Pediatric patients include those 21 years or younger; adult patients, older than 21 years.
Pediatric patients include those 21 years or younger.
Customize your JAMA Network experience by selecting one or more topics from the list below.
Richards MK, Dahl JP, Gow K, et al. Factors Associated With Mortality in Pediatric vs Adult Nasopharyngeal Carcinoma. JAMA Otolaryngol Head Neck Surg. 2016;142(3):217–222. doi:https://doi.org/10.1001/jamaoto.2015.3217
Nasopharyngeal carcinoma (NPC) is endemic in some Asian regions but is uncommon in the United States. Little is known about the racial, demographic, and biological characteristics of the disease in pediatric patients.
To improve understanding of the differences between pediatric and adult NPC and to determine whether race conferred a survival difference among pediatric patients with NPC.
Design, Setting, and Participants
This retrospective cohort study included all 17 317 patients with a primary diagnosis of NCP in the National Cancer Data Base from January 1, 1998, to December 31, 2011. Of these, 699 patients were 21 years or younger (pediatric); 16 618 patients, older than 21 years (adult). Data were analyzed after data collection.
Pediatric age at diagnosis of NPC.
Main Outcomes and Measures
Demographic, tumor, and treatment characteristics of pediatric patients with NPC were compared with those of adults using the χ2 test for categorical variables. An adjusted Cox proportional hazards regression model was used to examine survival differences in pediatric patients relative to adult patients. In addition, the risk for pediatric mortality by race was estimated.
Of the 17 317 patients, a total of 699 pediatric and 16 618 adult patients were identified with a primary diagnosis of NPC (female, 239 pediatric patients [34.2%] and 5153 adult patients [32.4%]). Pediatric patients were most commonly black (299 of 686 [43.6%]), whereas adults were most likely to be non-Hispanic white (9839 of 16 504 [60.0%]; P < .001). Pediatric patients were less likely to be Asian (39 of 686 [5.7%]) than were adults (3226 of 16 405 [19.7%]; P < .001). Pediatric patients were more likely to have regional nodal evaluation and to present with stage IV disease (227 of 643 [35.3%] and 330 of 565 [58.4%], respectively) than were adult patients (3748 of 15 631 [24.0%] and 6553 of 13 721 [47.8%], respectively; P < .001 for both comparisons). Pediatric patients had a lower risk for mortality relative to adults (hazard ratio, 0.37; 95% CI, 0.25-0.56). No difference in mortality by racial group was found among pediatric patients (hazard ratio, 1.10; 95% CI, 0.82-1.40).
Conclusions and Relevance
Pediatric patients with NPC were more commonly black and presented more frequently with stage IV disease. Pediatric patients had a decreased mortality risk relative to adults, even after adjusting for covariables. Asian race was not associated with increased mortality in pediatric patients with NPC. Racial differences are not associated with an increased risk for mortality among pediatric patients.
