- Email: [email protected]
Pediatric sinonasal malignancies: A population-based analysis
International Journal of Pediatric Otorhinolaryngology 98 (2017) 97e102
Contents lists available at ScienceDirect
International Journal of Pediatric Otorhinolaryngology journal homepage: http://www.ijporlonline.com/
Pediatric sinonasal malignancies: A population-based analysis Sei Y. Chung, B.S. a, Aykut A. Unsal, D.O. b, Suat Kılıç, B.A. a, Soly Baredes, M.D., F.A.C.S. a, c, James K. Liu, M.D., F.A.C.S. a, c, d, Jean Anderson Eloy, M.D., F.A.C.S. Professor and Vice Chairman, Director, Rhinology and Sinus Surgery, Director, Otolaryngology Research, Co-Director, Endoscopic Skull Base Surgery Program, Professor of Neurological Surgery, Professor of Ophthalmology and Visual Science a, c, d, e, * a
Department of Otolaryngology e Head and Neck Surgery, Rutgers New Jersey Medical School, Newark, NJ, USA Department of Otolaryngology & Facial Plastic Surgery, Rowan University School of Osteopathic Medicine, Stratford, NJ, USA c Center for Skull Base and Pituitary Surgery, Neurological Institute of New Jersey, Rutgers New Jersey Medical School, Newark, NJ, USA d Department of Neurological Surgery, Rutgers New Jersey Medical School, Newark, NJ, USA e Department of Ophthalmology and Visual Science, Rutgers New Jersey Medical School, Newark, NJ, USA b
a r t i c l e i n f o
a b s t r a c t
Article history: Received 25 January 2017 Received in revised form 21 April 2017 Accepted 21 April 2017 Available online 25 April 2017
Introduction: Pediatric Sinonasal Malignancies (PedsSNM) are rare and usually associated with a poor prognosis. We aim to investigate the epidemiology, tumor characteristics, and survival of PedsSNM using a population-based database to augment the scant literature on this topic. Methods: The Surveillance, Epidemiology, and End Results database was queried for patients 18 years diagnosed with PedsSNM between the years of 1973 and 2013. Data on incidence, tumor characteristics, and survival were analyzed. Results: In total, 210 patients with PedsSNM were identified. Demographically, 54.3% were female, 72.6% were white, and the mean age was 10.7 years. Overall incidence was 0.036 per 100,000 individuals between the years of 2000 and 2013. The nasal cavity was the most frequent primary site (37.1%) and rhabdomyosarcoma was the most frequent malignancy (50.5%). Five-, 10-, and 20-year disease-specific survival (DSS) rates were 60.2%, 46.1%, and 20.6%, respectively. Grade IV tumors made up the largest group (37.3%), and such tumors exhibited the worst 5-, 10-, and 20-year survival (P < 0.05). Distant disease predicted the worst 5-, 10-, and 20-year survival, followed by regional, then localized disease (P < 0.01). Patients treated with surgery alone had a higher 20-year survival (P ¼ 0.0425). No significant differences in survival were observed between race, gender, primary site, or histology. Conclusions: PedsSNM frequently presented as Grade IV tumors. The nasal cavity was the most common primary site and rhabdomyosarcoma was the most frequent histology. Patients receiving surgery alone had the highest survival; however, this may be a reflection of smaller, less aggressive tumors preferentially being treated with surgery alone. © 2017 Elsevier B.V. All rights reserved.
Keywords: Pediatric carcinoma Sinonasal carcinoma Pediatric malignancies Sinonasal malignancies Rhabdomyosarcoma Survival
1. Introduction While less than 5% of head and neck malignancies arise from the nasal cavity and paransasal sinuses, pediatric sinonasal
* Corresponding author. Department of Otolaryngology e Head and Neck Surgery, Neurological Institute of New Jersey, Rutgers New Jersey Medical School, 90 Bergen St., Suite 8100, Newark, 07103, NJ, USA. E-mail address: [email protected] (J.A. Eloy). http://dx.doi.org/10.1016/j.ijporl.2017.04.032 0165-5876/© 2017 Elsevier B.V. All rights reserved.
malignancies (PedsSNM) are even rarer [1,2]. The presenting symptoms of nasal obstruction, facial swelling, ophthalmic complaints, rhinorrhea, epistaxis, rhinosinusitis, and chronic upper respiratory tract infections are nonspecific and can be easily mistaken for more common conditions [1e4]. Diagnosis is often delayed by weeks to years and results in advanced disease stage at presentation [2e5]. PedsSNM have been shown to have distinct epidemiological, clinicopathologic, histological, and survival differences compared to adult sinonasal malignancies, emphasizing the value of examining these malignancies in pediatric patients
98
S.Y. Chung et al. / International Journal of Pediatric Otorhinolaryngology 98 (2017) 97e102
alone [1,2]. Given the rarity of PedsSNM, we used the US National Cancer Institute's Surveillance, Epidemiology, and End Results (SEER) database to identify a large, nationally representative cohort, with long-term surveillance. SEER is a multicenter, population-based record that has collected information on malignancies since 1973, and is exempt from institutional biases. Until recently, case reports, systematic reviews, and single-institution studies largely accounted for the scant literature on PedsSNM [4]. Shapiro et al. conducted the first database analysis of PedsSNM and provided some valuable information about their demographic and clinicopathologic characteristics. Their series was limited to 63 cases, and they could not provide any survival analyses due to their small sample size [3]. Gerth et al. have thus far been the only ones to provide survival analyses using SEER. Our study examines 210 PedsSNM patients, provides updated analyses of any previously reported survival trends, and compares our findings to that of the literature. Furthermore, we investigate the role of other potential prognostic indicators of survival not previously explored, such as histological grade at diagnosis. 2. Materials and methods The SEER 18 database was queried in order to obtain frequency, survival, and incidence data for all pediatric sinonasal tumors of the head and neck between the years 1973 through 2013. Sinonasal lymphoma cases were excluded from our analysis to facilitate staging and survival analyses. The SEER database is maintained by the National Cancer Institute (Bethesda, Maryland) and provided all the analyzed information including patient demographics, tumor characteristics, and incidence and survival rates. The SEER 18 database is an amalgamation of data collected from 18 registries from Connecticut, New Jersey, Georgia, Louisiana, Kentucky, New Mexico, Utah, California, Hawaii, Alaska, Oklahoma, Arizona, Michigan, Washington, and Iowa. Some of the registries are statewide registries, as in the case of New Jersey and Connecticut state registries. Others cover a particular metropolitan area, as in the case of the Los Angeles and San Franscico-Oakland registries. Overall, the database captures 28% of the population in the US. The data are void of any private identifiable health records, thus this study has been exempt from Institutional Review Board (IRB) approval as per the standing policy of the IRB of Rutgers New Jersey Medical School, Newark, New Jersey. 2.1. Patient and tumor characteristics International Classification of Disease for Oncology, 3rd Edition (ICD-O-3) topography codes corresponding to all sinonasal tumors were initially used to query case data from the SEER 18 database (C30.0, C31.0, C31.1, C31.2, C31.3, C31.8, C31.9). This dataset was further confined to only include patient cases who were under 18 years of age. The resultant data were then stratified by age groups, gender, race, histology, and primary site. Tumor properties were further analyzed by grade, extension of disease, and treatment modalities. All unknown values were included in the reported results; however, they have been omitted from distribution calculations. 2.2. Statistical analysis Pediatric sinonasal tumors diagnosed between 2000 and 2013 were reported for incidence trends as reported per 100,000 when adjusted to the standard 2000 US population (Census P25-1130). Calculated annual percent change (APC) was observed with 1year end points. Weight least squares were utilized for APC significance testing.
