Original Research

Pediatric Head and Neck Malignancies: Incidence and Trends, 1973-2010 Isaac Schwartz, MD1, Charlotte Hughes, MD, MPH2, and Matthew T. Brigger, MD, MPH1

No sponsorships or competing interests have been disclosed for this article.

Abstract Objective. Determine the current epidemiology of head and neck (H&N) cancer in the US pediatric population. Study Design. Analysis of national tumor registries comparing childhood H&N cancer trends to overall childhood cancer trends. Subjects and Methods. The SEER (Surveillance, Epidemiology, and End Results) database was accessed to gather epidemiologic data regarding pediatric H&N cancer between 1973 and 2010. Specific trends related to demographic background, histologic diagnosis, tumor location, and incidence, as well as general trends of all pediatric cancers, were extracted. Results. The total burden and incidence rates of pediatric cancer as well as H&N cancer continue to rise. Cancer was diagnosed in children \15 years old at a rate of 12.5 cases (95% confidence interval [CI], 11.9-13.1) per 100,000 from 1973 to 1975 and 17.3 (95% CI, 16.7-17.9) from 2007 to 2009, an increase of 38%. H&N cancers in the same age group increased from 1.1 (95% CI, 0.9-1.2) in 1973-1975 to 1.6 (95% CI, 1.4-1.8) in 2007-2009, an increase of 45%. Despite this apparent increased incidence, the proportion of H&N cancers to all cancers in the pediatric population has remained stable. Conclusion. As with all pediatric cancer, the public health burden of pediatric H&N cancer continues to rise. The proportion of H&N malignancy to all pediatric cancer is stable. With the incidence increasing, however, work to identify treatment strategies remains crucial.

Otolaryngology– Head and Neck Surgery 1–6 Ó American Academy of Otolaryngology—Head and Neck Surgery Foundation 2015 Reprints and permission: sagepub.com/journalsPermissions.nav DOI: 10.1177/0194599815575714 http://otojournal.org

in children compared to the adult population, it remains a significant cause of pediatric mortality. Approximately 1 in 285 children will be diagnosed with a malignancy prior to the age of 20.3 Despite improvements in therapeutic regimens, cancer remains the second-most common cause of death in children aged 5 to 14 years, second only to accidental trauma.1,2 Previous studies reported that the overall incidence of pediatric malignancies has increased in the United States.3,4-6 Of the pediatric cancer cases, 12% are reported to be located in the head and neck (H&N) region.7 Several studies documented increasing incidence of pediatric H&N malignancies.7-9 Through examination of large data sets, current understanding of pediatric cancer epidemiology is becoming more refined. Sultan et al found that carcinoma of the salivary glands made up a higher proportion of pediatric malignancies than adult malignancies.10 Wong et al reported increased trends of pediatric melanoma.11 The incidence of pediatric thyroid carcinoma has also been reported to be increasing.12,13 In a more broad, population-based evaluation of the SEER (Surveillance, Epidemiology, and End Results) database, Albright et al reported that the incidence of pediatric H&N malignancies increased at a greater rate than the incidence of overall pediatric malignancies from 1973 to 1996.7 There has not been a similarly broad examination of childhood H&N cancer since that time. Since the examination of data up to 1996 by Albright et al,7 there have been significant advances in the awareness and diagnosis of pediatric H&N cancer. The SEER database has undergone natural maturation and deliberate 1

Keywords cancer, epidemiology, head and neck neoplasms, United States, epidemiology, SEER program Received October 31, 2014; revised February 2, 2015; accepted February 11, 2015.