Nasopharyngeal carcinoma (NPC) is a rare malignant tumor derived from the epithelium lining the nasopharynx. Quiz Ref IDThe World Health Organization (WHO) classification system for NPC characterizes the tumors based on the following 3 primary histopathologic subtypes: keratinizing squamous cell carcinoma (type I), nonkeratinizing differentiated carcinoma (type II), and nonkeratinizing undifferentiated carcinoma (type III).1
The pathogenesis of NPC is related to a number of genetic and environmental factors. Subtypes of HLA antigens, chromosomal deletions affecting tumor suppressor genes, and a polymorphism of the CYP2E1 gene have all been associated with an increased risk for developing NPC.2 Infection with the Epstein-Barr virus is the most common environmental factor associated with the pathogenesis of types II and III NPC.2 Additional environmental risk factors include tobacco smoking and exposure to preserved foods, formaldehyde, and wood dust.2
Quiz Ref IDNasopharyngeal carcinoma is endemic to southern China, Southeast Asia, northern Africa, and the Arctic region; NPC is much less common in Japan and the western hemisphere. In endemic areas, the incidence of NPC has a unimodal age distribution with a peak from 50 to 60 years of age.3 In the Mediterranean countries and in select North American populations, a bimodal age distribution has a minor peak from 10 to 20 years of age and a second peak from 40 to 60 years of age.3 In the United States, adults of Chinese descent have the highest rates of NPC, followed distantly by other ethnic groups, with persons of European descent at the lowest rates of NPC.4 Nasopharyngeal cancer is 2 to 3 times more common in males than females, irrespective of race/ethnicity.5 In endemic areas, lower socioeconomic status is associated with a higher risk for NPC.6-8
In the western hemisphere, NPC is exceedingly rare in children and adolescents; the annual incidence of NPC in the United States has been estimated to be 0.5 per 1 million children.9Quiz Ref ID The WHO type III is the most common pathologic subtype of NPC in children, irrespective of geographic location or race/ethnicity, and the median age at NPC diagnosis in children is 13 years.4,10 As with adults, the incidence of NPC in children is highest in boys; however, the highest rates of NPC in persons younger than 20 years are in the black population.4,10,11
Although the demographic distribution and outcomes related to NPC in children and adolescents younger than 20 years in the United States are known, the characteristics of NPC and associated socioeconomic factors are less well described.11 The purpose of the present study was to use the National Cancer Data Base (NCDB) to examine such characteristics of NPC among pediatric patients. Our primary hypothesis was that, as with endemic areas, NPC in persons 21 years or younger in the Unites States would be associated with factors implying a lower socioeconomic status.
We used the NCDB to perform a retrospective cohort study of pediatric and adult patients with NPC in the United States. Pediatric patients from birth to 21 years of age were compared with adults older than 21 years. All of the data made available to us by the NCDB ranged from January 1, 1998, to December 31, 2011. No duplicate cases were included in the analysis. This study was approved by the institutional review board of the Seattle Children’s Hospital. Because data were deidentifed, informed consent was waived.
Patient demographics included sex, race, insurance status, household income, educational level, and rural vs urban place of residence (Table 1). Race was documented by the trained abstractors based on the medical record and based on standardized classifications within the NCDB data dictionary (http://ncdbpuf.facs.org/node/259). Race was assessed in this study because previous reports of unequal racial distributions of NPC among adult patients have been made.12-14 Whether this racial distribution is the same or different in children has yet to be determined. Insurance status was defined as private, governmental, or uninsured. Household income was divided into quartiles based on the patient’s home zip code. Educational level was based on the percentage of high school graduates in the patient’s home zip code. Rural vs urban place of residence was based on population density and proximity to a metropolitan area.
Tumor features included stage of disease, tumor behavior, tumor grade, and the presence of positive or negative margins at surgical resection. Disease staging was based on the National Comprehensive Cancer Network TNM classification.15 Tumor behavior was described as in situ or invasive, whereas tumor grade was categorized as well differentiated, moderately differentiated, poorly differentiated, or undifferentiated.15
Treatment factors included nodal evaluation and margin status. Nodal evaluation was categorized as performed or not performed. Margin status was defined as positive or negative. In addition, the types of treatment received, including radiotherapy, chemotherapy, and operative intervention, were compared between pediatric and adult patients with NPC.
Data were analyzed after data collection. To perform the analyses, pediatric patients 21 years or younger were compared with the adult population using univariate statistics with the χ2 test for categorical data (P < .05). An adjusted Cox proportional hazards regression model was used to estimate survival differences between the 2 groups. To account for the increased risk for mortality due to non-NPC causes among older patients, estimated survival was compared between the 2 groups with the adult patient group restricted to patients younger than 60 years. Finally, we stratified the pediatric group by race, and survival was again estimated using an adjusted Cox proportional hazards regression model. Adjustment factors were determined a priori and included sex, income, education, race, insurance status, urban vs rural place of residence, Charlson/Deyo comorbidity index (http://ncdbpuf.facs.org/content/charlsondeyo-comorbidity-index), tumor grade, and disease stage. Statistical analysis was completed using STATA software (version 12; StataCorp LP).