SEER*Stat 8.3.2 (National Cancer Institute, Bethesda, MD) was used to extract all patient data, which were subsequently organized on Microsoft Excel 2016 spreadsheets (Microsoft Corporation, Redmond, WA). Survival analysis for 5-, 10-, and 20-years was then accomplished using the binary scheme utilized in Kaplan-Meier analysis. All untraced patients were excluded from analyses, while traced patients were assigned ones and zeros based on their disease-specific survival (DSS) by the end of each time period, respectively. Demographic frequency distributions did however include all patient cases. Chi-square tests were used to compare categorical variables. JMP Statistical Discovery 12.2 (SAS Institute, Cary, NC) was then utilized to calculate log-rank analysis and Kaplan-Meier curves for all resultant data. Significance levels were set to f ¼ 0.05 for all tests. 3. Results 3.1. Patient and tumor characteristics A total of 210 PedsSNM patients, 18 years old or younger, were identified from the SEER database (Table 1). The mean age at diagnosis was 10.7 years with a standard deviation of 5.64 years. By gender, 45.7% were male and 54.3% were female. By race, 72.6% were white, 21.2% were black, 5.3% were Asian/Pacific Islander, and 1.0% were American Indian/Alaskan Native. This distribution represents prevalence only. By primary site, the nasal cavity was the most commonly involved subsite (37.1%), followed by the maxillary sinus (27.6%), and ethmoid sinus (17.1%). By histology, rhabdomyosarcoma was the most common type (50.5%), followed by olfactory neuroblastoma (17.6%), sarcoma not otherwise specified (NOS) (11.4%), and squamous cell carcinoma (2.9%). The remaining distribution of primary site and histologies is available in Table 2. Anatomic site distribution varied significantly by histology (Table 3) (p < 0.0001). Most olfactory neuroblastomas were in the nasal cavity (73.0%). Sarcoma NOS was more commonly seen in the maxillary sinus (41.7%), and rhabdomyosarcomas were distributed almost evenly throughout the different sinonasal sites. The most common histology varied by age (p ¼ 0.025). In the 0e3 year age group, rhabdomyosarcoma (76.0%) was by far the most common histology. However, in the 12e15 year age group, olfactory neuroblastoma (40.9%) was almost as common as rhabdomyosarcoma (45.5%). The distribution of tumor grade among cases demonstrated a propensity for grade IV disease (37.3%), followed by grade II (25.3%), grade III (22.7%), and grade I (14.7%) (Table 4). Table 1 Demographics of pediatric sinonasal malignancies.
Mean age of diagnosis ( ± SD) Total Age groups (years) 0e3 4e7 8e11 12e15 16e18 Gender Female Male Race White Black Asian or Pacific Islander American Indian/Alaska Native Unknown
N
%
10.7 ± 5.64 210
e 100.0%
35 27 36 58 54
16.7% 12.9% 17.1% 27.6% 25.7%
114 96
54.3% 45.7%
151 44 11 2 2
72.6% 21.2% 5.3% 1.0% e
S.Y. Chung et al. / International Journal of Pediatric Otorhinolaryngology 98 (2017) 97e102 Table 2 Frequency of pediatric sinonasal malignancies by site and histology. N Primary Site Nasal Cavity Maxillary sinus Ethmoid sinus Accessory sinus, NOS Sphenoid sinus Overlapping lesion of accessory sinuses Frontal sinus Histology Rhabdomyosarcoma Olfactory neuroblastoma Sarcoma NOS Squamous cell carcinoma Adenocarcinoma Neuroblastoma NOS Yolk sac tumor Carcinoma NOS Lymphoepithelial carcinoma Peripheral neuroectodermal tumor Spindle cell tumor Adenoid cystic carcinoma Neuroendocrine carcinoma NOS Small cell carcinoma NOS Transitional cell carcinoma NOS Mucoepidermoid carcinoma Acinar cell carcinoma Malignant rhabdoid tumor Malignant myoepithelioma Teratoma (malignant) NOS Chondroid chordoma Primitive neuroectodermal tumor Olfactory neurocytoma Neurilemoma (malignant) (OBS)
99
Table 4 Histological grade of pediatric sinonasal malignancies. %
78 58 36 21 10 7 0
37.1% 27.6% 17.1% 10.0% 4.8% 3.3% e
106 37 24 5 4 4 4 3 3 3 2 2 2 1 1 1 1 1 1 1 1 1 1 1
50.5% 17.6% 11.4% 2.4% 1.9% 1.9% 1.9% 1.4% 1.4% 1.4% 1.0% 1.0% 1.0% 0.5% 0.5% 0.5% 0.5% 0.5% 0.5% 0.5% 0.5% 0.5% 0.5% 0.5%
NOS ¼ not otherwise specified.
