.628 1.016
1.122–2.498 1.461–2.726 1.204–2.200 1.003–1.029
0.012 <0.001 0.002 0.017
OS
ACE-27 T classificationa N classificationa Age
1.829 2.007 1.591 1.040
1.240–2.699 1.445–2.788 1.146–2.209 1.025–1.055
0.002 <0.001 0.006 <0.001
DFS
ACE-27 T classificationa N classificationa Age
1.826 2.009 1.623 1.026
1.304–2.556 1.518–2.660 1.227–2.146 1.013–1.038
<0.001 <0.001 0.001 <0.001
Validation set
Independent set
OS
OS, overall survival; DFS, disease-free survival; ACE-27, Adult Comorbidity Evaluation-27: CI, confidence interval; HR, hazard ratio. a According to the 7th AJCC/UICC staging system. b Multivariate P-values were calculated using an adjusted Cox proportional-hazards model. The following parameters were included in the Cox proportion hazard model by backward elimination: age (continuous variable), gender (male vs. female), T classification (T3–4 vs. T1–2), N classification (N2–3 vs. N0–1), use of chemotherapy (with vs. without) and ACE-27 (grade >1 vs. grade 61).
53.8% vs. 44.2% (P = 0.132), respectively. Overall, the comorbid effect was most evident in stage II–III, and insignificant in stage I or IV.
Prognostic value of adding the ACE-27 score to the TNM staging system The three significantly independent prognostic variables including ACE-27 grade, age and TNM stage were used to construct a prognostic-score model. A cumulative risk score then was calculated for each patient in the training set, the validation set, the sum of the points ranged from 0 to 7. Then the patients were divided into four risk groups (Group 1: risk score 0–1; Group 2: risk score 2–4; Group 3: risk score 5; Group 4: risk score 6–7) according to their overall survival in the training set (Supplementary materials). In the validation set, the OS rates for Groups 1, 2, 3, and 4 were 98.1%, 88.5%, 74.2%, and 54.3% (P < 0.001) and the DFS rates were 96.4%, 81.4%, 61.1%, and 48.2% (P < 0.001), respectively (Fig. 2A, 2B). In the independent set, the OS rates for Groups 1, 2, 3, and 4 were 100%, 88.5%, 67.5%, and 44.0% (P < 0.001) and the DFS rates were 100%, 80.0%, 60.1%, and 33.7% (P < 0.001), respectively (Fig. 2C, 2D). For all patients, the OS rates for Groups 1, 2, 3, and 4 were 98.4%, 88.5%, 71.6%, and 48.1% (P < 0.001) and the DFS rates were 95.3%, 81.1%, 61.6%, and 41.1% (P < 0.001), respectively (Fig. 2E, 2F). For OS, the area under the ROC curves (AUCs) for the TNM stage model alone and combined risk grouping were 0.66 vs. 0.70 (P = 0.02) in the validation set (Fig. 3A), 0.67 vs. 0.73 (P = 0.002) in the independent set (Fig. 3B), and 0.67 vs. 0.71 (P < 0.001) for overall patients (Fig. 3C). Therefore, the combined risk grouping created by adding ACE-27 scores to TNM staging was superior to TNM staging alone in terms of predicting survival outcomes.
Discussion We explored the potential prognostic value of comorbidity in patients with NPC from two cancer centres in different areas of China, and to our knowledge, is the first such investigation in a large cohort of patients with NPC. This study provides strong evidence in support of comorbidity as an independent prognostic factor in patients with NPC. Prevalence, type and severity of comorbidity Previous studies reported the incidence of comorbidity in head and neck cancer to be 33–65%, with cardiovascular and pulmonary diseases being the most common [7,21–25]. Ramakrishnan et al. reported an incidence of comorbidity of 44% in 59 patients with NPC, with 19% patients having moderate or severe comorbidity and the cardiovascular system (27%) most commonly affected [9]. The comorbidity rate in this study was 40.4%, which is comparable to the rates reported previously. However, the most common comorbidities differed in the patients from different areas: the most common comorbidities in the training set and validation set were comparable, but were different to that of the independent set. The patients in the training set and validation set are from southern China, and the most common comorbidities in these groups were gastrointestinal diseases, with liver disease the leading comorbidity; these results are similar to the findings of another study [21]. This is not surprising, as epidemiological research has demonstrated that southern China has one of the highest incidences of chronic hepatitis B virus (HBV) infection, and the hepatitis B surface antigen (HBsAg) seropositivity rate in the general population is 10–12% [26]. Therefore, chronic HBV infection may be an important gastrointestinal comorbidity in patients with NPC in southern China.
Please cite this article in press as: Guo R et al. Comorbidity predicts poor prognosis in nasopharyngeal carcinoma: Development and validation of a predictive score model. Radiother Oncol (2015), http://dx.doi.org/10.1016/j.radonc.2014.12.002
R. Guo et al. / Radiotherapy and Oncology xxx (2015) xxx–xxx
5
Fig. 1. Kaplan–Meier estimates of overall survival (OS) and disease-free survival (DFS) for patients with nasopharyngeal cancer according to ACE-27 score. OS (A) and DFS (B) for the 658 patients in the training set, OS (C) and DFS (D) for the 658 patients in the validation set, and OS (E) and DFS (F) for the 652 patients in the independent set.
The incidence of comorbidity increased with increase in age in this study. Piccirillo et al. also investigated the prevalence of comorbidity across the age spectrum, and found increased age to be associated with an increase in the number and severity of comorbidities [27]. Similar results have been reported by other studies, in that comorbidity was present in younger patients, but the incidence of comorbidity increased with age [9,21,28,29]. Comorbidity is an independent prognostic factor in NPC Although the incidence of comorbidity and the most common comorbidities varied in different areas (training, validation and
independent sets), comorbidity significantly affected the survival of the patients with NPC in all sets, as a higher ACE-27 score was associated with poorer survival outcomes. The association between comorbidity and poorer overall survival may mainly be due to the physiological burden of chronic disease; however, the precise impact of comorbidity on disease-free survival remains unclear. It has been reported that comorbidity did not affect cancer-specific survival in patients treated with radiotherapy only [29,30]; however, diverging results were observed in patients treated with chemoradiotherapy and surgery [13,21]. In this study, the majority (964/1312; 73.5%) of patients with stage III or IV NPC (classified as T3–T4 and/or N2–N3 disease) underwent chemoradiotherapy.
Please cite this article in press as: Guo R et al. Comorbidity predicts poor prognosis in nasopharyngeal carcinoma: Development and validation of a predictive score model. Radiother Oncol (2015), http://dx.doi.org/10.1016/j.radonc.2014.12.002
6
Impact of comorbidity on prognosis in NPC
Fig. 2. Kaplan–Meier estimates of overall survival (OS) and disease-free survival (DFS) using the combined risk grouping in patients with nasopharyngeal cancer. OS (A) and DFS (B) for the 658 patients in the validation set, OS (C) and DFS (D) for the 652 patients in the independent set, and OS (E) and DFS (F) for all 1968 patients. Group 1: risk score 0–1; Group 2: risk score 2–4; Group 3: risk score 5; Group 4: risk score 6–7.