C

ancer remains the second-most common cause of death in the United States.1,2 It was estimated that 1,665,540 new malignancies would be diagnosed in 2014 and 585,720 people would die due to cancer in the United States alone.3 Although cancer is much less common

Department of Otolaryngology, Naval Medical Center San Diego, San Diego, California, USA 2 Department of Pediatrics, Naval Medical Center San Diego, San Diego, California, USA This article was presented at the 2014 AAO-HNSF Annual Meeting & OTO EXPO; September 21-24, 2014; Orlando, Florida. The views expressed in this article are those of the authors and do not reflect the official policy or position of the Department of the Navy, Department of Defense, or the United States Government. Corresponding Author: Isaac Schwartz MD, LT MC USN, Department of Otolaryngology, Naval Medical Center San Diego, 34800 Bob Wilson Drive Ste 200–ENT, San Diego, CA 92134, USA. Email: [email protected]

2

Our hypothesis was that trends in childhood H&N cancer have persisted since last reported in 2002. To test this hypothesis, we turned to the SEER registries. The SEER program maintains a repository of cancer statistics and is sponsored by the National Cancer Institute. The SEER research database contains data representing approximately 28% of the US population over a diverse geographic distribution, and it intentionally overrepresents minority populations to ensure sufficient numbers for statistical purposes.14 After obtaining an exemption from the Naval Medical Center San Diego Institutional Review Board due to the absence of any protected health information in this study, the SEER registries were queried for pediatric malignancies. H&N tumors were selected by primary tumor location using the appropriate codes from the International Classification of Disease for Oncology, 3rd edition. Malignancies of the eye, optic nerve, and central nervous system were excluded with the exception of retinoblastoma, which was included. Retinoblastomas are not generally evaluated or treated by the H&N surgeon, but they are nonetheless included in some of our analyses to directly compare our findings with previously published work. Queries included age of diagnosis, year of diagnosis, primary site of tumor, and International Classification of Disease for Oncology histologic code. Additionally, data on the annual rates of overall pediatric cancer from 1973 to 2010 were examined to provide a basis of comparison.

Number of Tumors

Methods

800 600 400 200 0

0

5

10

15

20

15

20

Age 400

Number of Tumors

expansion as well. Given these advances and lack of an updated population-level analysis of pediatric H&N cancer, this project was developed to investigate whether the reported disproportionate increase in H&N cancer has persisted. As such, this study aims to provide an update on the epidemiology of pediatric H&N malignancies in the United States, offering a direct comparison of current trends with those reported by Albright et al.

Otolaryngology–Head and Neck Surgery

300 200 100 0

0

5

10

Age

Male

Female

Figure 1. A, Age at diagnosis of head and neck (H&N) cancer, 1973-2010. The bimodal distribution is largely explained by retinoblastoma, and the disparity in males and females is largely attributable to papillary thyroid cancer (2B). B, Age at diagnosis of H&N cancer, excluding retinoblastoma and papillary thyroid carcinoma. This figure demonstrates how exclusion of these 2 malignancies changes the age distribution of the remaining H&N cancers.

Statistical Analysis The SEER database was queried using the SEER*Stat software, version 8.1.5, which provides access to the database and US population data. Descriptive statistics were derived using Microsoft Excel and GraphPad Prism 6 (GraphPad Software, San Diego, California). Prism 6 software was also used for hypothesis testing via performance of univariate linear regressions to analyze incidence trends over time. Findings were regarded as statistically significant with P  .05. As in the analysis of Albright et al,7 children .14 years old were excluded to reduce the effect of the well-described increase in papillary thyroid cancer in late-adolescent females. Annual counts, incidence rates, and proportions were based on US census data incorporated into the SEER*Stat program.15,16 Additionally, incidence rates were calculated in 3-year periods (1973-1975, 1994-1996, and 2007-2009) to provide accurate comparisons not affected by year-to-year variation. Subgroup evaluations excluding

retinoblastoma were performed to account for the high rate of this tumor in young children. In addition to the respective trends of H&N and overall cancers in children, the proportion of H&N to overall cancers was directly analyzed and plotted over the years 1973 to 2010.