3.2. Incidence and survival The incidence of PedsSNM from 2000 to 2013 was 0.036 per 100,000. Overall 5-, 10-, and 20-year DSS rates were 60.2%, 46.1%, and 20.6%, respectively (Table 5 and Fig. 1). Gender, race, and primary site of PedsSNM did not confer any survival differences at 5, 10, and 20 years. At all time intervals studied, DSS decreased as tumor grade increased (P < 0.05). When evaluating PedsSNM by extension of disease, localized cases conferred the best 5-, 10-, and 20-year DSS, while regional involvement reflected significantly worse survival. Distant disease was associated with the worst DSS at all time intervals (P < 0.01). After 20 years, 0% of distant disease cases survived, whereas 28.2% of cases with regional disease and 33.3% of cases with localized disease survived (Table 5). DSS rates were calculated based on the three most common histologies: rhabdomyosarcoma, olfactory neuroblastoma, and sarcoma NOS (Table 6). There were no significant differences in survival between these subtypes for all time intervals studied. Stratifying DSS by therapeutic modality demonstrated that individuals treated with surgery alone had the highest survival rates at 5, 10 and 20 years (80.7%, 73.9%, and 50.0%, respectively), followed by neither surgery nor radiotherapy (71.4%, 50.0%, and 33.3%, respectively), combination therapy with surgery and radiotherapy
Grade
N
%
Well differentiated; Grade I Moderately differentiated; Grade II Poorly differentiated; Grade III Undifferentiated; anaplastic; Grade IV Unknown
11 19 17 28 135
14.7% 25.3% 22.7% 37.3% e
(65.5%, 47.6%, and 25.3%, respectively), and lastly radiotherapy only (51.1%, 38.9%, and 8.0%, respectively). However, these differences were only statistically significant at 20 years (P ¼ 0.0425) (Table 7 and Fig. 2). The histologic breakdown of patients receiving each therapeutic modality differed significantly (p < 0.0001). Among patients receiving surgery and radiotherapy, rhabdomyosarcoma made up 40.6%, olfactory neuroblastoma made up 37.5%, and sarcoma NOS made up 4.7%. Among patients receiving radiotherapy alone, rhabdomyosarcoma made up 86.0%, olfactory neuroblastoma made up 2.0%, and sarcoma NOS made up 0.0%. Among patients receiving surgery alone, rhabdomyosarcoma made up 30.0%, olfactory neuroblastoma made up 13.3%, and sarcoma NOS made up 26.7%. Among patients receiving neither surgery nor radiotherapy, rhabdomyosarcoma made up 25.0%, olfactory neuroblastoma made up 25.0%, and sarcoma NOS made up 12.5%. 4. Discussion PedsSNM are rare, heterogeneous malignancies that behave aggressively and are associated with poor survival. The current study included 210 PedsSNM patients within a 40-year period. The mean age at diagnosis was 10.7 years with a standard deviation of 5.64 years, consistent with other mean ages reported in prior series [1e4]. A small case series initially reported a slight male predominance in PedsSNM [2]. When Turner et al. first examined sinonasal malignancies, a male to female ratio of 1.8:1 was reported [6]. However, they demonstrated a significantly decreasing incidence trend in males, and a stagnant incidence in females [6]. Gerth et al. subsequently reported a male to female incidence ratio of 1.2:1, and that the incidence rate in males was lower than in females [1]. Our study demonstrates the continuation of these incidence trends; the male to female ratio was 0.84:1. By race, most of our PedsSNM patients were white (72.6%), followed by black (21.2%), likely reflecting the racial distribution of the general US population. In our survival analysis, sex and ethnicity did not affect 5-, 10-, or 20-year DSS, as observed in prior studies [1,6]. PedsSNM most commonly arose from the nasal cavity, followed by the maxillary sinus, and ethmoid sinus. Gerth and colleagues did not find significant survival differences depending on primary site of PedsSNM [1]. Likewise, our analysis yielded the same findings. Most of the current literature demonstrates rhabdomyosarcoma to be the most common histological subtype of PedsSNM [1e5]. One contradictory study utilizing SEER found olfactory neuroblastoma to be the most common diagnosis, followed by rhabdomyosarcoma [4]. The present analysis agrees with the majority of previous studies; rhabdomyosarcoma was the most common PedsSNM subtype. Furthermore, there were no significant
Table 3 Site distribution by histology. n(%)
Nasal cavity
Maxillary sinus
Ethmoid sinus
Sphenoid sinus
Overlapping lesion of sinuses
Accessory sinus, NOS
Rhabdomyosarcoma Olfactory Neuroblastoma Sarcoma NOS
29(27.4%) 27(73.0%) 8(33.3%)
25(23.6%) 4(10.8%) 10(41.7%)
20(18.9%) 5(13.5%) 5(20.8%)
9(8.5%) 0 0
6(5.7%) 0 1(4.2%)
17(16.0%) 1(2.7%) 0
100
S.Y. Chung et al. / International Journal of Pediatric Otorhinolaryngology 98 (2017) 97e102
Table 5 Disease specific survival rates by demographics, tumor grade, and primary site. 5-Y
Overall Survival Gender Female Male Ethnicity White Black Asian or Pacific Islander American Indian/Alaska Native Unknown Grade Well differentiated; Grade I Moderately differentiated; Grade II Poorly differentiated; Grade III Undifferentiated; anaplastic; Grade IV Unknown Primary Site Nasal Cavity Maxillary sinus Ethmoid sinus Accessory sinus, NOS Sphenoid sinus Overlapping lesion of accessory sinuses Extension of Disease Localized Regional Distant a
10-Y
20-Y
Na
%
Na
%
Na
%
166 (P ¼ 0.5859) 91 75 (P ¼ 0.7104) 124 30 9 2 1 (P ¼ 0.0001) 9 13 12 22 110 (P ¼ 0.3047) 58 48 29 20 7 4 (P ¼ 0.0005) 22 82 33
60.2%
127 (P ¼ 0.1934) 73 54 (P ¼ 0.5185) 93 26 7 1 0 (P < 0.0001) 7 7 8 19 86 (P ¼ 0.6979) 38 38 22 19 6 4 (P ¼ 0.0001) 15 61 26
46.1%
89 (P ¼ 0.2014) 50 39 (P ¼ 0.6750) 64 20 4 1 0 (P ¼ 0.0480) 1 4 5 18 61 (P ¼ 0.6575) 24 28 16 16 4 1 (P ¼ 0.0021) 6 42 21
20.6%
61.9% 58.3% 63.1% 49.3% 55.6% 50.0% 100.0% 100.0% 84.6% 58.3% 27.3% 60.9% 72.3% 52.8% 61.2% 45.0% 42.9% 75.0% 86.4% 65.3% 39.4%
50.8% 39.6% 48.2% 41.2% 42.9% 0.0% e 100.0% 71.4% 37.5% 10.5% 48.3% 54.8% 39.8% 43.5% 42.1% 33.3% 75.0% 80.0% 51.3% 19.2%
25.1% 14.7% 23.6% 16.9% 0.0% 0.0% e 100.0% 50.0% 0.0% 5.6% 23.8% 27.5% 16.4% 20.6% 25.0% 0.0% 0.0% 33.3% 28.2% 0.0%
Not all cases had survival data; therefore, the number of cases may be less than in Tables 1e3.
Fig. 1. Kaplan-Meier curve of cumulative 20-year disease-specific survival.
differences in 5-, 10-, or 20-year DSS when stratified by histology, which supports Gerth's findings [1]. The greatest number of cases presented as a grade IV disease (37.3%), followed by grade II (25.3%), grade III (22.7%), and grade I
(14.7%), emphasizing the aggressive nature of PedsSNM at presentation. The present study suggests that higher tumor grade has a negative effect on survival. All grade I disease cases survived 20 years. However, for grade II disease, 5-year survival dropped by
S.Y. Chung et al. / International Journal of Pediatric Otorhinolaryngology 98 (2017) 97e102 Table 6 Disease specific survival rates of most common histologies. Histology
5-Y Na
Rhabdomyosarcoma Olfactory neuroblastoma Sarcoma NOS
10-Y Na
%
20-Y %
Na
%
(P ¼ 0.3083)
(P ¼ 0.5461)
(P ¼ 0.0967)
86 32 14
71 21 11
54 12 7
52.7% 68.8% 71.4%
41.0% 47.6% 63.6%
21.6% 8.3% 42.9%
a Not all cases had survival data; therefore, the number of cases may be less than in Tables 1e3.