The reasons explaining the association between comorbidity and DFS in patients with NPC are unclear. The possible reasons are: firstly, may simply reflect treatment biases on the part of the physician. The presence of comorbidity may necessitate the use of a less than ideal or less radical procedure, which may lead to poorer cancer-specific survival [8,16]. Secondly, an increase in the severity of comorbidities may increase the toxicity of specific treatments, which may consequently shorten the survival period to the extent of canceling out any possible gain obtained from the therapy [16,31,32]. Furthermore, the most frequently observed comorbidity was HBV infection-associated liver disease, which is similar to the finding of a previous study in which HBV infection-
associated liver disease was reported to negatively affect cancerspecific survival [33]. The latent mechanism responsible for the unfavourable prognosis of HBV-infected patients with NPC may be linked to certain types of immunological dysfunction, as HBV is associated with immune dysfunction, as demonstrated by hepatitis B-related nephritis and its association with lymphoma [34,35]. However, this hypothesis needs to be addressed in future studies. Varying impact of comorbidity by TNM stage Apart from comorbidity, TNM stage remained a significantly predictive covariate in the Cox regression model. However, the
Please cite this article in press as: Guo R et al. Comorbidity predicts poor prognosis in nasopharyngeal carcinoma: Development and validation of a predictive score model. Radiother Oncol (2015), http://dx.doi.org/10.1016/j.radonc.2014.12.002
R. Guo et al. / Radiotherapy and Oncology xxx (2015) xxx–xxx
7
Fig. 3. Sensitivity and specificity for the prediction of overall survival for the combined risk grouping (including the ACE-27 score) and the TNM stage alone model for (A) the 658 patients in the validation set, (B) 652 patients in the independent set and (C) all 1968 patients. AUC: area under the curve; ⁄P-values vs. TNM stage alone.
prognosis of patients with the same TNM stage but different severities of comorbidity was significantly different. This suggests that the current staging system is inadequate for defining prognosis and does not adequately reflect the biological heterogeneity of NPC. Furthermore, in the subgroup analyses, the impact of comorbidity on survival outcomes was strongly associated with TNM stage. In patients with stage II–III disease, a high ACE-27 score was significantly associated with outcome, while ACE-27 score had no significant effect on outcome in patients with stage I or IV disease. Our finding that comorbidity has little impact on the survival of patients with aggressive disease (stage IV) is also in accordance with previous studies in other cancers, in which comorbidity had the greatest prognostic impact among groups with the highest survival rate and the least impact in groups with the lowest survival rate [7,36]. Additionally, there was no significant association between comorbidity and either survival outcome in patients with stage I NPC. Firstly, treatment biases cannot exist in patients with stage I, as the treatment strategy for this group was radiotherapy only. Secondly, the failure rate for stage I is very low. However, the assessment of comorbidity and treatment selection appear to be particularly important in patients with stage II–III NPC, as the disease is less aggressive or less widely spread in these patients, and thus they have a better overall prognosis than patients with more advanced disease. Therefore, comorbidity interventions and more meticulous disease follow-up should be provided for patients with stage II–III disease in order to obtain maximum improvement in survival rates. Prognostic value of adding ACE-27 score to TNM staging In addition to comorbidity, T stage and N stage remained significant predictive covariates in the Cox regression model. Another major objective of this study was to verify the combined prognostic value of comorbidity and disease stage. We constructed a prognostic score model by combining the independent prognostic factors identified in the training set (Supplementary appendix), and compared its prognostic validity with TNM stage alone. ROC analysis of the validation and independent sets suggested that prognostic assessment could be improved by combining the ACE27 system with TNM staging. The ACE-27 system requires no special technological expertise for clinical use and can be applied by any healthcare professional using information that is readily available from a patient’s history or pretreatment evaluations. Therefore, a comorbidity index such as the ACE-27 system should be taken into account along with the clinical TNM stage when treating patients with NPC. Moreover, patients with comorbidities should be offered appropriate supportive care in order to ensure optimal cancer treatment; and such an approach necessitates a multidisciplinary approach to patient care. This study was a retrospective chart-based assessment; therefore, a selection bias may exist in the collection of comorbidity data. However, the results of this study indicate that comorbidity
may vary within patients with NPC from different areas, and highlights the significant impact of comorbidity on survival in patients with NPC. The results of this study also justify the need for collecting prospective comorbidity data in routine clinical practice, and indicate that a comorbidity index should be integrated into the current TNM staging system to improve treatment selection and outcome prediction. Conclusion For patients with comorbidity, the ACE-27 score is a significant predictor of prognosis in NPC. Assessment of comorbidity before treatment may help to improve outcome prediction and tailor individualized treatments for patients with NPC. Information on comorbidity could be added to staging systems and be incorporated into treatment strategy decision-making, especially for patients with stage II–III NPC. Conflict of interest statement The authors indicated no potential conflicts of interest. Acknowledgements This work was supported by grants from the Innovation Team Development Plan of the Ministry of Education (No. IRT1297), the National Natural Science Foundation of China (Nos. 81230056, 81201746), the Guangdong Province Universities and Colleges Pearl River Scholar Funded Scheme (2010), the Science and Technology Project of Guangzhou (No. 12BppZXaa2060002), the Key Laboratory Construction Project of Guangzhou City, China (No. 121800085), and the Key Scientific and Technological Innovation Program for Universities of Guangdong Province (No. cxzd1005). Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.radonc.2014.12. 002. References [1] Jemal A, Bray F, Center MM, et al. Global cancer statistics. CA Cancer J Clin 2011;61:69–90. [2] Wei WI, Sham JST. Nasopharyngeal carcinoma. Lancet 2005;365:2041–54. [3] Takes RP, Rinaldo A, Silver CE, et al. Future of the TNM classification and staging system in head and neck cancer. Head Neck 2010;32:1693–711. [4] Piccirillo JF, Feinstein AR. Clinical symptoms and comorbidity: significance for the prognostic classification of cancer. Cancer 1996;77:834–42. [5] Kazi R, Nutting CM, Rhys-Evans P, Harrington KJ. Significance and prognostic impact of co-morbidity in head and neck cancer. J Cancer Res Ther 2009;5:145–7.
Please cite this article in press as: Guo R et al. Comorbidity predicts poor prognosis in nasopharyngeal carcinoma: Development and validation of a predictive score model. Radiother Oncol (2015), http://dx.doi.org/10.1016/j.radonc.2014.12.002
8
Impact of comorbidity on prognosis in NPC
[6] Tanvetyanon T, Padhya T, McCaffrey J, et al. Prognostic factors for survival after salvage reirradiation of head and neck cancer. J Clin Oncol 2009;27:1983–91. [7] Alho OP, Hannula K, Luokkala A, et al. Differential prognostic impact of comorbidity in head and neck cancer. Head Neck 2007;29:913–8. [8] Boje CR. Impact of comorbidity on treatment outcome in head and neck squamous cell carcinoma – a systematic review. Radiother Oncol 2013. [9] Ramakrishnan Y, Paleri V, Shah R, et al. Comorbidity in nasopharyngeal carcinoma: a preliminary communication on the prevalence, descriptive distribution and impact on outcome. Clin Otolaryngol 2007;32: 484–8. [10] Edge SB, Compton CC. The American Joint Committee on Cancer: the 7th edition of the AJCC cancer staging manual and the future of TNM. Ann Surg Oncol 2010;17:1471–4. [11] Charlson ME, Pompei P, Ales KL, MacKenzie CR. A new method of classifying prognostic comorbidity in longitudinal studies: development and validation. J Chronic Dis 1987;40:373–83. [12] Deyo RA, Cherkin DC, Ciol MA. Adapting a clinical comorbidity index for use with ICD-9-CM administrative databases. J Clin Epidemiol 1992;45:613–9. [13] Paleri V, Wight RG. A cross-comparison of retrospective notes extraction and combined notes extraction and patient interview in the completion of a comorbidity index (ACE-27) in a cohort of United Kingdom patients with head and neck cancer. J Laryngol Otol 2002;116:937–41. [14] Paleri V, Wight RG. Applicability of the adult comorbidity evaluation - 27 and the Charlson indexes to assess comorbidity by notes extraction in a cohort of United Kingdom patients with head and neck cancer: a retrospective study. J Laryngol Otol 2002;116:200–5. [15] Singh B, Bhaya M, Zimbler M, et al. Impact of comorbidity on outcome of young patients with head and neck squamous cell carcinoma. Head Neck 1998;20:1–7. [16] Piccirillo JF, Tierney RM, Costas I, et al. Prognostic importance of comorbidity in a hospital-based cancer registry. JAMA 2004;291:2441–7. [17] Zhao C, Han FLuLX, et al. Ai Zheng 2004;23:1532–7. [18] Li WF, Sun Y, Chen M, et al. Locoregional extension patterns of nasopharyngeal carcinoma and suggestions for clinical target volume delineation. Chin J Cancer 2012;31:579–87. [19] Liang SB, Sun Y, Liu LZ, et al. Extension of local disease in nasopharyngeal carcinoma detected by magnetic resonance imaging: improvement of clinical target volume delineation. Int J Radiat Oncol Biol Phys 2009;75:742–50. [20] Yang HI, Yuen MF, Chan HL, et al. Risk estimation for hepatocellular carcinoma in chronic hepatitis B (REACH-B): development and validation of a predictive score. Lancet Oncol 2011;12:568–74.