We identified a total of 17 317 patients as having a primary diagnosis of NPC, including 699 pediatric patients and 16 618 adult patients. We found no difference between pediatric and adult patients in terms of their sex distribution (239 female pediatric patients [34.2%] 5153 female adult patients [31.0%]; P = .08). In addition, we found no difference with regard to insurance status between groups (P = .97). Pediatric patients had a significantly different racial profile relative to adult patients, with a predominance of black patients (299 of 686 [43.6%]) compared with adults, among whom white race was more common (9839 of 16 405 patients [60.0%]; P < .001). Pediatric patients were significantly less likely to be of Asian race (39 of 686 [5.7%]) than were adults (3226 of 16 405 [19.7%]; P < .001) (Table 1). Quiz Ref IDPediatric patients were more likely to live in a rural area, in an area with lower median household income, and in an area of lower educational achievement relative to adult patients (P = .01, P < .001, and P < .001, respectively) (Table 1).
Pediatric patients were more likely to present with an undifferentiated tumor (364 of 528 [68.9%]) than adult patients (3509 of 12 192 [28.8%]; P < .001). Quiz Ref IDPediatric patients were also more likely to present with stage IV disease (330 of 565 [58.4%]) than adult patients (6553 of 13 721 [47.8%]; P < .001), although there was no difference in positive margin status at resection. Pediatric patients had more undifferentiated squamous cell tumors (WHO type III) (271 of 435 [62.3%]) relative to adult patients (2312 of 12 170 [19.0%]; P < .001). Pediatric patients were more likely to undergo nodal sampling at the time of resection (227 of 643 [35.3%]) relative to adult patients (3748 of 15 631 [24.0%]; P < .001) (Table 1). In addition, pediatric patients were more likely to undergo radiotherapy and chemotherapy than adult patients. Pediatric patients more frequently received multiple-agent chemotherapy than did adult patients, who more often received single-agent therapy (P < .001) (Table 2).
Compared with adults, NPC-related mortality was decreased more than 60% among pediatric patients, based on Cox proportional hazards regression analysis (hazard ratio, 0.37; 95% CI 0.25-0.56) (Table 3 and Figure 1). When the adult population was restricted to patients younger than 60 years, again NPC-related mortality was nearly 60% less likely among pediatric patients (hazard ratio, 0.41; 95% CI 0.27-0.63). When pediatric patients were stratified by race, no difference was seen in mortality by race (hazard ratio, 1.10; 95% CI 0.82-1.40) (Table 3 and Figure 2).
The present study confirms that in the United States, the demographic distribution of NPC differs between the pediatric and adult populations. Using the NCDB population of 699 persons 21 years and younger with a diagnosis of NPC, we found that relative to adults, NPC in pediatric patients is more common in black patients and relatively rare in Asians. Richey et al11 and Sultan et al9 previously used a different data source, the National Cancer Institute Surveillance, Epidemiology, and End Results (SEER) tumor registry, to examine the population demographics for persons with NPC in the United States. The study by Richey et al11 identified 160 patients younger than 20 years with a diagnosis of NPC and demonstrated that the incidence of NPC was 1.61 per 1 million for black individuals and 0.95 per 1 million for Asian and other races. The study by Sultan et al9 of 129 pediatric patients younger than 20 years and 5885 adults found an incidence of 0.5 per 1 million person-years in pediatric patients and 8.4 per 1 million person-years in adults. The present study uses different statistical methods and agrees with the results presented by Richey et al,11 but it includes a much larger population of children with a diagnosis of NPC. Both previous studies failed to demonstrate any differences in NPC-related mortality among the different ethnic populations.