Table 7 Disease specific survival by therapy. 5-Y N
Surgery and radiotherapy Radiotherapy Surgery No surgery or radiotherapy
10-Y %
N
20-Y %
N
%
(P ¼ 0.0852)
(P ¼ 0.0794)
(P ¼ 0.0425)
69 45 31 7
50 36 23 4
36 25 12 3
65.6% 51.1% 80.7% 71.4%
47.6% 38.9% 73.9% 50.0%
25.3% 8.0% 50.0% 33.3%
more than 15%. After 10 and 20 years, DSS dropped further by 29% and 50%, respectively. These trends continue for grade III and IV lesions, with dramatic decreases in DSS at each time interval. AJCC and TNM staging were excluded from our study due to the lack of staging data available prior to 1988. Instead, we utilized older staging descriptors, such as “localized,” “regional,” or “distant” to account for cases dating back to 1973. Cases with distant metastatic disease, and cases with involvement of regional structures or lymph nodes demonstrated dramatically lower 5-, 10-, and 20-year DSS, than cases with localized disease. Consensus on treating PedsSNM does not yet exist;
101
otolaryngologists are currently using histopathologic-specific protocols that are continually changing and were mainly created for the general adult population. Furthermore, extirpation of tumor in the sinonasal region near the skull base is particularly challenging in children, and morbidity from complications can be detrimental and life-altering. Radiotherapy also has the potential to cause serious acute and long-term sequelae including osteoradionecrosis, blindness, mucositis, infection, hypopituitarism, and a second primary malignancy [1,5]. Because available treatment modalities have these associated possible ramifications, physicians and institutions have varying preferences in what they recommend to PedsSNM patients. Our findings revealed that the most common treatment modality used in PedsSNM was combination surgery and radiotherapy, which has been shown to be the treatment modality of choice in institutional studies [4,7]. The next most common therapies were radiotherapy only, surgery only, and neither surgery nor radiotherapy. A previous SEER study by Gerth et al. had stated that any surgery resulted in significantly higher survival compared to no surgery performed throughout their entire study period [1]. In contrast, the present analysis found no significant differences in DSS at 5 and 10 years between treatment groups. After 20 years, however, patients who had undergone any surgery had the highest survival rates versus those who had not. We also found that patients who received radiotherapy, either alone or in combination with surgery, had significantly lower DSS than patients who received surgery alone. Our results could signify that surgery should be the mainstay of treatment in PedsSNM. Also, the longterm morbidities and complications of radiotherapy may outweigh the benefits, and may possibly decrease survival in the pediatric population. However, these interpretations must be taken with great caution; our findings could merely be reflecting that patients receiving radiotherapy alone or surgery and radiotherapy have tumors that are larger, more aggressive, or possess other poor prognostic factors. Furthermore, rhabdomyosarcoma cases were
Fig. 2. Kaplan-Meier curve of cumulative 20-year disease-specific survival by treatment.
102
S.Y. Chung et al. / International Journal of Pediatric Otorhinolaryngology 98 (2017) 97e102
more likely to receive radiotherapy alone; olfactory neuroblastoma cases were more likely to receive surgery and radiotherapy; and sarcoma NOS cases were more likely to receive surgery alone. There was an inverse correlation between the percentage of aggressive histologies in therapeutic modality groups and greater survival rates. Therefore, histology may be one of the main determinants of these survival differences. Interestingly, patients receiving neither surgery nor radiotherapy exhibited significantly higher 20-year survival than those receiving combination therapy or radiotherapy alone. We posit that patients who underwent neither surgery nor radiotherapy, but had higher DSS, may possibly represent patients with disease that was amenable to chemotherapy only. Of note, our findings starkly contradict Gerth's SEER study, which reported that patients receiving no treatment had the very worst prognosis throughout their entire study period. Chemotherapy is thought to have a role in treating PedsSNM [4]; however, SEER does not provide data regarding chemotherapy use, limiting our analysis on treatment modality. For example, multidrug chemotherapy regimens may be effectively used as an adjunctive therapy for pediatric rhabdomyosarcoma, especially in advanced cases [1,2,4,8e10]. Olfactory neuroblastomas have also been demonstrated to respond to chemotherapy [1,8,11,12]. Further limitations of this study include those inherent to the SEER registries. We must assume this database contains accurate diagnoses and coding. Information such as comorbidities, surgical details, such as margin status, and radiotherapy dose for each case are missing from the SEER registries. Despite its limitations, SEER is an important database for studying rare tumors, with strong statistical power and cross-institutional data. 5. Conclusion Pediatric sinonasal malignancies are extremely rare and deserve attention separate from adult sinonasal malignancies. Although shown to historically occur more in males, these malignancies are now slightly predominant in females. They most frequently occurred in patients in their second decade of life. PedsSNM most often presented as grade IV tumors. The most common histology was rhabdomyosarcoma, and the most common primary site was the nasal cavity. While gender, race, primary site, and histological subtype did not affect survival, higher tumor grade and regional
involvement or distant metastasis at presentation predicted significantly worse disease-specific survival. Surgery alone was associated with the most favorable 20-year disease-specific survival rate. However, further investigation is warranted to delineate whether this is a true survival benefit or is merely reflecting smaller, less aggressive tumors that are amenable to surgery alone. Financial disclosures None. Conflicts of interest None. References [1] D.J. Gerth, J. Tashiro, S.R. Thaller, Pediatric sinonasal tumors in the United States: incidence and outcomes, J. Surg. Res. 190 (2014) 214e220. [2] J.S. Yi, G.S. Cho, M.J. Shim, J.Y. Min, Y.S. Chung, B.J. Lee, Malignant tumors of the sinonasal tract in the pediatric population, Acta Otolaryngol. 132 (Suppl 1) (2012) S21eS26. [3] N.L. Shapiro, N. Bhattacharyya, Staging and survival for sinus cancer in the pediatric population, Int. J. Pediatr. Otorhinolaryngol. 73 (2009) 1568e1571. [4] M.M. Benoit, N. Bhattacharyya, W. Faquin, M. Cunningham, Cancer of the nasal cavity in the pediatric population, Pediatrics 121 (2008) e141e145. [5] J.P. Zevallos, K.S. Jain, D. Roberts, A. El-Naggar, E.Y. Hanna, M.E. Kupferman, Sinonasal malignancies in children: a 10-year, single-institutional review, Laryngoscope 121 (2011) 2001e2003. [6] J.H. Turner, D.D. Reh, Incidence and survival in patients with sinonasal cancer: a historical analysis of population-based data, Head. Neck 34 (2012) 877e885. [7] R.P. Tufano, N.A. Mokadam, K.T. Montone, et al., Malignant Tumors of the Nose and Paranasal Sinuses: Hospital of the University of Pennsylvania Experience 1990-1997, Am. J. Rhinol. vol. 13 (1999) 117e123. [8] P. LoRusso, E. Tapazoglou, J.A. Kish, et al., Chemotherapy for paranasal sinus carcinoma. A 10-year experience at Wayne State University, Cancer 62 (1988) 1e5. [9] M.B. Gillespie, D.T. Marshall, T.A. Day, A.O. Mitchell, D.R. White, J.C. Barredo, Pediatric rhabdomyosarcoma of the head and neck, Curr. Treat. Options Oncol. 7 (2006) 13e22. [10] S. Sanghvi, P. Misra, N.R. Patel, E. Kalyoussef, S. Baredes, J.A. Eloy, Incidence trends and long-term survival analysis of sinonasal rhabdomyosarcoma, Am. J. Otolaryngol. 34 (2013) 682e689. [11] M.M. Fitzek, A.F. Thornton, M. Varvares, et al., Neuroendocrine tumors of the sinonasal tract. Results of a prospective study incorporating chemotherapy, surgery, and combined proton-photon radiotherapy, Cancer 94 (2002) 2623e2634. [12] J.M. Sheehan, J.P. Sheehan, J.A. Jane Sr., R.S. Polin, Chemotherapy for esthesioneuroblastomas, Neurosurg. Clin. N. Am. 11 (2000) 693e701.