[21] Liu CT, Chiu TJ, Huang TL, et al. Impact of comorbidity on survival for locally advanced head and neck cancer patients treated by radiotherapy or radiotherapy plus chemotherapy. Chang Gung Med J 2010;33:283–91. [22] Datema FR, Ferrier MB, van der Schroeff MP, de Jong RJ Baatenburg. Impact of comorbidity on short-term mortality and overall survival of head and neck cancer patients. Head Neck 2010;32:728–36. [23] Boje CR, Dalton SO, Gronborg TK, et al. The impact of comorbidity on outcome in 12 623 Danish head and neck cancer patients: a population based study from the DAHANCA database. Acta Oncol 2013;52:285–93. [24] Hall SF, Groome PA, Irish J, O’Sullivan B. Towards further understanding of prognostic factors for head and neck cancer patients: the example of hypopharyngeal cancer. Laryngoscope 2009;119:696–702. [25] Homma A, Sakashita T, Oridate N, et al. Importance of comorbidity in hypopharyngeal cancer. Head Neck 2010;32:148–53. [26] Te HS, Jensen DM. Epidemiology of hepatitis B and C viruses: a global overview. Clin Liver Dis 2010;14:1–21. vii. [27] Piccirillo JF, Vlahiotis A, Barrett LB, et al. The changing prevalence of comorbidity across the age spectrum. Crit Rev Oncol Hematol 2008;67:124–32. [28] Reid BC, Alberg AJ, Klassen AC, et al. Comorbidity and survival of elderly head and neck carcinoma patients. Cancer 2001;92:2109–16. [29] Sanabria A, Carvalho AL, Vartanian JG, et al. Comorbidity is a prognostic factor in elderly patients with head and neck cancer. Ann Surg Oncol 2007;14:1449–57. [30] Chen AY, Matson LK, Roberts D, Goepfert H. The significance of comorbidity in advanced laryngeal cancer. Head Neck 2001;23:566–72. [31] Piccirillo JF, Vlahiotis A. Comorbidity in patients with cancer of the head and neck: prevalence and impact on treatment and prognosis. Curr Oncol Rep 2006;8:123–9. [32] Piccirillo JF. Importance of comorbidity in head and neck cancer. Laryngoscope 2000;110:593–602. [33] Liu X, Li X, Jiang N, et al. Prognostic value of chronic hepatitis B virus infection in patients with nasopharyngeal carcinoma: Analysis of 1301 patients from an Endemic Area in China. Cancer 2013. [34] Peng G, Li S, Wu W, et al. PD-1 upregulation is associated with HBV-specific T cell dysfunction in chronic hepatitis B patients. Mol Immunol 2008;45:963–70. [35] Oliviero B, Cerino A, Varchetta S, et al. Enhanced B-cell differentiation and reduced proliferative capacity in chronic hepatitis C and chronic hepatitis B virus infections. J Hepatol 2011;55:53–60. [36] Read WL, Tierney RM, Page NC, et al. Differential prognostic impact of comorbidity. J Clin Oncol 2004;22:3099–103.
Please cite this article in press as: Guo R et al. Comorbidity predicts poor prognosis in nasopharyngeal carcinoma: Development and validation of a predictive score model. Radiother Oncol (2015), http://dx.doi.org/10.1016/j.radonc.2014.12.002
Contents lists available at ScienceDirect
Radiotherapy and Oncology journal homepage: www.thegreenjournal.com
Original article
Comorbidity predicts poor prognosis in nasopharyngeal carcinoma: Development and validation of a predictive score model Rui Guo a,1, Xiao-Zhong Chen b,1, Lei Chen a, Feng Jiang b, Ling-Long Tang a, Yan-Ping Mao a, Guan-Qun Zhou a, Wen-Fei Li a, Li-Zhi Liu c, Li Tian c, Ai-Hua Lin d, Jun Ma a,⇑ a
Department of Radiation Oncology, Sun Yat-sen University Cancer Center, State Key Laboratory of Oncology in Southern China, Collaborative Innovation Center for Cancer Medicine; Department of Radiation Oncology, Zhejiang Cancer Hospital; c Imaging Diagnosis and Interventional Center, Sun Yat-sen University Cancer Center, State Key Laboratory of Oncology in Southern China, Collaborative Innovation Center for Cancer Medicine; and d Department of Medical Statistics and Epidemiology, School of Public Health, Sun Yat-sen University, People’s Republic of China
b
a r t i c l e
i n f o
Article history: Received 4 November 2014 Received in revised form 5 December 2014 Available online xxxx Keywords: Nasopharyngeal carcinoma Comorbidity Prognosis Survival Multi-cancer centre
a b s t r a c t Background and purpose: The impact of comorbidity on prognosis in nasopharyngeal carcinoma (NPC) is poorly characterized. Material and methods: Using the Adult Comorbidity Evaluation-27 (ACE-27) system, we assessed the prognostic value of comorbidity and developed, validated and confirmed a predictive score model in a training set (n = 658), internal validation set (n = 658) and independent set (n = 652) using area under the receiver operating curve analysis. Results: Comorbidity was present in 40.4% of 1968 patients (mild, 30.1%; moderate, 9.1%; severe, 1.2%). Compared to an ACE-27 score 61, patients with an ACE-27 score >1 in the training set had shorter overall survival (OS) and disease-free survival (DFS) (both P < 0.001), similar results were obtained in the other sets (P < 0.05). In multivariate analysis, ACE-27 score was a significant independent prognostic factor for OS and DFS. The combined risk score model including ACE-27 had superior prognostic value to TNM stage alone in the internal validation set (0.70 vs. 0.66; P = 0.02), independent set (0.73 vs. 0.67; P = 0.002) and all patients (0.71 vs. 0.67; P < 0.001). Conclusions: Comorbidity significantly affects prognosis, especially in stages II and III, and should be incorporated into the TNM staging system for NPC. Assessment of comorbidity may improve outcome prediction and help tailor individualized treatment. Ó 2014 Elsevier Ireland Ltd. All rights reserved. Radiotherapy and Oncology xxx (2015) xxx–xxx
Nasopharyngeal carcinoma (NPC) has a distinct epidemiology and geographic distribution. Southern China including Hong Kong has the highest incidence of NPC, with an annual incidence of 20– 50 cases per 100,000 [1]. Currently, the extent of disease, as represented by the tumour-node-metastasis (TNM) staging system, is the main tool for determining clinical treatment strategies and predicting treatment outcome [2]. However, recent findings suggest that the current AJCC staging system is not satisfactorily accurate in terms of survival prediction [3]. Therefore, the identification of novel prognostic factors is of great importance for improving outcome prediction and tailoring individualized treatments in patients with NPC.
