Adenoid cystic carcinoma of the parotid and submandibular glands in children and young adults: A population-based study

Abstract

Objectives.

This study aims to analyze the behavior and treatment of adenoid cystic carcinoma (AdCC) in the pediatric and young adult population and to identify factors affecting overall survival (OS).

Materials and Methods.

The study analyzed salivary gland malignancy in patients aged 0–21 with AdCC histology using the National Cancer Database from 2004–2018.

Results.

A total of 72 patients (59.7% parotid, 36.1% submandibular, 1.4% sublingual, 2.8% unspecified) met criteria. Median age was 18 years [range: 0–21]. High grade dysplasia was present in 67% of cases. Therapy consisted of primary surgery for all cases, regional lymph node dissection (74%), radiotherapy (71%), chemotherapy (8%), and chemoradiation (7%). The 5-year OS rate was 93.2% (95% confidence interval (CI): 86.9–99.9%), respectively.

Patients who underwent associated lymph node dissection (LND) had improved OS (p=0.0083, log-rank test) with a 5-year OS at 82.4% (95%CI: 66.1–100%) vs. 97.6% (95%CI: 93.0–100%). A significant difference in OS was found with unfavorable outcomes after positive marginal status: 5Y OS 84.1% (95%CI: 71.0–99.7%) vs. 100% (95%CI: 100%; p<0.001). Adjuvant therapy did not seem to impact outcome.

Conclusion.

This study confirms that AdCC in children and young adults has an overall good prognosis despite frequent high grade. It suggests that cervical lymph node dissection may be of importance, but the value of systematic adjuvant therapy is not confirmed. These findings emphasize the importance and relevance of population-based studies in shaping clinical practice and informing the design of future prospective investigations.

Introduction

According to the American Cancer Society, salivary gland malignancy represents a relatively rare type of cancer, accounting for 6–8% of all head and neck cancers in the United States. The incidence of salivary gland malignancy varies depending on the type of cancer, location, and age. While the incidence of salivary gland malignancy in adults is approximately 1.5 cases per 100,000 per year, salivary gland malignancies are extremely rare in children with 0.2 to 0.8 cases per 1 million children per year [1–2]. Histology and biological behavior of salivary gland tumors differ with age and location. The two most frequent histologies in adult salivary gland malignancies are mucoepidermoid carcinoma and adenoid cystic carcinoma (AdCC), while in children, the most common histologies are mucoepidermoid carcinoma (46%) and acinic cell carcinoma (32%), followed by AdCC [3–5]. Overall, AdCC is a rare malignancy, with an incidence of approximately 1.2–4.4 cases per million per year in the general population, accounting for approximately 10–15% of all pediatric salivary gland malignancies [3–7]. Despite early descriptions, literature regarding AdCC’s characteristics in children is sparse. AdCC in children shares many histopathological features with AdCC in adults, including its biphasic pattern of basaloid and myoepithelial cells, tendency to infiltrate surrounding tissue, and propensity for local recurrence [8]. However, some differences have been noted in tumor location and molecular characteristics. AdCC in children is more commonly located in the head and neck region, particularly in the salivary glands, while AdCC in adults can also occur in the breast, lacrimal gland, lung, and other locations [9].

The National Comprehensive Cancer Network^® (NCCN^®) and the American Society of Clinical Oncology recommend wide tumor resection followed by adjuvant radiotherapy (aRT) for all AdCCs, regardless of stage or grade, due to the high-risk nature of this histology [6,10–11], with studies suggesting long-term follow-up should extend past 15 years [12]. The role of systematic lymph node dissection (LND) in children is still under debate [6,10]. In comparison to adult patients in high-risk groups, as defined by the NCCN: adenoid cystic histology, intermediate/high grade, T3/T4, positive margins, and/or lymph node involvement, children have been shown to have better overall survival (OS) rates [10, 13]. Furthermore, several studies have been conducted to identify prognostic factors for AdCCs in adults, including clinical stage, solid/higher grade, positive margin status, lymphovascular and/or perineural invasion, but there is a lack of similar research in the pediatric population [14–19]. In adults, AdCC is frequently considered as a high grade carcinoma poorly sensitive to conventional chemotherapy that requires frequently adjuvant radiotherapy in case of histologic high grade features [20–22]. Due to its rarity (around 5% of all salivary gland carcinomas in the recent European pediatric experience), few studies have analyzed this tumor in children.