Contents lists available at ScienceDirect
International Journal of Pediatric Otorhinolaryngology journal homepage: http://www.ijporlonline.com/
Pediatric sinonasal malignancies: A population-based analysis Sei Y. Chung, B.S. a, Aykut A. Unsal, D.O. b, Suat Kılıç, B.A. a, Soly Baredes, M.D., F.A.C.S. a, c, James K. Liu, M.D., F.A.C.S. a, c, d, Jean Anderson Eloy, M.D., F.A.C.S. Professor and Vice Chairman, Director, Rhinology and Sinus Surgery, Director, Otolaryngology Research, Co-Director, Endoscopic Skull Base Surgery Program, Professor of Neurological Surgery, Professor of Ophthalmology and Visual Science a, c, d, e, * a
Department of Otolaryngology e Head and Neck Surgery, Rutgers New Jersey Medical School, Newark, NJ, USA Department of Otolaryngology & Facial Plastic Surgery, Rowan University School of Osteopathic Medicine, Stratford, NJ, USA c Center for Skull Base and Pituitary Surgery, Neurological Institute of New Jersey, Rutgers New Jersey Medical School, Newark, NJ, USA d Department of Neurological Surgery, Rutgers New Jersey Medical School, Newark, NJ, USA e Department of Ophthalmology and Visual Science, Rutgers New Jersey Medical School, Newark, NJ, USA b
a r t i c l e i n f o
a b s t r a c t
Article history: Received 25 January 2017 Received in revised form 21 April 2017 Accepted 21 April 2017 Available online 25 April 2017
Introduction: Pediatric Sinonasal Malignancies (PedsSNM) are rare and usually associated with a poor prognosis. We aim to investigate the epidemiology, tumor characteristics, and survival of PedsSNM using a population-based database to augment the scant literature on this topic. Methods: The Surveillance, Epidemiology, and End Results database was queried for patients 18 years diagnosed with PedsSNM between the years of 1973 and 2013. Data on incidence, tumor characteristics, and survival were analyzed. Results: In total, 210 patients with PedsSNM were identified. Demographically, 54.3% were female, 72.6% were white, and the mean age was 10.7 years. Overall incidence was 0.036 per 100,000 individuals between the years of 2000 and 2013. The nasal cavity was the most frequent primary site (37.1%) and rhabdomyosarcoma was the most frequent malignancy (50.5%). Five-, 10-, and 20-year disease-specific survival (DSS) rates were 60.2%, 46.1%, and 20.6%, respectively. Grade IV tumors made up the largest group (37.3%), and such tumors exhibited the worst 5-, 10-, and 20-year survival (P < 0.05). Distant disease predicted the worst 5-, 10-, and 20-year survival, followed by regional, then localized disease (P < 0.01). Patients treated with surgery alone had a higher 20-year survival (P ¼ 0.0425). No significant differences in survival were observed between race, gender, primary site, or histology. Conclusions: PedsSNM frequently presented as Grade IV tumors. The nasal cavity was the most common primary site and rhabdomyosarcoma was the most frequent histology. Patients receiving surgery alone had the highest survival; however, this may be a reflection of smaller, less aggressive tumors preferentially being treated with surgery alone. © 2017 Elsevier B.V. All rights reserved.
Keywords: Pediatric carcinoma Sinonasal carcinoma Pediatric malignancies Sinonasal malignancies Rhabdomyosarcoma Survival
1. Introduction While less than 5% of head and neck malignancies arise from the nasal cavity and paransasal sinuses, pediatric sinonasal
* Corresponding author. Department of Otolaryngology e Head and Neck Surgery, Neurological Institute of New Jersey, Rutgers New Jersey Medical School, 90 Bergen St., Suite 8100, Newark, 07103, NJ, USA. E-mail address: [email protected] (J.A. Eloy). http://dx.doi.org/10.1016/j.ijporl.2017.04.032 0165-5876/© 2017 Elsevier B.V. All rights reserved.
malignancies (PedsSNM) are even rarer [1,2]. The presenting symptoms of nasal obstruction, facial swelling, ophthalmic complaints, rhinorrhea, epistaxis, rhinosinusitis, and chronic upper respiratory tract infections are nonspecific and can be easily mistaken for more common conditions [1e4]. Diagnosis is often delayed by weeks to years and results in advanced disease stage at presentation [2e5]. PedsSNM have been shown to have distinct epidemiological, clinicopathologic, histological, and survival differences compared to adult sinonasal malignancies, emphasizing the value of examining these malignancies in pediatric patients
98
S.Y. Chung et al. / International Journal of Pediatric Otorhinolaryngology 98 (2017) 97e102
alone [1,2]. Given the rarity of PedsSNM, we used the US National Cancer Institute's Surveillance, Epidemiology, and End Results (SEER) database to identify a large, nationally representative cohort, with long-term surveillance. SEER is a multicenter, population-based record that has collected information on malignancies since 1973, and is exempt from institutional biases. Until recently, case reports, systematic reviews, and single-institution studies largely accounted for the scant literature on PedsSNM [4]. Shapiro et al. conducted the first database analysis of PedsSNM and provided some valuable information about their demographic and clinicopathologic characteristics. Their series was limited to 63 cases, and they could not provide any survival analyses due to their small sample size [3]. Gerth et al. have thus far been the only ones to provide survival analyses using SEER. Our study examines 210 PedsSNM patients, provides updated analyses of any previously reported survival trends, and compares our findings to that of the literature. Furthermore, we investigate the role of other potential prognostic indicators of survival not previously explored, such as histological grade at diagnosis. 2. Materials and methods The SEER 18 database was queried in order to obtain frequency, survival, and incidence data for all pediatric sinonasal tumors of the head and neck between the years 1973 through 2013. Sinonasal lymphoma cases were excluded from our analysis to facilitate staging and survival analyses. The SEER database is maintained by the National Cancer Institute (Bethesda, Maryland) and provided all the analyzed information including patient demographics, tumor characteristics, and incidence and survival rates. The SEER 18 database is an amalgamation of data collected from 18 registries from Connecticut, New Jersey, Georgia, Louisiana, Kentucky, New Mexico, Utah, California, Hawaii, Alaska, Oklahoma, Arizona, Michigan, Washington, and Iowa. Some of the registries are statewide registries, as in the case of New Jersey and Connecticut state registries. Others cover a particular metropolitan area, as in the case of the Los Angeles and San Franscico-Oakland registries. Overall, the database captures 28% of the population in the US. The data are void of any private identifiable health records, thus this study has been exempt from Institutional Review Board (IRB) approval as per the standing policy of the IRB of Rutgers New Jersey Medical School, Newark, New Jersey. 2.1. Patient and tumor characteristics International Classification of Disease for Oncology, 3rd Edition (ICD-O-3) topography codes corresponding to all sinonasal tumors were initially used to query case data from the SEER 18 database (C30.0, C31.0, C31.1, C31.2, C31.3, C31.8, C31.9). This dataset was further confined to only include patient cases who were under 18 years of age. The resultant data were then stratified by age groups, gender, race, histology, and primary site. Tumor properties were further analyzed by grade, extension of disease, and treatment modalities. All unknown values were included in the reported results; however, they have been omitted from distribution calculations. 2.2. Statistical analysis Pediatric sinonasal tumors diagnosed between 2000 and 2013 were reported for incidence trends as reported per 100,000 when adjusted to the standard 2000 US population (Census P25-1130). Calculated annual percent change (APC) was observed with 1year end points. Weight least squares were utilized for APC significance testing.