⇑ Corresponding author at: State Key Laboratory of Oncology in South China, Department of Radiation Oncology, Sun Yat-sen University Cancer Center, 651 Dongfeng Road East, Guangzhou 510060, People’s Republic of China. E-mail address: [email protected] (J. Ma). 1 Rui Guo and Xiao-Zhong Chen contributed equally to this article.
Comorbidity refers to the presence of one or more co-existing conditions that are not caused by the primary tumour [4]. A relationship between the severity of comorbidity and treatment outcome has been observed in several other types of cancer, including head and neck, breast, prostate and colon cancers [5– 7]. The existence of comorbidity in a patient can affect treatment decisions in daily clinical practice, increase the chances of complications after treatment and, consequently, affect their final outcome [8]. Therefore, the evaluation of comorbidity as a component of prognostic risk assessment may drive improvements in the treatment and survival of patients with NPC. NPC has distinct epidemiology, pathology and clinical attributes compared to other head and neck cancers. To date, only one study has investigated the effect of comorbidity in NPC, and reported that comorbidity did not affect prognosis independently of the TNM staging system in 59 patients with NPC [9]. The precise role of comorbidity in NPC needs to be satisfactorily investigated in order to enable comprehensive risk evaluation before treatment. Additionally, there is no information available on whether comorbidity
http://dx.doi.org/10.1016/j.radonc.2014.12.002 0167-8140/Ó 2014 Elsevier Ireland Ltd. All rights reserved.
Please cite this article in press as: Guo R et al. Comorbidity predicts poor prognosis in nasopharyngeal carcinoma: Development and validation of a predictive score model. Radiother Oncol (2015), http://dx.doi.org/10.1016/j.radonc.2014.12.002
2
Impact of comorbidity on prognosis in NPC
could be reliably incorporated into the TNM staging system for patients with NPC. On this basis, we conducted a retrospective study to obtain insights into the effect of comorbidity on the survival of patients with NPC and investigated whether combining information on comorbidity with prognostic staging systems could improve our ability to assess the prognosis of patients with NPC. Patients and methods Patient characteristics and study design A total of 2055 patients with newly-diagnosed, non-metastatic, histologically-proven NPC who were treated at Sun Yat-sen University Cancer Center (Guangzhou, People’s Republic of China) and Zhejiang Cancer Hospital between February 2003 and January 2007 were retrospectively reviewed. Of these, 87 cases were excluded due to a lack of substantial information regarding comorbidity; the remaining 1968 cases were included in the current study. This retrospective study was approved by the Institutional Review Boards of Sun Yat-sen University Cancer Center and Zhejiang Cancer Hospital. All patients underwent pretreatment evaluations, including a complete medical history, physical examination, hematology and biochemistry profiles, coagulation test, electrocardiogram (ECG), infectious disease examination (Hepatitis, HIV, syphilis), magnetic resonance imaging (MRI) of the neck and nasopharynx, chest radiography, abdominal sonography and whole-body bone scan using single-photon emission computed tomography (ECT); 205 (10.4%) patients also underwent positron emission tomography CT (PETCT). All patients were restaged according to the 7th edition of the International Union against Cancer/American Joint Committee on Cancer (UICC/AJCC) system [10].
Follow-up and statistical analysis Patients were followed up every 3 months during the first 2 years, and every 6 months thereafter until death. The primary end point was overall survival (OS), defined from the first day of therapy to death due to any cause. The secondary end points included disease-free survival (DFS), defined as the time from the first day of therapy to progression or death, whichever occurred first. We randomly divided the 1316 patients from Sun Yat-sen University into a training set of 658 cases and a validation set of 658 cases using a computer-generated allocation sequence. To assess whether the comorbidities identified in these sets had similar prognostic values in different populations, we used the 652 cases from Zhejiang cancer hospital as an independent set to externally validate the results. We assessed the relationships between clinical characteristics and ACE-27 score using the v2 test or Fisher’s exact test. Actuarial rates were calculated using the Kaplan–Meier method and differences were compared using the log-rank test. Univariate and multivariate analyses using a Cox proportional hazards model were used to test the independent significance of different factors by backward elimination. To develop a more sensitive predictive tool, we constructed a prognostic score model by combining the independent prognostic factors identified in the training set (see appendix for details). The regression coefficient of each independent prognostic factor in the proportional hazards model was divided by the regression coefficient of the ACE-27, and rounded into an integer value to generate the risk score.[20]. We compared the prognostic validity of this model with TNM staging alone using receiver operating curve (ROC) analysis in the validation set and independent set. Stata Statistical Package (STATA 12; StataCorp LP, College Station, Texas, USA) was used for all analyses. Results
Adult Comorbidity Evaluation-27 system Comorbidity was assessed using the Adult Comorbidity Evaluation-27 (ACE-27) system, which was developed for patients with cancer, specifically head and neck cancer [11–15]. The ACE-27 system is a validated 27-item comorbidity index [16], by which specific conditions are graded into three degrees (grade 0 = none, grade 1 = minimal, grade 2 = moderate, grade 3 = severe) based on organ system decompensation. The overall comorbidity score is assigned based on the highest-ranking single ailment. When two or more moderate ailments occurred in different organ systems, the overall comorbidity score was graded as severe. The ACE-27 scores were manually calculated by one physician who reviewed detailed information on the patients’ baseline medical condition and comorbidities. Treatment All patients were treated with definitive radiation therapy. Details of the radiotherapy techniques applied at Sun Yat-sen University Cancer Center have been reported previously [17–19]. Among the 1968 patients, 1238 (62.9%) underwent two-dimensional radiotherapy (2D-CRT) and 730 (37.1%) underwent threedimensional conformal radiotherapy (3-DCRT) or intensity-modulated radiotherapy (IMRT). During the study period, as per institutional guidelines, the majority (964/1312; 73.5%) of patients with stage III or IV NPC (classified as T3-T4 and/or N2-N3 disease) underwent chemotherapy. Inductive chemotherapy (IC) or adjuvant chemotherapy (AC) consisted of cisplatin with 5-fluorouracil or cisplatin with taxoids every 3 weeks for two to three cycles. Concomitant chemotherapy consisted of cisplatin at weeks 1, 4 and 7 of radiotherapy or weekly cisplatin.