Results A total of 85,561 tumors were identified in the pediatric population from 1973 to 2010; 10,214 of these were found to be in the H&N. H&N cancers accounted for 11.9% of all pediatric cancer. Figure 1A illustrates the age of diagnosis for pediatric H&N cancers, demonstrating a bimodal distribution. Examination of age at diagnosis that excluded retinoblastoma and papillary thyroid carcinoma demonstrates that the peak in infancy and the sex differences seen in adolescence can be explained by these 2 diagnoses (Figure 1B). The range of histologic diagnoses from 1973 to 2010 is shown in Table 1. The 3 most common diagnoses were lymphoma (23.8%),

Schwartz et al

3

Table 1. Pediatric Head and Neck Tumor Histologic Diagnoses. Total

Sex, No.

Cancer

No.

%

Boys

Girls

Acinar cell carcinoma Adenocarcinoma Adenoid cystic carcinoma Germ cell neoplasms Lymphoma Hodgkin Non-Hodgkin Histiocytosis Melanoma Mucoepidermoid carcinoma Neural tumors Neuroblastoma Retinoblastoma Schwannoma Other neural malignancies Skeletal sarcomas Osteosarcoma Cartilage malignancies Ewing sarcoma Other skeletal sarcomas Soft tissue sarcomas Rhabdomyosarcoma Nonrhabdomyosarcoma Squamous cell carcinoma Thyroid cancer Follicular carcinoma Papillary carcinoma Medullary carcinoma Other tumors

133 65 38 70 2432 1407 1025 65 526 271 1963 115 1533 173 142

1.3 0.6 0.4 0.7 23.8 13.8 10.0 0.6 5.1 2.7 19.2 1.1 15.0 1.7 1.4

45 24 12 35 1527 804 723 36 280 119 986 49 794 75 68

88 41 26 35 905 603 302 29 246 152 977 66 739 98 74

98 45 76 25 1029 745 284 138 2767 251 2384 132 473

1.0 0.4 0.7 0.2 10.1 7.3 2.8 1.4 27.1 2.5 23.3 1.3 4.6

58 28 41 12 533 387 146 90 542 43 444 55 249

40 17 35 13 496 358 138 48 2225 208 1940 77 224

100,000 in 2007-2009 (Table 2), representing a total increase of 38%. H&N cancers in the same age group increased from 1.1 (95% CI, 0.9-1.2) per 100,000 in 1973-1975 to 1.6 (95% CI, 1.4-1.8) per 100,000 in 2007-2009, a total increase of 45%. The annual incidence increased for overall pediatric malignancies (P \ .001) and in H&N malignancies (P = .003; Figure 2A). This analysis was performed in children aged 0 to 19 years as well, which demonstrated a similar increase in incidence over time (P \ .0001 in total and H&N cancers). This supports the notion that focusing on children aged 0 to 14 years does not dramatically alter our findings (Figure 2B). Rates of H&N cancer by sex do not appear to diverge (Figure 2C). Figure 3, importantly, demonstrates that the proportion of pediatric H&N cancers to total cancers did not significantly change between 1973 and 2010 (P = .08). To ensure the applicability of this analysis to the H&N surgeon, it was repeated excluding retinoblastoma. This exclusion did not significantly alter the results (P = .16; Figure 3C).

Discussion

papillary thyroid cancer (23.3%), and neural tumors (19.2%), which included retinoblastomas. The overall incidence of pediatric cancer reported in the SEER registries continued to rise in children aged 0 to 14 years, with 12.5 per 100,000 in 1973-1975 (95% confidence interval [CI], 11.7-13.9) and 17.3 (95% CI, 16.7-17.9) per