SEER*Stat 8.3.2 (National Cancer Institute, Bethesda, MD) was used to extract all patient data, which were subsequently organized on Microsoft Excel 2016 spreadsheets (Microsoft Corporation, Redmond, WA). Survival analysis for 5-, 10-, and 20-years was then accomplished using the binary scheme utilized in Kaplan-Meier analysis. All untraced patients were excluded from analyses, while traced patients were assigned ones and zeros based on their disease-specific survival (DSS) by the end of each time period, respectively. Demographic frequency distributions did however include all patient cases. Chi-square tests were used to compare categorical variables. JMP Statistical Discovery 12.2 (SAS Institute, Cary, NC) was then utilized to calculate log-rank analysis and Kaplan-Meier curves for all resultant data. Significance levels were set to f ¼ 0.05 for all tests. 3. Results 3.1. Patient and tumor characteristics A total of 210 PedsSNM patients, 18 years old or younger, were identified from the SEER database (Table 1). The mean age at diagnosis was 10.7 years with a standard deviation of 5.64 years. By gender, 45.7% were male and 54.3% were female. By race, 72.6% were white, 21.2% were black, 5.3% were Asian/Pacific Islander, and 1.0% were American Indian/Alaskan Native. This distribution represents prevalence only. By primary site, the nasal cavity was the most commonly involved subsite (37.1%), followed by the maxillary sinus (27.6%), and ethmoid sinus (17.1%). By histology, rhabdomyosarcoma was the most common type (50.5%), followed by olfactory neuroblastoma (17.6%), sarcoma not otherwise specified (NOS) (11.4%), and squamous cell carcinoma (2.9%). The remaining distribution of primary site and histologies is available in Table 2. Anatomic site distribution varied significantly by histology (Table 3) (p < 0.0001). Most olfactory neuroblastomas were in the nasal cavity (73.0%). Sarcoma NOS was more commonly seen in the maxillary sinus (41.7%), and rhabdomyosarcomas were distributed almost evenly throughout the different sinonasal sites. The most common histology varied by age (p ¼ 0.025). In the 0e3 year age group, rhabdomyosarcoma (76.0%) was by far the most common histology. However, in the 12e15 year age group, olfactory neuroblastoma (40.9%) was almost as common as rhabdomyosarcoma (45.5%). The distribution of tumor grade among cases demonstrated a propensity for grade IV disease (37.3%), followed by grade II (25.3%), grade III (22.7%), and grade I (14.7%) (Table 4). Table 1 Demographics of pediatric sinonasal malignancies.
Mean age of diagnosis ( ± SD) Total Age groups (years) 0e3 4e7 8e11 12e15 16e18 Gender Female Male Race White Black Asian or Pacific Islander American Indian/Alaska Native Unknown
N
%
10.7 ± 5.64 210
e 100.0%
35 27 36 58 54
16.7% 12.9% 17.1% 27.6% 25.7%
114 96
54.3% 45.7%
151 44 11 2 2
72.6% 21.2% 5.3% 1.0% e
S.Y. Chung et al. / International Journal of Pediatric Otorhinolaryngology 98 (2017) 97e102 Table 2 Frequency of pediatric sinonasal malignancies by site and histology. N Primary Site Nasal Cavity Maxillary sinus Ethmoid sinus Accessory sinus, NOS Sphenoid sinus Overlapping lesion of accessory sinuses Frontal sinus Histology Rhabdomyosarcoma Olfactory neuroblastoma Sarcoma NOS Squamous cell carcinoma Adenocarcinoma Neuroblastoma NOS Yolk sac tumor Carcinoma NOS Lymphoepithelial carcinoma Peripheral neuroectodermal tumor Spindle cell tumor Adenoid cystic carcinoma Neuroendocrine carcinoma NOS Small cell carcinoma NOS Transitional cell carcinoma NOS Mucoepidermoid carcinoma Acinar cell carcinoma Malignant rhabdoid tumor Malignant myoepithelioma Teratoma (malignant) NOS Chondroid chordoma Primitive neuroectodermal tumor Olfactory neurocytoma Neurilemoma (malignant) (OBS)
99
Table 4 Histological grade of pediatric sinonasal malignancies. %
78 58 36 21 10 7 0
37.1% 27.6% 17.1% 10.0% 4.8% 3.3% e
106 37 24 5 4 4 4 3 3 3 2 2 2 1 1 1 1 1 1 1 1 1 1 1
50.5% 17.6% 11.4% 2.4% 1.9% 1.9% 1.9% 1.4% 1.4% 1.4% 1.0% 1.0% 1.0% 0.5% 0.5% 0.5% 0.5% 0.5% 0.5% 0.5% 0.5% 0.5% 0.5% 0.5%
NOS ¼ not otherwise specified.