The median duration of follow-up for the entire patient group was 64.4 months (range, 2.7–126.5 months). A total of 222 (11.3%) patients developed local failure or regional failure, 271 (13.8%) developed distant metastases, and 429 (21.8%) patients died. The 5-year DFS and OS rates for the entire cohort were 71.2% and 79.1%, respectively. The overall distribution of patients with stages I, II, III, and IVAB disease was 7.1%, 26.2%, 39.8%, and 26.9%, respectively. The clinical characteristics of the patients in the training, validation and independent sets are shown in Table 1. Prevalence, type and severity of comorbidity Review of the patients’ medical records showed that a total of 795 patients had comorbidities (40.4%). Of the 795 patients, 592 (30.1%) had an ACE-27 score of 1; 179 (9.1%), an ACE-27 score of 2; and 24 (1.2%), an ACE-27 score of 3. Gastrointestinal disease, substance abuse and cardiovascular disease were most common, with an incidence of 19.4%, 16.1% and 7.1%, respectively. No patients were diagnosed with acquired immune deficiency syndrome or human immunodeficiency virus infection. Data on the prevalence and severity of the comorbidities in the training, validation and independent sets are presented in Table 2. There was no correlation between disease stage or histology and comorbidity burden in the training, validation and independent sets. Not surprisingly, comorbidity varied significantly by age and gender (P < 0.05; Table 1). Prognostic value of comorbidity In the training set (n = 658), the 5-year OS and DFS rates for patients with an ACE-27 score 61 vs. > 1 were 81.1% vs. 59.3%
Please cite this article in press as: Guo R et al. Comorbidity predicts poor prognosis in nasopharyngeal carcinoma: Development and validation of a predictive score model. Radiother Oncol (2015), http://dx.doi.org/10.1016/j.radonc.2014.12.002
3
R. Guo et al. / Radiotherapy and Oncology xxx (2015) xxx–xxx Table 1 Characteristics of the 1968 patients with nasopharyngeal carcinoma. Characteristic
Training set (n = 658) ACE-27 6 1
Age (y) 650 >50 Gender Male Female Histology WHO type I WHO type II/III T classification* T1 T2 T3 T4 N classification* N0 N1 N2 N3 Clinical stage* I II III IVA-B OS 5-year survival DFS 5-year survival
Validation set (n = 658)
ACE-27 > 1
P
ACE-27 6 1
ACE-27 > 1
0.011 426 (70.2) 181 (29.8)
27 (52.9) 24 (47.1)
454 (74.8) 153 (25.2)
48 (94.1) 3 (5.9)
4 (0.7) 603 (99.3)
0 (0) 51 (100.0)
135 142 197 133
(22.2) (23.4) (32.5) (21.9)
10 (19.6) 7 (13.7) 19 (37.3) 15 (29.4)
179 (29.5) 274 (45.1) 105 (17.3) 49 (8.1)
10 (19.6) 26 (51.0) 8 (15.7) 7 (13.7)
53 (8.7) 156 (25.7) 228 (37.6) 170 (28.0)
3 (5.9) 10 (19.6) 16 (31.4) 22 (43.1)
81.1
59.3
74.1
43.9
Independent set (n = 652) P
ACE-27 6 1
ACE-27 > 1
320 (56.9) 242 (43.1)
38 (42.2) 52 (57.8)
383 (68.1) 179 (31.9)
82 (91.1) 8 (8.9)
3 (0.5) 559 (99.5)
0 (0) 90 (100)
88 (15.7) 267 (47.5) 125 (22.2) 82 (14.6)
17 44 18 11
(18.9) (48.9) (20.0) (12.2)
116 (20.6) 206 (36.7) 192 (34.2) 48 (8.5)
20 28 28 14
(22.2) (31.1) (31.1) (15.6)
21 (3.7) 159 (28.3) 259 (46.1) 123 (21.9)
4 (4.4) 23 (25.6) 38 (42.2) 25 (27.8)
78.1
62.1
70.5
49.8
0.008 432 (72.5) 164 (27.5)
35 (56.5) 27 (43.5)
428 (71.8) 168 (28.2)
58 (93.5) 4 (6.5)
3 (0.5) 593 (99.5)
0 (0) 62 (100)
140 135 193 128
10 15 21 16
<0.001
0.009
<0.001
0.561
<0.001
0.576
0.308
0.669
0.591 (23.5) (22.7) (32.4) (21.5)
(16.1) (24.2) (33.9) (25.8)
0.503
0.797
0.765 185 (31.0) 252 (42.3) 111 (18.6) 48 (8.1)
16 (25.8) 26 (41.9) 14 (22.6) 6 (9.7)
55 (9.2) 155 (26.0) 219 (36.7) 167 (28.0)
4 (6.5) 13 (21.0) 23 (37.1) 22 (35.3)
82.6
66.4
74.6
55.3
0.150
0.173
0.545
<0.001
0.626
0.001
<0.001
P
<0.001
0.001
<0.001
RT, Radiotherapy; 2-DRT, two-dimensional radiotherapy; 3-DCRT, three-dimensional conformal radiotherapy; IMRT, intensity-modulated radiotherapy; ACE-27, Adult Comorbidity Evaluation-27. * According to the 7th AJCC/UICC staging system.
Table 2 Presence and severity of comorbidities in the study population training set, validation set and independent set. Disease classification
Training set n = 658
ACE-27 score
Grade 1
Grade 2
Grade 3
Validation set n = 658 Grade 1
Grade 2
Grade 3
Grade 1
Grade 2
Grade 3
Gastrointestinal system Substance abuse Cardiovascular system Respiratory system Renal system Endocrine system Neurological system Psychiatric Rheumatologic Immunological system Malignancy Body weight Overall ACE-27 score
164 (24.9) 48 (7.3) 31 (4.7) 14 (2.1) 18 (2.7) 11 (1.7) 2 (0.3) 1 (0.2) 0 (0.0) 0 (0.0) 2 (0.3) 0 (0.0) 237 (36.0)
5 (0.8) 39 (5.9) 1 (0.2) 2 (0.3) 0 (0.0) 1 (0.2) 0 (0.0) 0 (0.0) 0 (0.0) 0 (0.0) 3 (0.5) 1 (0.2) 48 (7.3)
0 0 0 0 0 0 0 0 0 0 1 0 3
150 (22.8) 41 (6.2) 35 (5.3) 11 (1.7) 16 (2.4) 9 (1.4) 4 (0.6) 0 (0.0) 2 (0.3) 0 (0.0) 2 (0.3) 1 (0.2) 204 (31.0)
6 (0.9) 41 (6.2) 6 (0.9) 2 (0.3) 0 (0.0) 1 (0.2) 0 (0.0) 0 (0.0) 0 (0.0) 0 (0.0) 2 (0.3) 0 (0.0) 58 (8.8)
0 0 0 0 0 0 0 0 0 0 4 0 4
41 (6.3) 103 (15.8) 48 (7.4) 4 (0.6) 1 (0.2) 7 (1.1) 0 (0.0) 0 (0.0) 2 (0.3) 0 (0.0) 0 (0.0) 0 (0.0) 150 (23.0)
15 (2.3) 43 (6.6) 13 (2.0) 4 (0.6) 0 (0.0) 6 (0.9) 1 (0.2) 1 (0.2) 1 (0.2) 0 (0.0) 3 (0.5) 0 (0.0) 73 (11.2)
2 (0.3) 0 (0.0) 5 (0.8) 1 (0.2) 1 (0.2) 0 (0.0) 1 (0.2) 0 (0.0) 0 (0.0) 0 (0.0) 2 (0.3) 0 (0.0) 17 (2.6)
(0.0) (0.0) (0.0) (0.1) (0.0) (0.0) (0.0) (0.0) (0.0) (0.0) (0.2) (0.0) (0.5)
Independent set n = 652
(0.0) (0.0) (0.0) (0.0) (0.0) (0.0) (0.0) (0.0) (0.0) (0.0) (0.6) (0.0) (0.6)
ACE-27, Adult Comorbidity Evaluation-27.