Cancer is a leading cause of death in the pediatric population. Patients with H&N tumors are likely to present or be referred to otolaryngologists. As such, understanding the epidemiology of H&N cancer is of clear importance to the H&N surgeon as well as to our patients. An important finding of this study comes from our direct analysis of the proportion of H&N cancers to all cancers in the pediatric population. We found that the proportion of H&N cancers to overall cancers in children to be stable over the decades. In other words, the rise of pediatric H&N cancer diagnoses is consistent with and proportional to that of overall childhood cancer. This is in contrast with previous reports that childhood H&N cancer incidence was increasing disproportionately. Previous reports suggest that the rate of cancer diagnoses—specifically, the rate of H&N cancer diagnoses— is increasing in the US pediatric population. This study did corroborate this finding, noting a significant increase in the rate of new cancer diagnoses from 1973 to 2010 in all children’s cancers as well as in the H&N. As with all population registry–level data, there is a variety of possible known and unknown confounders making this trend difficult to analyze. The development of cancer centers nationwide, the increasing

Table 2. Rates of Cancer Diagnoses in Children \15 Years of Age. 1973-1975

1994-1996

2007-2009

Cancer

Total

Rate (95% CI)

Total

Rate (95% CI)

Total

Rate (95% CI)

Overall Boys Girls Head and neck Boys Girls

1785 1001 784 152 84 68

12.5 (11.9-13.1) 13.8 (12.9-14.6) 11.1 (10.4-11.9) 1.1 (0.9-1.2) 1.2 (0.9-1.4) 1 (0.8-1.2)

2410 1340 1070 214 112 102

14.4 (13.8-15) 15.6 (14.8-16.4) 13.1 (12.3-13.9) 1.3 (1.1-1.5) 1 (1.1-1.6) 1.3 (1-1.5)

3006 1604 1402 279 139 140

17.3 (16.7-17.9) 18 (17.2-19) 16.5 (15.7-17.4) 1.6 (1.4-1.8) 1.6 (1.3-1.9) 1.7 (1.4-2)

Abbreviation: CI, confidence interval.

Otolaryngology–Head and Neck Surgery 20

Number of Tumors

4000

15 10 5

1000

20

10

00 20

90

10 20

00 20

90

15 Proportion of H&N (%)

All Pediatric Cancers

2.5 2.0 1.5 1.0

p = 0.16

10

5

10 20

00 20

90 19

80 19

19

70

0

0.5

Year 10 20

00 20

90 19

19

80

0.0 70

19

20 10

20 00

19 90

19 80

Year

Year

19

80

6

19

0 19 70

8

70

5

10

19

10

p = 0.08

12

80

15

H&N

14

19

Proportion of H&N (%)

H&N

20

H&N

19

70 19

10 20

00 20

90 19

80 19

19

Year Total Cancers

25

Rate (per 100,000)

2000

0

Year All Cancers

Rate (per 100,000)

3000

0

70

Tumor Rate (per 100,000)

4

Year Male

Female

Figure 2. A, Annual cancer rates, age 0-14 years. Rates of pediatric overall cancer and head and neck (H&N) cancer have increased. Linear regressions demonstrate statistical significance of both rate increases. P values: all cancers, P \.001; H&N cancers, P = .003. B, Cancer rates by year, age 0-19 years, 19732010. The trends of children aged 0-14 and 0-19 years demonstrate similar rates of change over time. P values: all cancers, P \ .0001; H&N cancers, P \.0001. C, Rate of H&N cancer, age 0-14 years, by sex. Rates of pediatric H&N cancer are similar between boys and girls.

Figure 3. A, Total cancers, age 0-14 years. Annual counts of all pediatric and head and neck (H&N) pediatric cancer. Spikes in 1992 and 2000 correlate with additions of geographic populations in the SEER registries. B, Annual proportion of H&N malignancies, age 0-14 years. Percentage of all pediatric tumors found in the H&N. The linear regression of this proportion is shown in green. Dotted lines represent the 95% confident interval of the linear regression (P = .08). C, Annual proportion of H&N malignancies, excluding retinoblastoma. Linear regression of this proportion shown in green. Dotted lines represent the 95% confident interval of the linear regression (P = .16).

Schwartz et al availability and precision of advanced imaging technology, and the maturation of the SEER database could all contribute to the finding that childhood cancer diagnoses appear to be on the rise. As such, it is not clear how much of this overall increasing trend can be attributed to a real increase in cancer, to earlier diagnosis of malignancy, or to improved reporting of new diagnoses to more inclusive and expansive tumor registries. However, from a population level, these updated data provide important insights to the clinician. In particular, an increased awareness of the incidence and epidemiologic trends, combined with a basic understanding of the distribution of histologic diagnoses, provides an important basis for diagnosis and counseling. This study aims to give an update on and reexamine the issue of rates of H&N cancer in children and to address concerns that the rate of H&N cancer is increasing at a relatively faster rate than that of overall cancer in the pediatric population. The SEER data show that 11.9% of overall pediatric cancers are in the H&N region, which is consistent with the analysis of Albright et al7 in their analysis of data from 1973 to 1996 (12%). The distribution of H&N cancer diagnoses from 1973 to 2010 was similar to that from 1973 to 1996. Certain tumors display sex bias—notably, thyroid cancers, in which far more tumors were diagnosed in girls, and lymphoma, for which males demonstrate a relative predilection. This finding is consistent with previous reports.3,7,8,12,13 Albright et al,7 when looking at SEER data from 1973 to 1996, determined that the increase of pediatric H&N malignancies appeared to outpace that of overall pediatric cancer. This alarming trend prompted the query of whether such an increase was sustained. The interpretation was based on a relative comparison of incidence between the time periods of 1973-1976 and 1993-1996, derived by comparing corresponding increases in incidence of pediatric H&N cancer and overall childhood cancer. Of note, there was no direct proportional comparison between the diagnosis subgroups, nor was a statistical analysis performed. Using a modified, more robust approach directly examining the proportional incidence rates over the whole period provides further detail and allows a graphic representation and statistical analysis of the data, as shown in Figure 3. Our analysis indicates that the proportion of H&N cancer to overall cancer in children does not appear to be increasing when the period of 1973-2010 is examined. In fact, the proportion has remained remarkably stable over the decades studied. While our analysis does note that the rates of pediatric H&N cancer do continue to increase, the pattern appears to be consistent with that of all pediatric cancer. As such, using a refined analysis method dispels the notion that the proportional incidence of pediatric H&N cancer is outpacing overall pediatric cancer. The stability of the proportion of H&N cancer in relation to overall pediatric cancer is important from an epidemiologic standpoint. While the finding is somewhat intuitive and anecdotally congruent with what is seen in clinical practice, a persistent shift in proportion would be concerning and require long-term surveillance. Although the proportion

5 remains stable, the overall public health impact of pediatric cancer cannot go unrecognized. Given the potential longterm disability and risk for development of future malignancies associated with successful treatment in children, a keen awareness of the short- and long-term individual and societal effects is requisite. The strengths of this study are its inclusivity of data and its timeline. By looking at data from a national tumor registry, we were able to analyze large numbers of tumors, over a period of almost 4 decades. This allowed us to examine the epidemiology of pediatric H&N cancer from the perspective of a national population. In addition to the above-described potential confounding issues inherent in population-level data analyses, the primary limitations of this approach stem from the required procedures to develop and maintain such a registry. The necessary reliance of proper and standardized coding from the diverse SEER contributors can be subject to error, although the SEER data are carefully standardized and validated. Additionally, such procedures result in a decrease of individual detail that comes from taking a broad, populationbased approach.

Conclusion Pediatric H&N cancer remains a public health concern in the United States. Similar to that of pediatric cancer in general, the overall incidence appears to be increasing. Analysis of the SEER database reveals that previously raised concerns about the incidence of H&N cancer increasing at a greater rate than that of overall pediatric cancer have not persisted. Given the overall increase in diagnosis, further work in identifying treatment strategies for children suffering from H&N cancer is crucial. Author Contributions Isaac Schwartz, obtained SEER data, designed study, analyzed data, generated figures, wrote Methods, Results, and Discussion sections; Charlotte Hughes, initiated project, designed study, literature review, wrote introduction, revised manuscript; Matthew T. Brigger, initiated project, designed study, revised article.

Disclosures Competing interests: None. Sponsorships: None. Funding source: None.

References 1. Heron M. Deaths: leading causes for 2008. Natl Vital Stat Rep. 2012;60:1-94. 2. Murphy SL, Xu J, Kochanek KD. Deaths: final data for 2010. Natl Vital Stat Rep. 2013;61:1. 3. American Cancer Society. Cancer Facts & Figures 2014. Atlanta, GA: American Cancer Society; 2014. 4. Linet MS, Ries LA, Smith MA, Tarone RE, Devesa SS. Cancer surveillance series: recent trends in childhood cancer incidence and mortality in the United States. J Natl Cancer Inst. 1999;91:1051-1058.

6 5. Gurney JG, Davis S, Severson RK, Fang JY, Ross JA, Robison LL. Trends in cancer incidence among children in the US. Cancer. 1996;78:532-541. 6. Linabery A, Ross J. Trends in childhood cancer incidence in the US (1992-2004). Cancer. 2008;112:416-432. 7. Albright J, Topham A, Reilly J. Pediatric head and neck malignancies: US incidence and trends over 2 decades. Arch Otolaryngol Head Neck. 2002;128:655-659. 8. Cunningham MJ, Myers EN, Bluestone CD. Malignant tumors of the head and neck in children: a twenty-year review. Int J Pediatr Otorhinolaryngol. 1987;13:279-292. 9. Schantz S, Yu G. Head and neck cancer incidence trends in young Americans, 1973-1997, with a special analysis for tongue cancer. Arch Otolaryngol Head Neck. 2002;128:268-274. 10. Sultan I, Rodriguez GC, Al Sharabati S, Guzzo M, Casanova M, Ferrari A. Salivary gland carcinomas in children and adolescents: a population-based study, with comparison to adult cases. Head Neck. 2011;33:1476-1481. 11. Wong J, Harris J, Rodriguez GC, Johnson K. Incidence of childhood and adolescent melanoma in the United States: 1973-2009. Pediatrics. 2013;131:846-854. 12. Holmes L, Hossain J, Opara F. Pediatric thyroid carcinoma incidence and temporal trends in the USA (1973-2007): race or shifting diagnostic paradigm? ISRN Oncol. 2012;2012: 906197.

Otolaryngology–Head and Neck Surgery 13. Hogan A, Zhuge Y, Perez E, Koniaris L, Lew J, Sola J. Pediatric thyroid carcinoma: incidence and outcomes in 1753 patients. J Surg Res. 2009;156:167-172. 14. Park H, Lloyd S, Decker R, Wilson L, Yu J. Overview of the Surveillance, Epidemiology, and End Results database: evolution, data variables, and quality assurance. Curr Probl Cancer. 2012;36:183-190. 15. Surveillance, Epidemiology, and End Results Program. SEER*Stat Database: incidence—SEER 9 regs research data, Nov 2012 sub (1973-2010) \Katrina/Rita population adjustment.—linked to county attributes—total US, 1969-2011 counties. National Cancer Institute, Division of Cancer Control and Population Sciences, Surveillance Research Program, Surveillance Systems Branch. http://www.seer.cancer.gov. Released April 2013. Based on the November 2012 submission. 16. Surveillance, Epidemiology, and End Results Program. SEER*Stat Database: incidence—SEER 18 regs research data 1 Hurricane Katrina impacted Louisiana cases, Nov 2012 sub (1973-2010 varying)—linked to county attributes—total US, 1969-2011 counties. National Cancer Institute, Division of Cancer Control and Population Sciences, Surveillance Research Program, Surveillance Systems Branch. http://www.seer.cancer.gov. Released April 2013. Based on the November 2012 submission.