3.2. Incidence and survival The incidence of PedsSNM from 2000 to 2013 was 0.036 per 100,000. Overall 5-, 10-, and 20-year DSS rates were 60.2%, 46.1%, and 20.6%, respectively (Table 5 and Fig. 1). Gender, race, and primary site of PedsSNM did not confer any survival differences at 5, 10, and 20 years. At all time intervals studied, DSS decreased as tumor grade increased (P < 0.05). When evaluating PedsSNM by extension of disease, localized cases conferred the best 5-, 10-, and 20-year DSS, while regional involvement reflected significantly worse survival. Distant disease was associated with the worst DSS at all time intervals (P < 0.01). After 20 years, 0% of distant disease cases survived, whereas 28.2% of cases with regional disease and 33.3% of cases with localized disease survived (Table 5). DSS rates were calculated based on the three most common histologies: rhabdomyosarcoma, olfactory neuroblastoma, and sarcoma NOS (Table 6). There were no significant differences in survival between these subtypes for all time intervals studied. Stratifying DSS by therapeutic modality demonstrated that individuals treated with surgery alone had the highest survival rates at 5, 10 and 20 years (80.7%, 73.9%, and 50.0%, respectively), followed by neither surgery nor radiotherapy (71.4%, 50.0%, and 33.3%, respectively), combination therapy with surgery and radiotherapy
Grade
N
%
Well differentiated; Grade I Moderately differentiated; Grade II Poorly differentiated; Grade III Undifferentiated; anaplastic; Grade IV Unknown
11 19 17 28 135
14.7% 25.3% 22.7% 37.3% e
(65.5%, 47.6%, and 25.3%, respectively), and lastly radiotherapy only (51.1%, 38.9%, and 8.0%, respectively). However, these differences were only statistically significant at 20 years (P ¼ 0.0425) (Table 7 and Fig. 2). The histologic breakdown of patients receiving each therapeutic modality differed significantly (p < 0.0001). Among patients receiving surgery and radiotherapy, rhabdomyosarcoma made up 40.6%, olfactory neuroblastoma made up 37.5%, and sarcoma NOS made up 4.7%. Among patients receiving radiotherapy alone, rhabdomyosarcoma made up 86.0%, olfactory neuroblastoma made up 2.0%, and sarcoma NOS made up 0.0%. Among patients receiving surgery alone, rhabdomyosarcoma made up 30.0%, olfactory neuroblastoma made up 13.3%, and sarcoma NOS made up 26.7%. Among patients receiving neither surgery nor radiotherapy, rhabdomyosarcoma made up 25.0%, olfactory neuroblastoma made up 25.0%, and sarcoma NOS made up 12.5%. 4. Discussion PedsSNM are rare, heterogeneous malignancies that behave aggressively and are associated with poor survival. The current study included 210 PedsSNM patients within a 40-year period. The mean age at diagnosis was 10.7 years with a standard deviation of 5.64 years, consistent with other mean ages reported in prior series [1e4]. A small case series initially reported a slight male predominance in PedsSNM [2]. When Turner et al. first examined sinonasal malignancies, a male to female ratio of 1.8:1 was reported [6]. However, they demonstrated a significantly decreasing incidence trend in males, and a stagnant incidence in females [6]. Gerth et al. subsequently reported a male to female incidence ratio of 1.2:1, and that the incidence rate in males was lower than in females [1]. Our study demonstrates the continuation of these incidence trends; the male to female ratio was 0.84:1. By race, most of our PedsSNM patients were white (72.6%), followed by black (21.2%), likely reflecting the racial distribution of the general US population. In our survival analysis, sex and ethnicity did not affect 5-, 10-, or 20-year DSS, as observed in prior studies [1,6]. PedsSNM most commonly arose from the nasal cavity, followed by the maxillary sinus, and ethmoid sinus. Gerth and colleagues did not find significant survival differences depending on primary site of PedsSNM [1]. Likewise, our analysis yielded the same findings. Most of the current literature demonstrates rhabdomyosarcoma to be the most common histological subtype of PedsSNM [1e5]. One contradictory study utilizing SEER found olfactory neuroblastoma to be the most common diagnosis, followed by rhabdomyosarcoma [4]. The present analysis agrees with the majority of previous studies; rhabdomyosarcoma was the most common PedsSNM subtype. Furthermore, there were no significant
Table 3 Site distribution by histology. n(%)
Nasal cavity
Maxillary sinus
Ethmoid sinus
Sphenoid sinus
Overlapping lesion of sinuses
Accessory sinus, NOS
Rhabdomyosarcoma Olfactory Neuroblastoma Sarcoma NOS
29(27.4%) 27(73.0%) 8(33.3%)
25(23.6%) 4(10.8%) 10(41.7%)
20(18.9%) 5(13.5%) 5(20.8%)
9(8.5%) 0 0
6(5.7%) 0 1(4.2%)
17(16.0%) 1(2.7%) 0
100
S.Y. Chung et al. / International Journal of Pediatric Otorhinolaryngology 98 (2017) 97e102
Table 5 Disease specific survival rates by demographics, tumor grade, and primary site. 5-Y
Overall Survival Gender Female Male Ethnicity White Black Asian or Pacific Islander American Indian/Alaska Native Unknown Grade Well differentiated; Grade I Moderately differentiated; Grade II Poorly differentiated; Grade III Undifferentiated; anaplastic; Grade IV Unknown Primary Site Nasal Cavity Maxillary sinus Ethmoid sinus Accessory sinus, NOS Sphenoid sinus Overlapping lesion of accessory sinuses Extension of Disease Localized Regional Distant a
10-Y
20-Y
Na
%
Na
%
Na
%
166 (P ¼ 0.5859) 91 75 (P ¼ 0.7104) 124 30 9 2 1 (P ¼ 0.0001) 9 13 12 22 110 (P ¼ 0.3047) 58 48 29 20 7 4 (P ¼ 0.0005) 22 82 33
60.2%
127 (P ¼ 0.1934) 73 54 (P ¼ 0.5185) 93 26 7 1 0 (P < 0.0001) 7 7 8 19 86 (P ¼ 0.6979) 38 38 22 19 6 4 (P ¼ 0.0001) 15 61 26
46.1%
89 (P ¼ 0.2014) 50 39 (P ¼ 0.6750) 64 20 4 1 0 (P ¼ 0.0480) 1 4 5 18 61 (P ¼ 0.6575) 24 28 16 16 4 1 (P ¼ 0.0021) 6 42 21
20.6%
61.9% 58.3% 63.1% 49.3% 55.6% 50.0% 100.0% 100.0% 84.6% 58.3% 27.3% 60.9% 72.3% 52.8% 61.2% 45.0% 42.9% 75.0% 86.4% 65.3% 39.4%
50.8% 39.6% 48.2% 41.2% 42.9% 0.0% e 100.0% 71.4% 37.5% 10.5% 48.3% 54.8% 39.8% 43.5% 42.1% 33.3% 75.0% 80.0% 51.3% 19.2%
25.1% 14.7% 23.6% 16.9% 0.0% 0.0% e 100.0% 50.0% 0.0% 5.6% 23.8% 27.5% 16.4% 20.6% 25.0% 0.0% 0.0% 33.3% 28.2% 0.0%
Not all cases had survival data; therefore, the number of cases may be less than in Tables 1e3.
Fig. 1. Kaplan-Meier curve of cumulative 20-year disease-specific survival.
differences in 5-, 10-, or 20-year DSS when stratified by histology, which supports Gerth's findings [1]. The greatest number of cases presented as a grade IV disease (37.3%), followed by grade II (25.3%), grade III (22.7%), and grade I
(14.7%), emphasizing the aggressive nature of PedsSNM at presentation. The present study suggests that higher tumor grade has a negative effect on survival. All grade I disease cases survived 20 years. However, for grade II disease, 5-year survival dropped by
S.Y. Chung et al. / International Journal of Pediatric Otorhinolaryngology 98 (2017) 97e102 Table 6 Disease specific survival rates of most common histologies. Histology
5-Y Na
Rhabdomyosarcoma Olfactory neuroblastoma Sarcoma NOS
10-Y Na
%
20-Y %
Na
%
(P ¼ 0.3083)
(P ¼ 0.5461)
(P ¼ 0.0967)
86 32 14
71 21 11
54 12 7
52.7% 68.8% 71.4%
41.0% 47.6% 63.6%
21.6% 8.3% 42.9%
a Not all cases had survival data; therefore, the number of cases may be less than in Tables 1e3.
Table 7 Disease specific survival by therapy. 5-Y N
Surgery and radiotherapy Radiotherapy Surgery No surgery or radiotherapy
10-Y %
N
20-Y %
N
%
(P ¼ 0.0852)
(P ¼ 0.0794)
(P ¼ 0.0425)
69 45 31 7
50 36 23 4
36 25 12 3
65.6% 51.1% 80.7% 71.4%
47.6% 38.9% 73.9% 50.0%
25.3% 8.0% 50.0% 33.3%
more than 15%. After 10 and 20 years, DSS dropped further by 29% and 50%, respectively. These trends continue for grade III and IV lesions, with dramatic decreases in DSS at each time interval. AJCC and TNM staging were excluded from our study due to the lack of staging data available prior to 1988. Instead, we utilized older staging descriptors, such as “localized,” “regional,” or “distant” to account for cases dating back to 1973. Cases with distant metastatic disease, and cases with involvement of regional structures or lymph nodes demonstrated dramatically lower 5-, 10-, and 20-year DSS, than cases with localized disease. Consensus on treating PedsSNM does not yet exist;
101
otolaryngologists are currently using histopathologic-specific protocols that are continually changing and were mainly created for the general adult population. Furthermore, extirpation of tumor in the sinonasal region near the skull base is particularly challenging in children, and morbidity from complications can be detrimental and life-altering. Radiotherapy also has the potential to cause serious acute and long-term sequelae including osteoradionecrosis, blindness, mucositis, infection, hypopituitarism, and a second primary malignancy [1,5]. Because available treatment modalities have these associated possible ramifications, physicians and institutions have varying preferences in what they recommend to PedsSNM patients. Our findings revealed that the most common treatment modality used in PedsSNM was combination surgery and radiotherapy, which has been shown to be the treatment modality of choice in institutional studies [4,7]. The next most common therapies were radiotherapy only, surgery only, and neither surgery nor radiotherapy. A previous SEER study by Gerth et al. had stated that any surgery resulted in significantly higher survival compared to no surgery performed throughout their entire study period [1]. In contrast, the present analysis found no significant differences in DSS at 5 and 10 years between treatment groups. After 20 years, however, patients who had undergone any surgery had the highest survival rates versus those who had not. We also found that patients who received radiotherapy, either alone or in combination with surgery, had significantly lower DSS than patients who received surgery alone. Our results could signify that surgery should be the mainstay of treatment in PedsSNM. Also, the longterm morbidities and complications of radiotherapy may outweigh the benefits, and may possibly decrease survival in the pediatric population. However, these interpretations must be taken with great caution; our findings could merely be reflecting that patients receiving radiotherapy alone or surgery and radiotherapy have tumors that are larger, more aggressive, or possess other poor prognostic factors. Furthermore, rhabdomyosarcoma cases were
Fig. 2. Kaplan-Meier curve of cumulative 20-year disease-specific survival by treatment.
102
S.Y. Chung et al. / International Journal of Pediatric Otorhinolaryngology 98 (2017) 97e102
more likely to receive radiotherapy alone; olfactory neuroblastoma cases were more likely to receive surgery and radiotherapy; and sarcoma NOS cases were more likely to receive surgery alone. There was an inverse correlation between the percentage of aggressive histologies in therapeutic modality groups and greater survival rates. Therefore, histology may be one of the main determinants of these survival differences. Interestingly, patients receiving neither surgery nor radiotherapy exhibited significantly higher 20-year survival than those receiving combination therapy or radiotherapy alone. We posit that patients who underwent neither surgery nor radiotherapy, but had higher DSS, may possibly represent patients with disease that was amenable to chemotherapy only. Of note, our findings starkly contradict Gerth's SEER study, which reported that patients receiving no treatment had the very worst prognosis throughout their entire study period. Chemotherapy is thought to have a role in treating PedsSNM [4]; however, SEER does not provide data regarding chemotherapy use, limiting our analysis on treatment modality. For example, multidrug chemotherapy regimens may be effectively used as an adjunctive therapy for pediatric rhabdomyosarcoma, especially in advanced cases [1,2,4,8e10]. Olfactory neuroblastomas have also been demonstrated to respond to chemotherapy [1,8,11,12]. Further limitations of this study include those inherent to the SEER registries. We must assume this database contains accurate diagnoses and coding. Information such as comorbidities, surgical details, such as margin status, and radiotherapy dose for each case are missing from the SEER registries. Despite its limitations, SEER is an important database for studying rare tumors, with strong statistical power and cross-institutional data. 5. Conclusion Pediatric sinonasal malignancies are extremely rare and deserve attention separate from adult sinonasal malignancies. Although shown to historically occur more in males, these malignancies are now slightly predominant in females. They most frequently occurred in patients in their second decade of life. PedsSNM most often presented as grade IV tumors. The most common histology was rhabdomyosarcoma, and the most common primary site was the nasal cavity. While gender, race, primary site, and histological subtype did not affect survival, higher tumor grade and regional
involvement or distant metastasis at presentation predicted significantly worse disease-specific survival. Surgery alone was associated with the most favorable 20-year disease-specific survival rate. However, further investigation is warranted to delineate whether this is a true survival benefit or is merely reflecting smaller, less aggressive tumors that are amenable to surgery alone. Financial disclosures None. Conflicts of interest None. References [1] D.J. Gerth, J. Tashiro, S.R. Thaller, Pediatric sinonasal tumors in the United States: incidence and outcomes, J. Surg. Res. 190 (2014) 214e220. [2] J.S. Yi, G.S. Cho, M.J. Shim, J.Y. Min, Y.S. Chung, B.J. Lee, Malignant tumors of the sinonasal tract in the pediatric population, Acta Otolaryngol. 132 (Suppl 1) (2012) S21eS26. [3] N.L. Shapiro, N. Bhattacharyya, Staging and survival for sinus cancer in the pediatric population, Int. J. Pediatr. Otorhinolaryngol. 73 (2009) 1568e1571. [4] M.M. Benoit, N. Bhattacharyya, W. Faquin, M. Cunningham, Cancer of the nasal cavity in the pediatric population, Pediatrics 121 (2008) e141e145. [5] J.P. Zevallos, K.S. Jain, D. Roberts, A. El-Naggar, E.Y. Hanna, M.E. Kupferman, Sinonasal malignancies in children: a 10-year, single-institutional review, Laryngoscope 121 (2011) 2001e2003. [6] J.H. Turner, D.D. Reh, Incidence and survival in patients with sinonasal cancer: a historical analysis of population-based data, Head. Neck 34 (2012) 877e885. [7] R.P. Tufano, N.A. Mokadam, K.T. Montone, et al., Malignant Tumors of the Nose and Paranasal Sinuses: Hospital of the University of Pennsylvania Experience 1990-1997, Am. J. Rhinol. vol. 13 (1999) 117e123. [8] P. LoRusso, E. Tapazoglou, J.A. Kish, et al., Chemotherapy for paranasal sinus carcinoma. A 10-year experience at Wayne State University, Cancer 62 (1988) 1e5. [9] M.B. Gillespie, D.T. Marshall, T.A. Day, A.O. Mitchell, D.R. White, J.C. Barredo, Pediatric rhabdomyosarcoma of the head and neck, Curr. Treat. Options Oncol. 7 (2006) 13e22. [10] S. Sanghvi, P. Misra, N.R. Patel, E. Kalyoussef, S. Baredes, J.A. Eloy, Incidence trends and long-term survival analysis of sinonasal rhabdomyosarcoma, Am. J. Otolaryngol. 34 (2013) 682e689. [11] M.M. Fitzek, A.F. Thornton, M. Varvares, et al., Neuroendocrine tumors of the sinonasal tract. Results of a prospective study incorporating chemotherapy, surgery, and combined proton-photon radiotherapy, Cancer 94 (2002) 2623e2634. [12] J.M. Sheehan, J.P. Sheehan, J.A. Jane Sr., R.S. Polin, Chemotherapy for esthesioneuroblastomas, Neurosurg. Clin. N. Am. 11 (2000) 693e701.