(P < 0.001) and 74.1% vs. 43.9% (P < 0.001), respectively (Fig. 1A, 1B). In the validation set (n = 658), the 5-year OS and DFS rates for patients with an ACE-27 score 61 vs. > 1 were 82.6% vs. 66.4% (P = 0.001) and 74.6% vs. 55.3% (P = 0.001) (Fig. 1C, 1D). In the independent set (n = 652), the 5-year OS and DFS rates for patients with an ACE-27 score 61 vs. >1 were 78.1.6% vs. 62.1% (P < 0.001) and 70.5% vs. 49.8% (P < 0.001) (Fig. 1E, 1F). Overall, patients with an ACE-27 score >1 had poorer OS and DFS than patients with an ACE-27 score 61. In the multivariate analysis, ACE-27 score was a significant, independent prognostic factor for OS (hazard ratio [HR] = 2.027) and DFS (HR = 2.365) in the training set, OS (HR = 1.660) and DFS (HR = 1.674) in the validation set, and OS (HR = 1.829) and DFS (HR = 1.826) in the independent set (all P < 0.05; Table 3).
Varying impact of comorbidity by stage In order to analyse the prognostic value of comorbidity in patients with the same stage of disease, all patients (n = 1968) with NPC were stratified according to TNM stage. In stage I, the 5-year OS and DFS rates for patients with an ACE-27 score 61 vs. >1 were 98.3% vs. 87.5% (P = 0.084) and 95.9% vs. 78.8% (P = 0.079), respectively. In stage II, the 5-year OS and DFS rates for patients with an ACE-27 score 61 vs. >1 were 90.1% vs. 77.1% (P = 0.002) and 84.0% vs. 63.8% (P < 0.001), respectively. In stage III, the 5-year OS and DFS rates for patients with an ACE-27 score 61 vs. >1 were 83.3% vs. 55.3% (P < 0.001) and 75.1% vs. 41.9% (P < 0.001), respectively. In stage IV, the 5-year OS and DFS rates for patients with an ACE-27 score 61 vs. >1 were 62.5% vs. 54.4% (P = 0.107) and
Please cite this article in press as: Guo R et al. Comorbidity predicts poor prognosis in nasopharyngeal carcinoma: Development and validation of a predictive score model. Radiother Oncol (2015), http://dx.doi.org/10.1016/j.radonc.2014.12.002
4
Impact of comorbidity on prognosis in NPC
Table 3 Multivariate analysis of the impact of all variables on survival in the training, validation and independent sets. Patients
Endpoint
Variable
HR
95% CI
P-Valueb
Training set
OS
ACE-27 T classificationa N classificationa Age
2.027 2.147 2.363 1.021
1.259–3.265 1.500–3.073 1.690–3.305 1.006–1.035
0.004 <0.001 <0.001 0.006
DFS
ACE-27 T classificationa N classificationa
2.365 1.760 2.161
1.574–3.554 1.307–2.370 1.610–2.900
<0.001 <0.001 <0.001
Age ACE-27 T classificationa N classificationa Age
1.012 1.660 2.155 1.920 1.031
0.999–1.024 1.043–2.643 1.481–3.136 1.355–2.719 1.016–1.047
0.069 0.033 <0.001 <0.001 <0.001
DFS
ACE-27 T classificationa N classificationa Age
1.674 1.996 1.628 1.016
1.122–2.498 1.461–2.726 1.204–2.200 1.003–1.029
0.012 <0.001 0.002 0.017
OS
ACE-27 T classificationa N classificationa Age
1.829 2.007 1.591 1.040
1.240–2.699 1.445–2.788 1.146–2.209 1.025–1.055
0.002 <0.001 0.006 <0.001
DFS
ACE-27 T classificationa N classificationa Age
1.826 2.009 1.623 1.026
1.304–2.556 1.518–2.660 1.227–2.146 1.013–1.038
<0.001 <0.001 0.001 <0.001
Validation set
Independent set
OS
OS, overall survival; DFS, disease-free survival; ACE-27, Adult Comorbidity Evaluation-27: CI, confidence interval; HR, hazard ratio. a According to the 7th AJCC/UICC staging system. b Multivariate P-values were calculated using an adjusted Cox proportional-hazards model. The following parameters were included in the Cox proportion hazard model by backward elimination: age (continuous variable), gender (male vs. female), T classification (T3–4 vs. T1–2), N classification (N2–3 vs. N0–1), use of chemotherapy (with vs. without) and ACE-27 (grade >1 vs. grade 61).
53.8% vs. 44.2% (P = 0.132), respectively. Overall, the comorbid effect was most evident in stage II–III, and insignificant in stage I or IV.
Prognostic value of adding the ACE-27 score to the TNM staging system The three significantly independent prognostic variables including ACE-27 grade, age and TNM stage were used to construct a prognostic-score model. A cumulative risk score then was calculated for each patient in the training set, the validation set, the sum of the points ranged from 0 to 7. Then the patients were divided into four risk groups (Group 1: risk score 0–1; Group 2: risk score 2–4; Group 3: risk score 5; Group 4: risk score 6–7) according to their overall survival in the training set (Supplementary materials). In the validation set, the OS rates for Groups 1, 2, 3, and 4 were 98.1%, 88.5%, 74.2%, and 54.3% (P < 0.001) and the DFS rates were 96.4%, 81.4%, 61.1%, and 48.2% (P < 0.001), respectively (Fig. 2A, 2B). In the independent set, the OS rates for Groups 1, 2, 3, and 4 were 100%, 88.5%, 67.5%, and 44.0% (P < 0.001) and the DFS rates were 100%, 80.0%, 60.1%, and 33.7% (P < 0.001), respectively (Fig. 2C, 2D). For all patients, the OS rates for Groups 1, 2, 3, and 4 were 98.4%, 88.5%, 71.6%, and 48.1% (P < 0.001) and the DFS rates were 95.3%, 81.1%, 61.6%, and 41.1% (P < 0.001), respectively (Fig. 2E, 2F). For OS, the area under the ROC curves (AUCs) for the TNM stage model alone and combined risk grouping were 0.66 vs. 0.70 (P = 0.02) in the validation set (Fig. 3A), 0.67 vs. 0.73 (P = 0.002) in the independent set (Fig. 3B), and 0.67 vs. 0.71 (P < 0.001) for overall patients (Fig. 3C). Therefore, the combined risk grouping created by adding ACE-27 scores to TNM staging was superior to TNM staging alone in terms of predicting survival outcomes.
Discussion We explored the potential prognostic value of comorbidity in patients with NPC from two cancer centres in different areas of China, and to our knowledge, is the first such investigation in a large cohort of patients with NPC. This study provides strong evidence in support of comorbidity as an independent prognostic factor in patients with NPC. Prevalence, type and severity of comorbidity Previous studies reported the incidence of comorbidity in head and neck cancer to be 33–65%, with cardiovascular and pulmonary diseases being the most common [7,21–25]. Ramakrishnan et al. reported an incidence of comorbidity of 44% in 59 patients with NPC, with 19% patients having moderate or severe comorbidity and the cardiovascular system (27%) most commonly affected [9]. The comorbidity rate in this study was 40.4%, which is comparable to the rates reported previously. However, the most common comorbidities differed in the patients from different areas: the most common comorbidities in the training set and validation set were comparable, but were different to that of the independent set. The patients in the training set and validation set are from southern China, and the most common comorbidities in these groups were gastrointestinal diseases, with liver disease the leading comorbidity; these results are similar to the findings of another study [21]. This is not surprising, as epidemiological research has demonstrated that southern China has one of the highest incidences of chronic hepatitis B virus (HBV) infection, and the hepatitis B surface antigen (HBsAg) seropositivity rate in the general population is 10–12% [26]. Therefore, chronic HBV infection may be an important gastrointestinal comorbidity in patients with NPC in southern China.
Please cite this article in press as: Guo R et al. Comorbidity predicts poor prognosis in nasopharyngeal carcinoma: Development and validation of a predictive score model. Radiother Oncol (2015), http://dx.doi.org/10.1016/j.radonc.2014.12.002
R. Guo et al. / Radiotherapy and Oncology xxx (2015) xxx–xxx
5
Fig. 1. Kaplan–Meier estimates of overall survival (OS) and disease-free survival (DFS) for patients with nasopharyngeal cancer according to ACE-27 score. OS (A) and DFS (B) for the 658 patients in the training set, OS (C) and DFS (D) for the 658 patients in the validation set, and OS (E) and DFS (F) for the 652 patients in the independent set.
The incidence of comorbidity increased with increase in age in this study. Piccirillo et al. also investigated the prevalence of comorbidity across the age spectrum, and found increased age to be associated with an increase in the number and severity of comorbidities [27]. Similar results have been reported by other studies, in that comorbidity was present in younger patients, but the incidence of comorbidity increased with age [9,21,28,29]. Comorbidity is an independent prognostic factor in NPC Although the incidence of comorbidity and the most common comorbidities varied in different areas (training, validation and
independent sets), comorbidity significantly affected the survival of the patients with NPC in all sets, as a higher ACE-27 score was associated with poorer survival outcomes. The association between comorbidity and poorer overall survival may mainly be due to the physiological burden of chronic disease; however, the precise impact of comorbidity on disease-free survival remains unclear. It has been reported that comorbidity did not affect cancer-specific survival in patients treated with radiotherapy only [29,30]; however, diverging results were observed in patients treated with chemoradiotherapy and surgery [13,21]. In this study, the majority (964/1312; 73.5%) of patients with stage III or IV NPC (classified as T3–T4 and/or N2–N3 disease) underwent chemoradiotherapy.
Please cite this article in press as: Guo R et al. Comorbidity predicts poor prognosis in nasopharyngeal carcinoma: Development and validation of a predictive score model. Radiother Oncol (2015), http://dx.doi.org/10.1016/j.radonc.2014.12.002
6
Impact of comorbidity on prognosis in NPC
Fig. 2. Kaplan–Meier estimates of overall survival (OS) and disease-free survival (DFS) using the combined risk grouping in patients with nasopharyngeal cancer. OS (A) and DFS (B) for the 658 patients in the validation set, OS (C) and DFS (D) for the 652 patients in the independent set, and OS (E) and DFS (F) for all 1968 patients. Group 1: risk score 0–1; Group 2: risk score 2–4; Group 3: risk score 5; Group 4: risk score 6–7.
The reasons explaining the association between comorbidity and DFS in patients with NPC are unclear. The possible reasons are: firstly, may simply reflect treatment biases on the part of the physician. The presence of comorbidity may necessitate the use of a less than ideal or less radical procedure, which may lead to poorer cancer-specific survival [8,16]. Secondly, an increase in the severity of comorbidities may increase the toxicity of specific treatments, which may consequently shorten the survival period to the extent of canceling out any possible gain obtained from the therapy [16,31,32]. Furthermore, the most frequently observed comorbidity was HBV infection-associated liver disease, which is similar to the finding of a previous study in which HBV infection-
associated liver disease was reported to negatively affect cancerspecific survival [33]. The latent mechanism responsible for the unfavourable prognosis of HBV-infected patients with NPC may be linked to certain types of immunological dysfunction, as HBV is associated with immune dysfunction, as demonstrated by hepatitis B-related nephritis and its association with lymphoma [34,35]. However, this hypothesis needs to be addressed in future studies. Varying impact of comorbidity by TNM stage Apart from comorbidity, TNM stage remained a significantly predictive covariate in the Cox regression model. However, the
Please cite this article in press as: Guo R et al. Comorbidity predicts poor prognosis in nasopharyngeal carcinoma: Development and validation of a predictive score model. Radiother Oncol (2015), http://dx.doi.org/10.1016/j.radonc.2014.12.002
R. Guo et al. / Radiotherapy and Oncology xxx (2015) xxx–xxx
7
Fig. 3. Sensitivity and specificity for the prediction of overall survival for the combined risk grouping (including the ACE-27 score) and the TNM stage alone model for (A) the 658 patients in the validation set, (B) 652 patients in the independent set and (C) all 1968 patients. AUC: area under the curve; ⁄P-values vs. TNM stage alone.
prognosis of patients with the same TNM stage but different severities of comorbidity was significantly different. This suggests that the current staging system is inadequate for defining prognosis and does not adequately reflect the biological heterogeneity of NPC. Furthermore, in the subgroup analyses, the impact of comorbidity on survival outcomes was strongly associated with TNM stage. In patients with stage II–III disease, a high ACE-27 score was significantly associated with outcome, while ACE-27 score had no significant effect on outcome in patients with stage I or IV disease. Our finding that comorbidity has little impact on the survival of patients with aggressive disease (stage IV) is also in accordance with previous studies in other cancers, in which comorbidity had the greatest prognostic impact among groups with the highest survival rate and the least impact in groups with the lowest survival rate [7,36]. Additionally, there was no significant association between comorbidity and either survival outcome in patients with stage I NPC. Firstly, treatment biases cannot exist in patients with stage I, as the treatment strategy for this group was radiotherapy only. Secondly, the failure rate for stage I is very low. However, the assessment of comorbidity and treatment selection appear to be particularly important in patients with stage II–III NPC, as the disease is less aggressive or less widely spread in these patients, and thus they have a better overall prognosis than patients with more advanced disease. Therefore, comorbidity interventions and more meticulous disease follow-up should be provided for patients with stage II–III disease in order to obtain maximum improvement in survival rates. Prognostic value of adding ACE-27 score to TNM staging In addition to comorbidity, T stage and N stage remained significant predictive covariates in the Cox regression model. Another major objective of this study was to verify the combined prognostic value of comorbidity and disease stage. We constructed a prognostic score model by combining the independent prognostic factors identified in the training set (Supplementary appendix), and compared its prognostic validity with TNM stage alone. ROC analysis of the validation and independent sets suggested that prognostic assessment could be improved by combining the ACE27 system with TNM staging. The ACE-27 system requires no special technological expertise for clinical use and can be applied by any healthcare professional using information that is readily available from a patient’s history or pretreatment evaluations. Therefore, a comorbidity index such as the ACE-27 system should be taken into account along with the clinical TNM stage when treating patients with NPC. Moreover, patients with comorbidities should be offered appropriate supportive care in order to ensure optimal cancer treatment; and such an approach necessitates a multidisciplinary approach to patient care. This study was a retrospective chart-based assessment; therefore, a selection bias may exist in the collection of comorbidity data. However, the results of this study indicate that comorbidity
may vary within patients with NPC from different areas, and highlights the significant impact of comorbidity on survival in patients with NPC. The results of this study also justify the need for collecting prospective comorbidity data in routine clinical practice, and indicate that a comorbidity index should be integrated into the current TNM staging system to improve treatment selection and outcome prediction. Conclusion For patients with comorbidity, the ACE-27 score is a significant predictor of prognosis in NPC. Assessment of comorbidity before treatment may help to improve outcome prediction and tailor individualized treatments for patients with NPC. Information on comorbidity could be added to staging systems and be incorporated into treatment strategy decision-making, especially for patients with stage II–III NPC. Conflict of interest statement The authors indicated no potential conflicts of interest. Acknowledgements This work was supported by grants from the Innovation Team Development Plan of the Ministry of Education (No. IRT1297), the National Natural Science Foundation of China (Nos. 81230056, 81201746), the Guangdong Province Universities and Colleges Pearl River Scholar Funded Scheme (2010), the Science and Technology Project of Guangzhou (No. 12BppZXaa2060002), the Key Laboratory Construction Project of Guangzhou City, China (No. 121800085), and the Key Scientific and Technological Innovation Program for Universities of Guangdong Province (No. cxzd1005). Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.radonc.2014.12. 002. References [1] Jemal A, Bray F, Center MM, et al. Global cancer statistics. CA Cancer J Clin 2011;61:69–90. [2] Wei WI, Sham JST. Nasopharyngeal carcinoma. Lancet 2005;365:2041–54. [3] Takes RP, Rinaldo A, Silver CE, et al. Future of the TNM classification and staging system in head and neck cancer. Head Neck 2010;32:1693–711. [4] Piccirillo JF, Feinstein AR. Clinical symptoms and comorbidity: significance for the prognostic classification of cancer. Cancer 1996;77:834–42. [5] Kazi R, Nutting CM, Rhys-Evans P, Harrington KJ. Significance and prognostic impact of co-morbidity in head and neck cancer. J Cancer Res Ther 2009;5:145–7.
Please cite this article in press as: Guo R et al. Comorbidity predicts poor prognosis in nasopharyngeal carcinoma: Development and validation of a predictive score model. Radiother Oncol (2015), http://dx.doi.org/10.1016/j.radonc.2014.12.002
8
Impact of comorbidity on prognosis in NPC
[6] Tanvetyanon T, Padhya T, McCaffrey J, et al. Prognostic factors for survival after salvage reirradiation of head and neck cancer. J Clin Oncol 2009;27:1983–91. [7] Alho OP, Hannula K, Luokkala A, et al. Differential prognostic impact of comorbidity in head and neck cancer. Head Neck 2007;29:913–8. [8] Boje CR. Impact of comorbidity on treatment outcome in head and neck squamous cell carcinoma – a systematic review. Radiother Oncol 2013. [9] Ramakrishnan Y, Paleri V, Shah R, et al. Comorbidity in nasopharyngeal carcinoma: a preliminary communication on the prevalence, descriptive distribution and impact on outcome. Clin Otolaryngol 2007;32: 484–8. [10] Edge SB, Compton CC. The American Joint Committee on Cancer: the 7th edition of the AJCC cancer staging manual and the future of TNM. Ann Surg Oncol 2010;17:1471–4. [11] Charlson ME, Pompei P, Ales KL, MacKenzie CR. A new method of classifying prognostic comorbidity in longitudinal studies: development and validation. J Chronic Dis 1987;40:373–83. [12] Deyo RA, Cherkin DC, Ciol MA. Adapting a clinical comorbidity index for use with ICD-9-CM administrative databases. J Clin Epidemiol 1992;45:613–9. [13] Paleri V, Wight RG. A cross-comparison of retrospective notes extraction and combined notes extraction and patient interview in the completion of a comorbidity index (ACE-27) in a cohort of United Kingdom patients with head and neck cancer. J Laryngol Otol 2002;116:937–41. [14] Paleri V, Wight RG. Applicability of the adult comorbidity evaluation - 27 and the Charlson indexes to assess comorbidity by notes extraction in a cohort of United Kingdom patients with head and neck cancer: a retrospective study. J Laryngol Otol 2002;116:200–5. [15] Singh B, Bhaya M, Zimbler M, et al. Impact of comorbidity on outcome of young patients with head and neck squamous cell carcinoma. Head Neck 1998;20:1–7. [16] Piccirillo JF, Tierney RM, Costas I, et al. Prognostic importance of comorbidity in a hospital-based cancer registry. JAMA 2004;291:2441–7. [17] Zhao C, Han FLuLX, et al. Ai Zheng 2004;23:1532–7. [18] Li WF, Sun Y, Chen M, et al. Locoregional extension patterns of nasopharyngeal carcinoma and suggestions for clinical target volume delineation. Chin J Cancer 2012;31:579–87. [19] Liang SB, Sun Y, Liu LZ, et al. Extension of local disease in nasopharyngeal carcinoma detected by magnetic resonance imaging: improvement of clinical target volume delineation. Int J Radiat Oncol Biol Phys 2009;75:742–50. [20] Yang HI, Yuen MF, Chan HL, et al. Risk estimation for hepatocellular carcinoma in chronic hepatitis B (REACH-B): development and validation of a predictive score. Lancet Oncol 2011;12:568–74.
[21] Liu CT, Chiu TJ, Huang TL, et al. Impact of comorbidity on survival for locally advanced head and neck cancer patients treated by radiotherapy or radiotherapy plus chemotherapy. Chang Gung Med J 2010;33:283–91. [22] Datema FR, Ferrier MB, van der Schroeff MP, de Jong RJ Baatenburg. Impact of comorbidity on short-term mortality and overall survival of head and neck cancer patients. Head Neck 2010;32:728–36. [23] Boje CR, Dalton SO, Gronborg TK, et al. The impact of comorbidity on outcome in 12 623 Danish head and neck cancer patients: a population based study from the DAHANCA database. Acta Oncol 2013;52:285–93. [24] Hall SF, Groome PA, Irish J, O’Sullivan B. Towards further understanding of prognostic factors for head and neck cancer patients: the example of hypopharyngeal cancer. Laryngoscope 2009;119:696–702. [25] Homma A, Sakashita T, Oridate N, et al. Importance of comorbidity in hypopharyngeal cancer. Head Neck 2010;32:148–53. [26] Te HS, Jensen DM. Epidemiology of hepatitis B and C viruses: a global overview. Clin Liver Dis 2010;14:1–21. vii. [27] Piccirillo JF, Vlahiotis A, Barrett LB, et al. The changing prevalence of comorbidity across the age spectrum. Crit Rev Oncol Hematol 2008;67:124–32. [28] Reid BC, Alberg AJ, Klassen AC, et al. Comorbidity and survival of elderly head and neck carcinoma patients. Cancer 2001;92:2109–16. [29] Sanabria A, Carvalho AL, Vartanian JG, et al. Comorbidity is a prognostic factor in elderly patients with head and neck cancer. Ann Surg Oncol 2007;14:1449–57. [30] Chen AY, Matson LK, Roberts D, Goepfert H. The significance of comorbidity in advanced laryngeal cancer. Head Neck 2001;23:566–72. [31] Piccirillo JF, Vlahiotis A. Comorbidity in patients with cancer of the head and neck: prevalence and impact on treatment and prognosis. Curr Oncol Rep 2006;8:123–9. [32] Piccirillo JF. Importance of comorbidity in head and neck cancer. Laryngoscope 2000;110:593–602. [33] Liu X, Li X, Jiang N, et al. Prognostic value of chronic hepatitis B virus infection in patients with nasopharyngeal carcinoma: Analysis of 1301 patients from an Endemic Area in China. Cancer 2013. [34] Peng G, Li S, Wu W, et al. PD-1 upregulation is associated with HBV-specific T cell dysfunction in chronic hepatitis B patients. Mol Immunol 2008;45:963–70. [35] Oliviero B, Cerino A, Varchetta S, et al. Enhanced B-cell differentiation and reduced proliferative capacity in chronic hepatitis C and chronic hepatitis B virus infections. J Hepatol 2011;55:53–60. [36] Read WL, Tierney RM, Page NC, et al. Differential prognostic impact of comorbidity. J Clin Oncol 2004;22:3099–103.
Please cite this article in press as: Guo R et al. Comorbidity predicts poor prognosis in nasopharyngeal carcinoma: Development and validation of a predictive score model. Radiother Oncol (2015), http://dx.doi.org/10.1016/j.radonc.2014.12.002