This study aims to scrutinize the behavioral patterns and therapeutic interventions in young patients with salivary glands AdCC registered in the National Cancer Database and identify factors affecting OS in this disease [5, 23–24].

Materials and Methods

Study Design and Data Source

This study was conducted retrospectively and was exempt from institutional board review by our quaternary center Institutional Review Board (#21–08528-XM). De-identified data for this study were obtained from the National Cancer Database (NCDB), which is a clinical oncology database jointly created by the American College of Surgeons Commission on Cancer (CoC) and the American Cancer Society to improve the quality of cancer care. The NCDB sources data from hospital registries of over 1,500 CoC accredited facilities. Neither the NCDB nor the represented hospitals are responsible for the statistical analyses or conclusions presented in this study [25].

Study Population Selection

The NCDB was queried for patients 0 to 21 years old with a diagnosis of salivary gland malignancy with adenoid cystic carcinoma histology from 2004 through 2018. Tumors were defined by ICD-O-3 codes (International Classification of Disease for Oncology, Third Edition; World Health Organization) and narrowed to the histology type adenoid cystic carcinoma (ICD-O-3 code 8200) [26]. The following primary sites were included: parotid (ICD-O-3 code C079), submandibular (C080), and sublingual (C081), and any malignant neoplasm of major salivary gland, unspecified (C089).

Data Elements

Demographics & Social Determinants of Health

The study evaluated various demographics and social determinants of health: age, race, gender, insurance status, quartile median household income for the patient’s area of residence, quartile measure of educational attainment for the patient’s area of residence (percentage of the population without a high school degree), urban/rural status, Charlson-Deyo value, and crowfly distance (the “great circle” distance in miles between the patient’s residence and the reporting hospital) [27]. Due to the disease’s rarity, for analysis purpose race was grouped into black, white, other (Asian, Oriental, Samoan), and unknown. Insurance status was grouped into government (Medicaid, Medicare, other government), not insured, private insurance/managed care, and unknown.

Tumor & Treatment Characteristics

The study assessed several tumor and treatment characteristics: histological grade, primary site, local node involvement, stage, primary margins status, lymph node dissection (LND), and adjuvant therapies. Histological grade is defined by ICD-O-3 codes: with “well”, “moderately”, and “poorly” differentiated approximated to grades I, II, III; and “anaplastic” and “undifferentiated” correlated to grade IV [26]. The tumors were categorized according to the American Joint Committee on Cancer (AJCC) staging system that was being utilized during the time of diagnosis. Upon final staging, the tumors falling within stages 4A, 4B, and 4C were consolidated into a solitary grouping denoted as Stage 4. Extent of cervical LND was further categorized in 3 classes, based on the number of lymph nodes removed during the operation: no dissection (0 lymph node), limited dissection (1–3 lymph nodes), and full dissection (>3 lymph nodes).

Outcomes

The primary outcomes of interest included risk factors affecting overall survival (OS) rates. Vital status from the NCDB identified the event for OS, defined as alive, dead, or unknown at the time of last contact according to the NCDB data dictionary [27]. Secondary outcomes of interest included associations with social determinants of health, tumor characteristics, and treatment characteristics.

Statistical analysis

Chi-squared test (or Fisher’s exact test if any expected count was <5) was used for testing association between two factors (p<0.05). Continuous data were expressed as medians, with a Wilcoxon rank-sum test used for testing differences between two groups, and a Kruskal–Wallis test used for testing differences among three or more groups (p<0.05).

Kaplan-Meier curves to report the months from diagnosis to death according to histology and other stratification criteria. Log-rank test was used for comparing survival curves. Five-year and ten-year survival rates were calculated and compared between the groups using Chi-squared test. Variables tested were as follows: primary site (parotid versus submandibular), gender, race (black versus white), age group (0–15 years versus 16–21 years), insurance status (government versus private), stage (low stages T1/T2 versus high stages T3/T4), primary margins status, lymph node dissection status (no lymph nodes versus at least one lymph node), and adjuvant therapy given (none vs. adjuvant therapy with chemotherapy and/or radiation). Univariate logistic and Cox regression were used for assessing significant associations (p<0.05) between prognostic factors and death. All analyses were performed using R software (R Foundation for Statistical Computing, 4.2.1).

Results

Demographics

A total of 967 patients with salivary gland malignancy aged between 0 and 21 years old were identified in the NCDB. AdCC accounted for 72 (7.4%) of the overall salivary gland malignancy cases. Of the AdCC cases, 63.9% were female and 59.7% presented with a malignancy in their parotid gland, 36.1% in the submandibular gland, 1.4% in the sublingual gland, and 2.8% of tumors were found in other unspecified salivary gland locations (Table 1).

Table 1:

Demographics, tumor, and treatment characteristics data for AdCCs by primary tumor site in a pediatric and young adult population.

	Primary Tumor Site	Parotid	Submandibular	Sublingual	Unspecified	Total
		N=43 (59.7%)	N=26 (36.1%)	N=1 (1.14%)	N=2 (2.8%)	N=72
Demographics	Age (Median [Min, Max])	18 [0, 21]	17 [8, 21]	18 [18, 18]	15 [13, 17]	18 [0, 21]
	Age Group
	A: 0–5	1 (2.3%)	-	-	-	1 (1.4%)
	B: 6–10	2 (4.6%)	1 (3.8%)	-	-	3 (4.2%)
	C: 11–15	9 (20.9%)	6 (23.1%)	-	1 (50.0%)	16 (22.2%)
	D: 16–21	31 (72.1%)	19 (70.1%)	1 (100%)	1 (50.0%)	52 (72.2%)
	Race
	White	27 (62.8%)	21 (80.8%)	-	1 (50.0%)	49 (68.1%)
	Black	8 (18.6%)	2 (7.7%)	1 (100%)	-	11 (15.3%)
	Other	4 (9.3%)	2 (7.7%)		-	6 (8.3%)
	Unknown	4 (9.3%)	1 (3.8%)	-	1 (50.0%)	6 (8.3%)
	Gender
	Female	28 (65.1%)	17 (65.4%)	1 (100%)	-	46 (63.9%)
	Male	15 (34.9%)	9 (34.6%)	-	2 (100%)	26 (36.1%)
Tumor & Treatment Characteristics	AJCC Staging
	1	10 (23.3%)	8 (30.8%)	-	-	18 (25.0%)
	2	9 (20.9%)	13 (50.0%)	-	1 (50.0%)	23 (31.9%)
	3	10 (23.3%)	2 (7.7%)	-	-	12 (16.7%)
	4	8 (18.6%)	3 (11.5%)	1 (100%)	1 (50.0%)	13 (8.1%)
	Unknown	6 (14.0%)	-	-	-	6 (8.3%)
	Tumor size	30 (8, 89)	32 (11,85)	40	42.5 (40–45)	30 (8, 90)
	mm (Median, Min, Max)
	Positive cervical lymph nodes	4 (5.5%)	1 (1.4%)	1 (1.4%)	-	6 (8.33%)
	Metastasis	1 (1.4%)	2 (2.8%)	-	-	3 (4.2%)
	Grade
	Well differentiated	6 (14.0%)	5 (19.2%)	-	-	11 (15.3%)
	Moderately differentiated	7 (16.3%)	3 (11.5%)	1 (100%)	-	11 (15.3%)
	Poorly differentiated	-	2 (7.7%)	-	-	2 (2.8%)
	High grade dysplasia	30 (69.8%)	16 (61.5%)	-	2 (100%)	48 (66.7%)
	Positive margins	16 (22.2%)	13 (18%)	1 (1.4%)	-	30 (4.2%)
	Lymph Node Dissection
	No	10 (23.3%)	7 (26.9%)	-	1 (50.0%)	18 (25.0%)
	Limited	27 (62.8%)	13 (50.0%)	1 (50.0%)	-	41 (56.9%)
	Full	6 (14.0%)	6 (23.1%)	-	-	12 (16.7%)
	Unknown	-	-	-	1 (50.0%)	1 (1.4%)
	Radiation
	Yes	35 (81.4%)	16 (61.5%)	-	-	51 (70.8%
	No	8 (18.6%)	10 (38.5%)	1 (100%)	2 (100%)	21 (29.2%)
	Chemotherapy
	Yes	4 (9.3%)	2 (7.7%)	-	-	6 (8.3%)
	No	39 (90.7%)	24 (92.3%)	1 (100%)	2 (100%)	66 (91.7%)
	Overall Survival
	Yes	39 (90.7%)	22 (84.6%)	1 (100%)	2 (100%)	64 (88.9%)
	No	3 (7.0%)	4 (15.4%)	-	-	7 (9.7%)
	Unknown	1 (2.3%)	-	-	-	1 (1.4%)

The median age was 18 years (range, 0 to 21). There were no statistically significant differences in demographic data between patients with parotid and submandibular gland tumors.

Treatment characteristics

Details of tumor and treatment characteristics are presented in Table 1. Therapy consisted of primary surgery for all cases associated with a cervical lymph node dissection (53 cases), radiotherapy (51 cases), chemotherapy (6 cases), and chemoradiation (5 cases).

A significant association was found between lymph node dissection (LND) and gender for all salivary gland AdCCs (p=0.0198, Table 2); however, there was no significant association with stage (p=0.0514). For all salivary gland AdCCs, histological grade and margin status associations were also tested, and no statistically significant association was found (p=0.857).

Table 2:

Association testing for all salivary gland AdCCs for lymph node dissection with tumor stage, gender, and overall survival (OS).

Lymph node dissection	No (0 nodes) N=18	Limited (1–3 nodes) N=41	Yes (>3 nodes) N=12	Unknown N=1	P-value
AJCC Tumor Stage
1	5 (27.8%)	10 (24.4%)	3 (25.0%)	0 (0%)	0.0514
2	3 (16.7%)	16 (39.0%)	3 (25.0%)	1 (100%)
3	3 (16.7%)	9 (22.0%)	0 (0%)	0 (0%)
4	4 (22.2%)	3 (7.3%)	6 (50.0%)	0 (0%)
Unknown	3 (16.7%)	3 (7.3%)	0 (0%)	0 (0%)
Gender
Female	7 (38.9%)	31 (75.6%)	8 (66.7%)	0 (0%)	0.0198
Male	11 (61.1%)	10 (24.4%)	4 (33.3%)	1 (100%)
Overall Survival
Alive	13 (72.2%)	39 (95.1%)	11 (91.7%)	1 (100%)	0.022
Dead	5 (27.8%)	1 (2.4%)	1 (8.3%)	0 (0%)
Missing	0 (0%)	1 (2.4%)	0 (0%)	0 (0%)

For parotid gland AdCCs, a significant association was found between LND and gender (p=0.024), with females receiving a higher rate of LND. Furthermore, LND again did not have a significant association with stage (Table 3). No significant associations were observed regarding treatment characteristics for submandibular AdCCs.

Table 3:

For parotid gland AdCCs, lymph node dissection compared to overall survival broken out by gender and AJCC tumor stage.

Lymph Node Dissection	No (0 nodes)	Limited (1–3 nodes)	Yes (>3 nodes)
	N=10	N=27	N=6
Alive
Female
1	-	4 (14.8%)	-
2	-	6 (22.2%)	-
3	1 (10.0%)	6 (22.2%)	-
4	-	4 (14.8%)	4 (66.7%)
Male
1	-	3 (11.1%)	1 (16.7%)
2	1 (10.0%)	1 (3.7%)	-
3	2 (20.0%)	1 (3.7%)	-
4	3 (30.0%)	1 (3.7%)	1 (16.7%)
Dead
Female
1	1 (10.0%)	-	-
4	1 (10.0%)	-	-
Male
1	1 (10.0%)	-	-
Unknown
Female
2	-	1 (3.7%)	-

Tumor characteristics & Social Determinants of Health

Regarding the 43 patients with parotid AdCCs, all well-differentiated grade, or grade I records, were from patients that were identified as white race, with 22.2% of all white race malignancies classified as well-differentiated. A statistically significant association was observed between histological grade and race, although the statistical significance of the association between histological grade and race of parotid glands was found to be unstable due to limitations in the current method of significance testing.

For the 26 submandibular AdCCs, patients with private insurance accounted for all of the stage 1 and 3 patients, as well as a majority of the stage 2 patients (71.4%). Government insurance patients accounted for the remaining stage 2 patients. Stage 4 patients were split between government and uninsured patients, with zero private insurance stage 4 patients (Table 4). Regarding chemotherapy, 40% of the patients with government insurance received chemotherapy and these were the only two submandibular AdCC patients who received chemotherapy. Insurance status was significantly associated with stage (p=0.0137, Table 4) and chemotherapy administration (p=0.0462). Furthermore, there was no statistically significant association with insurance status for patients who received radiation (p=0.313) or chemoradiation (p=0.231).

Table 4:

Association testing between submandibular AdCC staging against gender and insurance status.

AJCC Tumor Stage	1 (N=8)	2 (N=13)	3 (N=2)	4 (N=3)	P-value
Insurance Status
Private Insurance/Managed Care	8 (100%)	10 (76.9%)	2 (100%)	0 (0%)	0.0137
Government	0 (0%)	3 (23.1%)	0 (0%)	2 (66.7%)
Not Insured	0 (0%)	0 (0%)	0 (0%)	1 (33.3%)
Missing	0 (0%)	1 (7.7%)	0 (0%)	1 (33.3%)
Gender
Female	2 (11.8%)	11 (64.7%)	1 (5.9%)	3 (17.6%)	0.0157
Male	6 (66.7%)	2 (22.2%)	1 (11.1%)	0 (0%)

Further submandibular AdCCs subgroup analysis showed that the majority of males were diagnosed at stage 1 (66.7%), while no males were diagnosed at stage 4. In comparison, the majority of females were diagnosed at stage 2 (64.7%), and all stage 4 diagnoses were female patients, with a notable 17.6% of the female population being diagnosed at stage 4 (Table 4). There was a statistically significant association between stage and gender (p=0.0157, Table 4).

There were no statistically significant differences in social determinants or tumor characteristics data between parotid and submandibular glands.

Outcome

Across all locations of AdCCs, the overall median follow-up duration for the entire study population was 71.75 months. There were a total of 7 deaths: 4 deaths at 5 years and 3 additional deaths at 10 years (Table 1). The 5-year and 10-year OS rate was 93.2% (95% confidence interval (CI): 86.9–99.9%) and 85.0% (95%CI: 75.1–96.3%), respectively. The cause of death was not documented.

The OS of marginal status across all periods was not significant (p=0.22, log-rank, figure 1A). Nevertheless, a significant difference in the OS was found with unfavorable outcomes for positive marginal status: 5 Y OS 84.1% (95% confidential intervals (CI): 71.0–99.7%) vs. 100% (95%CI: 100%; p=0.029).

Figure 1:

Overall survival across all time periods: (A) marginal status (B) lymph node dissection status.

Individuals who underwent LND had significantly better OS (p=0.0083, log-rank, figure 1B). with a 5-year OS: 5 Y OS 82.4% (95%CI: 66.1–100%) without vs 97.6% (95%CI: 93.0–100%) with LND.

No statistically significant differences were observed in the OS between individuals who did not receive adjuvant therapy and those who received adjuvant chemotherapy and/or radiation therapy: 5 Y OS 88.2% (95%CI: 74.2–100%) with no adjuvant therapy vs. 95.1% (95%CI: 88.8–100%; p=0.418) with adjuvant therapy.

There were no significant differences in OS rates between parotid and submandibular glands: 5 Y OS 97.3% (95%CI: 92.2–100%) for parotid gland vs. 85.7% (95%CI: 72.0–100%; p=0.152) for submandibular gland.

No other significant differences were observed in 5 OS outcome rates when comparing gender, age group, insurance status (private versus government), and staging (high and low). Univariate logistic and Cox regression for each factor with the event status, OS, was not conclusive. Neither with forward and backward selection for all variables.

Discussion

This quite large study focused in this specific histotype give important information on this tumor in children. Overall, AdCC represents 7.4% of all registered salivary gland carcinoma in this database and mainly concerns adolescent-young adults age at a median age of 18 years. Due to this rarity, there is limited data on AdCC in pediatric populations, and studies on AdCC in children are often based on case reports or small case series [5, 23]. While AdCC in children has been reported in both males and females, with no significant gender bias, our study revealed a female predisposition. Consistent with prior reports indicating a median age of onset of approximately 11 years in children [28–29], the present findings demonstrated an increased incidence with age.

The available data suggest that individuals of black and other races tend to have intermediate or high-grade tumors, while lacking well-differentiated tumors, noting diagnosis at a later grade compared to white race. Recently, it has been suggested that genetic factors may play a role in the development of AdCC, as there have been reported family cases in first-degree relatives presenting with the MYB::NFIB gene fusion [30].Similar to previous report in the adult population, gender differences were observed in regard to submandibular AdCCs, for which females tended to receive later diagnosis and comprised a larger proportion of late-stage cases compared to males [31–32]. In terms of LND for pediatric parotid AdCCs, males were less likely to have the procedure done, while females had a higher frequency of limited dissection, followed by an increased frequency of full dissection. However, the extent of LND performed did not show any significant association with stage, indicating possible variations in practice among institutions [33]. This highlights the need for greater awareness of potential gender biases and the importance of standardized care for all patients, regardless of gender.

As suggested by previous studies on cancer patients, disparities by health insurance status were revealed, with private insurance patients tending to receive earlier diagnoses than those with government insurance [34–36]. Although government insurance provides coverage for all pediatric patients, it is crucial to ensure that the quality and speed of care is comparable to that of their peers. In terms of treatment, access to specialized care and expertise may be a social determinant of health for children with AdCC, as this is a rare and complex disease that requires multidisciplinary management.

The OS rate in our pediatric cohort confirms that the OS in children is superior to that observed in adults despite frequent high grade histological features in this population (66.7% of high grade dysplasia) [37–39]. However, positive margins are an unfavorable outcome risk factor that could increase the risk of recurrence. Given the poorer local control and survival rates observed in historical studies of salivary gland tumors [40], this finding highlights the impact on overall survival in the pediatric population as well. Notably, there is no association between histological grade and margin status, which reduces the efficacy of preoperative planning and risk mitigation based on grade. Currently, surgical resection can be supplemented with non-targeted fluorescence-guided resection [41]. While IgG-targeted fluorescence resection research advances in head and neck cancers have shown promising survival outcomes [42–44], these findings highlight the importance of further research for targeted resection options for aggressive salivary gland tumors, such as AdCCs.

At this point in time, there is no statistically significant correlation between chemotherapy and/or radiation therapy and overall survival (OS, p >.05). Notably, adjuvant radiation was delivered in more than 70% of patients. Moreover, our findings, which are consistent with a previous study on salivary gland tumors in pediatric patients, indicate that the use of systematic additional radiation therapy and/or chemotherapy for pediatric salivary gland carcinomas does not confer any survival benefit compared to those who did not receive these treatments [37]. Notably, a recent Surveillance, Epidemiology, and End Results (SEER) database analysis for 352 young patients showed that aRT had a favorable impact in primary malignant pediatric parotid tumors, regardless of the tumor type [24]. Therefore, care must be taken to draw conclusions or set guidelines when few events are present. In addition, a recent study of patients with salivary gland AdCCs of all ages found no survival benefit for early-stage AdCCs or minor salivary gland AdCCs with aRT, and only a survival benefit for late-stage submandibular AdCCs [45].

In our experience, the only favorable prognostic factor for OS in young patients with AdCCs is the cervical lymph node dissection. Our data indicates that patients with AdCCs should receive some type of dissection, irrespective of tumor staging. Given individuals with no LND had significantly lower OS, including those with lower stage malignancies, this further reiterates the importance of LND. Notably, the extent of the LND does not appear to impact outcome as survival rates were similar between individuals with limited versus full neck dissections. Taken together, these findings corroborate a recent report suggesting that the prognostic factors that have been identified in the adult population may not be systematically applicable to the pediatric population.

Pediatric patients typically exhibit tumors of lower grades and manifest at earlier stages. Conversely, adults with tumors of comparable type, grade, and stage often encounter concomitant health conditions that may exacerbate with treatment, such as renal failure or pre-existing cardiac ailments. Adults are subject to cofactors like aging, lifestyle, and environmental influences, which can influence survival rates. Nonetheless, even within the adult population, the role of aRT remains a subject of debate [46].

Strengths and Limitations

The NCDB is a valuable tool for studying rare diseases like AdCCs in the pediatric population, and its use in this study lends generalizability to the results. However, due to the rarity of AdCCs, all statistics are limited by small sample size and event counts, and missing data further limits analysis. Additionally, tumor characteristics such as the practice of histological grading is variable among pathologists, and an increased risk exists with a rare pediatric tumor [26]. Furthermore, previous reports in the adult population showed lymphovascular invasion to be a strong prognostic factor [37, 47–50]. Unfortunately, this data was missing from the NCDB. Caution must be exercised in interpreting survival curves as there were only seven events overall, and the database may not capture care received at outside facilities, leading to potential underreporting. Therefore, the study’s findings regarding survival should be taken with care since only seven events were observed in total. Any conclusions drawn from the overall survival analyses using the NCDB should also take into consideration the potential presence of confounding variables and biases [51]. Due to limitations of NCDB, it was also not possible to assess recurrence rate and disease-free survival. Another constraint pertains to the lack of somatic genetic data, which restricts our understanding of the impact of genetic mutations on the behavior of tumors and their prognostic significance in children [52–54]. Institutions are required by the NCDB guidelines to gather follow-up information for 90% of patients over five years. Moreover, the study’s population is restricted to patients who received treatment at CoC hospitals. While this limits the study’s generalizability, the CoC involves prominent hospitals and represents a significant portion of the cancer population in the US, particularly for rare malignancies. However, it must be noted the dataset excludes large hospital systems that are not part of the CoC, and unaccounted-for confounding variables may affect the results [55].

Conclusion

As more pediatric patients with AdCC diagnoses are included in the NCDB, it is expected that certain trends will become clearer, and the power of the findings in this publication will be strengthened. We recommend longer follow-up periods to account for the differences observed in overall survival over time as some late tumor events could occur even 5 years after diagnosis (Figure 1). Some of the associations identified in this study could inform less invasive initial surgical treatments and identify characteristics for long-term care and follow-up. Cervical lymph node dissection in conjunction seems important. The role of systematic aRT is not fully determined. Therefore, more research is required to achieve consensus on clinical decision-making and determine the optimal extent of dissection.

Abbreviations:

AdCC

adenoid cystic carcinoma

OS

overall survival

CI

confidence interval

LND

lymph node dissection

NCCN^®

National Comprehensive Cancer Network^®

aRT

adjuvant radiotherapy

NCDB

National Cancer Database

CoC

American College of Surgeons Commission on Cancer

SEER

Surveillance, Epidemiology, and End Results