Data from the NCDB also demonstrate that pediatric patients presenting with NPC tend to live in a rural area, in an area with a lower median income, or in an area with lower educational achievement. These findings are consistent with previous findings indicating that NPC is more common in adults of lower socioeconomic status.12 Given these findings, we can hypothesize that pediatric patients with NPC in such areas may be at increased risk for developing NPC owing to environmental exposures, such as tobacco smoke or preserved foods. However, tobacco exposure is most commonly associated with type I NPC, which is the least common type of NPC in children. In addition, the method of food preservation closely associated with the development of NPC is most commonly used in traditional southern Chinese culture.12 The development of NPC in the pediatric population in the United States may be related to Epstein-Barr virus infection or a specific genetic predisposition found more commonly in the black population.
Our results indicate that persons 21 years or younger with NPC in the United States more frequently present with undifferentiated tumors and more advanced disease. This finding is consistent with those of previous studies focusing on the population characteristics of NPC in the United States and in endemic areas. Sultan et al9 noted in the SEER database that children were more likely to present with advanced disease. One recent study by Yan et al13 examined the characteristics of NPC in children and adolescents at a large cancer center in southern China and found that more than 90% of patients with NPC who were younger than 21 years had stage III or stage IV disease at presentation. Despite the fact that children with NPC tend to have less differentiated, more advanced tumors, these patients have a lower risk for all-cause mortality relative to adults with NPC. The improved survival may have been related to the presence of less morbidity in the pediatric population relative to the adults; however, pediatric patients were significantly more likely to undergo radiotherapy and chemotherapy. The more aggressive treatment regimens may have contributed to an overall improved survival.
An important strength of this study is the large number of pediatric patients. The NCDB is a hospital-based registry that captures approximately 70% of all new cancer cases, including 42% of childhood cancers in the United States.16 Data capture began in 1989 and now includes more than 30 million records.16 However, some limitations to this study must be noted. First, owing to limitations in database coding, we are unable to distinguish between disease-specific and all-cause mortality, which is important when comparing adult and pediatric patient populations. We attempted to mitigate this limitation by restricting that analysis to younger adult patients (<60 years) and by controlling for adult comorbidities. Second, disease incidence could not be calculated because the NCDB does not capture a representative sample of the population. Finally, facility identification information was not available, so we could not determine whether patients were treated at adult or pediatric institutions. These data may have been useful in evaluating for differences in treatment practices. Long-term follow-up data regarding second malignant neoplasms after radiotherapy may be limited in this data set.
Although uncommon, pediatric NPC appears to affect a different patient demographic relative to adult NPC. Nasopharyngeal carcinoma in children is associated with rural location, low socioeconomic status, and more advanced disease at presentation. Despite these differences, pediatric NPC is associated with a lower mortality rate than adult disease.
Corresponding Author: Morgan K. Richards, MD, MPH, Division of Pediatric General and Thoracic Surgery, Department of General and Thoracic Surgery, Seattle Children’s Hospital, 4800 Sand Point Way NE, PO Box OA.9.220, Seattle, WA 98105.
Submitted for Publication: August 19, 2015; final revision received October 31, 2015; accepted November 11, 2015.
Published Online: January 14, 2016. doi:10.1001/jamaoto.2015.3217.
Author Contributions: Drs Richards and Gow had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.
Study concept and design: Richards, Gow, Goldin, Nuchtern, Langer, Vasudevan, Parikh.
Acquisition, analysis, or interpretation of data: Richards, Dahl, Gow, Doski, Goldfarb, Beierle, Hawkins, Parikh.
Drafting of the manuscript: Richards, Dahl, Gow, Goldin, Hawkins, Parikh.
Critical revision of the manuscript for important intellectual content: All authors.
Statistical analysis: Richards, Goldin.
Administrative, technical, or material support: Gow, Goldin, Nuchtern.
Study supervision: Gow, Goldin, Goldfarb, Vasudevan, Hawkins, Parikh.
Conflict of Interest Disclosures: None reported.
Create a personal account or sign in to: