Endoscopic Resection of Pediatric Skull Base Tumors: An Evidence-Based Review

Abstract

Objectives  To perform a systematic review examining experiences with endoscopic resection of skull base lesions in the pediatric population, with a focus on outcomes, recurrence, and surgical morbidities.

Methods  PubMed/MEDLINE, Cochrane Library, Embase, and Web of Science databases were evaluated. Studies were assessed for level of evidence. Bias risk was evaluated using the Cochrane Bias tool, Grades of Recommendation, Assessment, Development and Evaluation (GRADE), and Methodological Index for Non-Randomized Studies (MINORS) criteria. Patient characteristics, pathology, site of primary disease, presenting symptoms, stage, procedure specific details, and complications were evaluated. Results were reported using the Preferred Reporting Systems for Systematic Reviews and Meta-Analysis guidelines.

Results  Ninety-three studies met criteria for inclusion, encompassing 574 patients with skull base tumors. The GRADE and MINORS criteria determined the overall evidence to be moderate quality. The most common benign and malignant pathologies included juvenile nasopharyngeal angiofibromas ( n  = 239) and chondrosarcomas ( n  = 11) at 41.6 and 1.9%, respectively. Of all juvenile nasopharyngeal angiofibroma tumors, most presented at stage IIIa and IIIb (25.8 and 27.3%, respectively). Nasal obstruction (16.5%) and headache (16.0%) were common symptoms at initial presentation. Surgical approaches included endoscopic endonasal ( n  = 193, 41.2%) and endoscopic extended transsphenoidal ( n  = 155, 33.1%). Early (< 6 weeks) and late (>6 weeks) complications included cerebrospinal fluid leak ( n  = 36, 17.3%) and endocrinopathy ( n  = 43, 20.7%). Mean follow-up time was 37 months (0.5–180 months), with 86.5% showing no evidence of disease and 2.1% having died from disease at last follow-up.

Conclusion  Endoscopic skull base surgery has been shown to be a safe and effective method of treating a variety of pediatric skull base tumors. If appropriately employed, the minimally invasive approach can provide optimal results in the pediatric population.

Introduction

Advances in endoscopic visualization and instrumentation have led to increasing utilization of minimally invasive approaches among skull base surgeons over the past two decades. Along with this increasing expertise, numerous series have noted decreased complication rates and morbidities associated with endoscopic approaches in appropriately selected lesions. Despite this increasing experience, far less inquiry has been dedicated to examining the appropriate indications as well as outcomes and complications among endoscopic skull base surgery (ESBS) performed on the pediatric population.

Institution-specific series and pathology-specific reviews have examined aspects of endoscopic resection among pediatric patients. One analysis examining outcomes among patients undergoing resection for juvenile nasopharyngeal angiofibroma (JNA) noted endoscopic resection to have significantly lesser intraoperative blood loss and possibly lower recurrence rates, although differences in the latter disappeared when performing analysis using individual patient data. ¹ Multiple other intrainstitutional studies have explored outcomes and complications for pediatric transsphenoidal surgery. ^{2 3 4 5 6} To our knowledge, there have been no systematic reviews evaluating the endoscopic approach for skull base lesions focused on the pediatric population. Our objective was to perform a systematic review examining experiences with endoscopic resection of skull base lesions in the pediatric population, with a focus on outcomes, recurrence, and surgical morbidities.

Methods

This systematic review was based on a comprehensive search of the literature using Medline (1964 to 30 July 2017; using the “Keyword” search function), Pubmed (1964 to 30 July 2017; using the “All Fields” search function), Embase (1964 to 30 July 2017; using the “All Fields” search function), Cochrane Library (2003 to 30 July 2017; using the “All Text” search function), and Web of Science (1990 to 30 July 2017; using the “Topic” search function) using medical subject heading (MeSH) terms related to endoscopic treatment of pediatric skull base tumors both benign and malignant. Our search strategy was as follows:

(“skull base” OR “base, skull” OR “Cranial Base” OR “base, cranial” OR “Basis cranii” OR “Base of Skull” OR “Basicranium”) AND (“pediatric” OR “pediatrics” OR “child” OR “child, preschool” OR “adolescent” OR “children” OR “adolescents”) AND (“Neoplasia” OR “Neoplasias” OR “Neoplasm” OR “Tumors” OR “tumor” OR “Benign Neoplasms” OR “Neoplasms, Benign” OR “Benign Neoplasm” OR “Neoplasm, Benign” OR “Malignancy” OR “Malignancies” OR “Cancer” OR “Cancers”) AND (“Surgical Procedures, Endoscopic” OR “Procedure, Endoscopic Surgical” OR “Procedures, Endoscopic Surgical” OR “Surgical Procedure, Endoscopic” OR “Endoscopy, Surgical” OR “Surgical Endoscopy” OR “Endoscopic Surgical Procedure” OR “Endoscopic Surgical Procedures” OR “endoscopy” OR “endoscopic” OR “endoscopy assisted” OR “endoscopic assisted” OR “endoscope” OR “endoscope assisted” OR “endonasal” OR “microscopy” OR “microscope” OR “microscopy assisted” OR “microscope assisted”)

A detailed keyword search is highlighted in Table 1 . This systematic review followed the guidelines set by the Preferred Reporting Systems for Systematic Reviews and Meta-Analysis (PRISMA) and is outlined in Fig. 1 . We included studies published from 1995 to July 2017 evaluating the use of endoscopic approaches in the management of skull base tumors (both benign and malignant) among patients ranging from newborn to 18 years of age. Patient data for those undergoing endoscopic approaches were collected from all included studies (including those employing both open and endoscopic approaches). Additionally, review articles, papers written in languages other than English, articles detailing the endoscopic method in cadavers or non-human subjects, or in adults were all excluded.

Table 1. Search strategy.

Step	Search term	Articles
1	“skull base” OR “base, skull” OR “Cranial Base” OR “base, cranial” OR “Basis cranii” OR “Base of Skull” OR “Basicranium”	24509
2	“pediatric” OR “pediatrics” OR “child” OR “child, preschool” OR “adolescent” OR “children” OR “adolescents”	3919528
3	“Neoplasia” OR “Neoplasias” OR “Neoplasm” OR “Tumors” OR “tumor” OR “Benign Neoplasms” OR “Neoplasms, Benign” OR “Benign Neoplasm” OR “Neoplasm, Benign” OR “Malignancy” OR “Malignancies” OR “Cancer” OR “Cancers”	3598610
4	“Surgical Procedures, Endoscopic” OR “Procedure, Endoscopic Surgical” OR “Procedures, Endoscopic Surgical” OR “Surgical Procedure, Endoscopic” OR “Endoscopy, Surgical” OR “Surgical Endoscopy” OR “Endoscopic Surgical Procedure” OR “Endoscopic Surgical Procedures” OR “endoscopy” OR “endoscopic” OR “endoscopy assisted” OR “endoscopic assisted” OR “endoscope” OR “endoscope assisted” OR “endonasal” OR “microscopy” OR “microscope” OR “microscopy assisted” OR “microscope assisted”	1191186
5	1 AND 2 AND 3 AND 4	1013

Fig. 1.

Search strategy based on Preferred Reporting Items for Systematic Reviews and Meta-Analyses.

Patient characteristics such as age and sex were collected. Additionally, disease specific qualities such as pathology, site of primary disease, presenting symptoms, stage, follow-up duration, outcome at follow-up, number of recurrences and time to recurrence were collected. Procedure specific details such as procedure performed, surgical approach, surgical time, blood loss, early and late complications, and number of procedures performed were also collected and analyzed. All studies were assessed for bias and level of evidence using the Grades of Recommendation, Assessment, Development and Evaluation (GRADE) criteria, the Cochrane Bias Tool, and the Methodological Index for Non-Randomized Studies (MINORS) criteria.

Statistical Analysis

A chi-square analysis was used for comparison of categorical variables with threshold for significance set at p  < 0.05. SPSS version 20 (an IBM Company, Chicago, Illinois, United States) was used for statistical analysis.

Results

The initial search of the literature yielded 711 original studies, review articles, and abstracts. In total, 93 studies met criteria for inclusion into this qualitative synthesis. Between 1995 and July 2017, 574 patients were identified from 41 retrospective cohort studies, 34 case reports, 15 case series, and 3 prospective cohort studies. No randomized controlled trials were identified in the literature. Using the Cochrane Bias Tool, we judged eight studies to be at low risk, twenty-eight studies at moderate risk, seven studies at high risk, and fifty studies at unclear risk of bias. Using the GRADE and MINORS criteria, level of evidence is considered to be of moderate quality. Level of evidence as well as study conclusions are shown in Supplemental Table S1 (online only).

Patient demographics are presented in Table 2 . In total, 574 patients were included in this systematic review with a mean age of 12.8 and range of newborn—18 years. Of the studies that reported gender, 72.3% were male and 27.7% were female. Follow-up data were available for 352 patients with a mean follow-up time of 37 months and range of 0.5 to 180 months. Regarding disease status at follow-up, 243 patients (86.5%) were found to have no evidence of disease, 31 patients (11.0%) were alive with disease, and 6 (2.1%) patients had died from disease.

Table 2. Summary of included studies.

Characteristic	Data
Number of studies included, n (range of years)	93 (1995–2017)
Study design, n (%)
Retrospective cohort	41 (44.1)
Case reports	34 (36.6)
Case series	15 (16.1)
Prospective cohort	3 (3.2)
Number of patients, n	574
Average age (range in years), median	12.8 (newborn—18), 14
Male, n (%)	415 (72.3)
Female, n (%)	159 (27.7)
Total number of procedures	643
Type of procedure, n (%) (available for 468 patients)
EEA	193 (41.2)
EETA	155 (33.1)
Combined	42 (9.0)
Open	42 (9.0)
ETA	33 (7.1)
ETO	2 (0.9)
MTO	1 (0.2)
Extent of resection, n (%)
GTR	337 (58.7)
STR	83 (14.5)
PR	29 (5.0)
Other	12 (2.1)
Biopsy only	3 (0.5)
Palliative	1 (0.2)
Unspecified	109 (19.0)
Follow-up duration, average (range in months) (available for 352 patients)	37 (0.5–180)
Disease status on follow-up, n (%)	281
NED	243 (86.5)
AWD	31 (11.0)
DOD	6 (2.1)
DOC	1 (0.4)
Total number of recurrence, n (range for patients)	72 (0–4)
Time to recurrence, average (range in months) (available for 24 patients)	19.75 (3–82)

Abbreviations: AWD, alive with disease; DOC, deceased from other causes; DOD, deceased from disease; EEA, endoscopic endonasal; EETA, endoscopic extended transsphenoid; ETA, endoscopic transfacial; ETO, endoscopic transoral; GTR, gross total resection; MTO, microscopic transoral; NED, no evidence of disease; PR, partial resection; STR, subtotal resection.

Aggregate data for endoscopic approaches are shown in Table 2 . Specific procedures were reported for 468 patients, of which 643 procedures were performed. The most common approach included endoscopic endonasal ( n  = 193, 41.2%), endoscopic extended transsphenoidal ( n  = 155, 33.1%), combined ( n  = 42, 9.0%), and open ( n  = 42, 9.0%). In total, JNA was the most common benign pathology included in this study ( n  = 239, 41.6%), followed by craniopharyngiomas ( n  = 89, 15.5%) and pituitary adenomas ( n  = 52, 9.1%). Of all JNA tumors, stage IIIa and IIIb were the most common stage at presentation according to Andrews classification, at 25.8 and 27.3%, respectively ( Fig. 2 ). Chondrosarcomas ( n  = 11, 1.9%) and sarcomas ( n  = 4, 0.7%) were the most common malignant microscopic pathologies ( Table 3 ). Nasal cavity and ethmoid sinus tumors encompassed 55.5 and 16.3% of all anterior skull base tumors, respectively. Of all central skull base tumors, nasopharynx and suprasellar locations were the most common locations at 18.6 and 14.9%, respectively. Posterior cranial fossa tumors most commonly involved the brainstem at 31.3%. Table 4 further illustrates additional areas of tumor involvement.

Fig. 2.

Stage of presentation for juvenile nasopharyngeal angiofibromas (JNA) based on Andrews classification.

Table 3. Microscopic pathology.

Benign	n (%)	Malignant	n (%)
Juvenile nasopharyngeal angiofibroma	239 (41.6)	Chondrosarcoma	11 (2)
Craniopharyngioma	89 (15.5)	Sarcoma, unspecified	4 (1)
Pituitary adenoma	52 (9.1)	Osteoblastoma	3
Chordoma	29 (5.1)	Adenocarcinoma	2
Rathke's cleft cyst	24 (4.2)	Lymphoma	2
Fibro-osseous disease	23 (4.0)	Metastasis	2
Dermoid/Epidermoid	16 (2.8)	Olfactory neuroblastoma (esthesioneuroblastoma)	2
Fibroma	6 (1.0)	Rhabdomyosarcoma	2
Teratoma	5 (1)	Adenoid cystic	1
Glioma	4 (1)	Ewing's sarcoma	1
Schwannoma	4 (1)	Mucoepidermoid carcinoma	1
Histiocytosis	3	Nasopharyngeal carcinoma	1
Cholesterol granuloma	2	Osteogenic sarcoma	1
Ependymoma	2	SNUC	1
Germinoma	2
Hemangioma	2	Other pathology	30 (5.2)
Osteoma	2
Giant cell tumor	1
Infantile myofibromatosis	1
Juvenile pilocytic astrocytoma	1
Leiomyoma	1
Myofibroma	1
Pseudotumor	1

Abbreviation: SNUC, sinonasal undifferentiated carcinoma.

Table 4. Site of primary tumor involvement.

Anterior skull base	n (%)	Central skull base	n (%)
Nasal cavity	116 (14.5)	Nasopharynx	102 (12.8)
Ethmoid sinus	34 (4.3)	Suprasella	82 (10.3)
Orbit	27 (3.4)	Pterygopalatine fossa	66 (8.3)
Anterior fossa	22 (2.8)	Infratemporal fossa	61 (7.6)
Cribriform plate	7 (0.9)	Clivus	56 (7.0)
Frontal lobe	2	Sphenoid sinus	50 (6.3)
Frontal sinus	1	Cavernous sinus	44 (5.5)
		Pituitary fossa	41 (5.1)
		Maxillary sinus	17 (2.1)
Posterior skull base		Parasella	9 (1.1)
Brainstem	5	Middle cranial fossa	7 (0.9)
Meckel's cave	4	Infra-sella	3
Oropharynx	3	Inferior orbital fissure	2
Jugular tubercle	2	Petrous carotid	2
Hypoglossal canal	1	Sphenopalatine foramen	2
Posterior fossa, unspecified	1	Prevertebral space	1
		Parapharyngeal space	1
		Orbital apex	1
Intracranial extension, unspecified	25 (3.1)	Retrosella	1

Table 5 and Fig. 3A outline the most common signs and symptoms of skull base tumors on presentation: nasal obstruction (16.5%), headache (16.0%), endocrinopathy (14.3%), and epistaxis (14.0%). The most common early complications (< 6 weeks) after endoscopy included cerebrospinal fluid (CSF) leak, headache, and infection/sepsis at 17.3, 12.5, and 8.7%, respectively. Endocrinopathy ( n  = 43, 20.7%), chronic nasal symptoms ( n  = 5, 2.4%), and cranial nerve (CN) palsy ( n  = 4, 1.9%) were the most common late complications (>6 weeks) after endoscopy ( Table 6 , Fig. 3B ).

Table 5. Initial signs and symptoms at presentation.

Symptom	n (%)
Nasal obstruction	115 (16.5)
Headache	112 (16)
Endocrinopathy	100 (14.3)
Epistaxis	98 (14)
Visual changes	80 (11.5)
CN palsy	46 (6.6)
Nausea/emesis	17 (2.4)
Rhinorrhea/nasal discharge	15 (2.1)
Proptosis	14 (2.0)
IC pressure	14 (2.0)
Facial swelling	13 (1.9)
Anosmia/hyposmia	10 (1.4)
Dysphagia	6 (0.9)
Hearing loss	6 (0.9)
Otitis media/ET dysfunction	6 (0.9)
Seizure	4 (0.6)
Facial deformity	4 (0.6)
Neck pain	3 (0.4)
CSF leak	3 (0.4)
AMS/syncope	3 (0.4)
Personality changes	3 (0.4)
Sinusitis	3 (0.4)
Respiratory distress	3 (0.4)
Ataxia/dysequilibrium	2 (0.3)
Weakness/fatigue	2 (0.3)
Dizziness/vertigo	2 (0.3)
Draining sinus cyst	2 (0.3)
Neurologic symptoms, unspecified	2 (0.3)
Eye swelling	2 (0.3)
Incidental	2 (0.3)
Brainstem dysfunction, unspecified	1 (0.1)
Tongue atrophy	1 (0.1)
Facial pain	1 (0.1)
Neck mass	1 (0.1)
OSA	1 (0.1)
UE weakness	1 (0.1)

Abbreviation: AMS, altered mental status; CN, cranial nerve; CSF, cerebrospinal fluid; ET, Eustachian tube; IC, intracranial; OSA, obstructive sleep apnea; UE, upper extremity.

Fig. 3.

The most common presenting symptoms of skull base tumors ( A ) as well as the most common complications after endoscopic management ( B ). Abbreviations: N Ob, nasal obstruction; HA, headache; Endo, endocrinopathy; Epi, epistaxis; Vision, visual changes; CN, cranial nerve; N/V, nausea/vomiting; CSF, cerebrospinal.

Table 6. Complications.

Early complications (<6 weeks)	n (%)	Late complications (>6 weeks)	n (%)
CSF leak	36 (17.3)	Endocrinopathy	43 (20.7)
Headache	26 (12.5)	Chronic nasal symptoms (crusting synechia)	5 (2.4)
Infection/sepsis	18 (8.7)	CN palsy	4 (1.9)
Visual disturbance/loss	13 (6.3)	Visual disturbance	3 (1.4)
Endocrinopathy/hypothalamic injury	13 (6.3)	Xerophthalmia	3 (1.4)
CN palsy	11 (5.3)	Alopecia	1 (0.5)
Epistaxis	9 (4.3)	Dacryocystitis	1 (0.5)
Carotid sacrifice/vascular injury	7 (3.4)	ET dysfunction	1 (0.5)
Trismus	4 (1.9)	DVT	1 (0.5)
Nausea/vomiting	4 (1.9)	Mucocele	1 (0.5)
Wound infection	1 (0.5)
Xerophthalmia	1 (0.5)
Meningitis	1 (0.5)
Facial swelling	1 (0.5)

Abbreviations: CSF, cerebrospinal fluid; CN, cranial nerve; DVT, deep vein thrombosis; ET, Eustachian tube.

Juvenile Nasopharyngeal Angiofibroma

A total of 239 patients presented with nasopharyngeal angiofibroma representing 41.6% of all patients in this analysis. The average age of presentation was 13.3 years. The most common presenting feature was epistaxis reported in 79.1% of these patients. In total, 53.5% of patients (128 in total) received embolization prior to surgery. With regard to stage at the time of presentation, a majority of the patients presented at Andrew's stage IIIa and IIIB, 25.8 and 27.3%, respectively, is depicted in Fig. 2 . Average operative blood loss was ∼791cc.

Gross total resection (GTR) was achieved in 79.0% of these patients (132 in all), subtotal resection (STR) in 11.9% (20 patients), and partial resection (PR) in 1.8% of patients (3 patients). Extent of resection was not reported in 84 patients. The most common complications following surgery included CN palsy (3.3%), xerophthalmia (1.7%), and trismus; CSF leak and chronic nasal crusting were each reported in 1.3% of all patients with JNA. Recurrence was reported in 21 patients, comprising 8.7% of patients. Follow-up information was available for only 156 patients, 80.1% of whom were free of disease on follow-up, and another 12.2% were alive with disease.

Craniopharyngioma

A total of 89 pediatric patients with craniopharyngioma underwent endoscopic surgery comprising 15.5% of the total patient population. These were evenly divided between males and females (44 and 45, respectively). The average age of presentation was 10.8 years. Three patients received neoadjuvant radiation and another three received adjuvant radiation, one patient receiving chemotherapy, one receiving chemoradiotherapy, and one receiving adjuvant embolization. GTR was achieved in 71.9% of patients, STR in 23.6% of patients, and 4.5% of patients had no reported extent of resection. A total of 21 recurrences were reported, comprising 23.6% of all treated patients with the disease. In terms of complications reported, 14.6% of patients experienced endocrinopathy, most commonly DI, representing 76.9% of this patient population. Only 6.7% experienced CSF leak, and there was one report of an intracranial arterial injury (P1) causing subarachnoid hemorrhage and small thalamic infarct. Follow-up information was lacking for the majority of these patients; average follow-up was roughly 23 months. For the 24 patients for whom follow-up information was reported, 83.4% of patients were free of disease and 16.6% were alive with disease on follow-up.

Pituitary Adenoma

There were 52 patients with pituitary adenomas, accounting for 9.6% of all patients. The average age of presentation was 13.5 years old. GTR was achieved in 84.6% of patients, STR in 7.7% of patients, PR in 3.8% of patients, and 3.8% of patients did not have reported extent of resection. Endocrinopathies were reported in 11.5% of patients after endoscopic removal of these tumors. The average follow-up period was 16.5 months, 23.1% of patients were free of disease, one patient was alive with disease (1.9%), and 75.0% of patients had no reported outcome on follow-up. There were only three reported recurrences.

Chordoma/Chondrosarcoma

A sum of 40 patients (7.0%) with chordoma and chondrosarcoma were included. These two pathologies were often reported together in the literature, frequently not making distinction in diagnosis; therefore, the decision was made to combine demographic, treatment, and outcome measures. Average age of presentation was 11.0 years with range of 3 to 18 years old. Approximately 17.5% of patients had neoadjuvant radiation therapy (7 patients) and one patient had neoadjuvant chemotherapy. With regard to adjuvant therapy, another 17.5% of patients underwent radiation therapy, one completed chemoradiotherapy and another only chemotherapy. Extent of resection was available for 26 patients, 34.6% of who achieved GTR (9 patients). The majority of these patients, 53.8%, had STR and the rest (11.5%) underwent PR. Patients with these tumors had a relative risk of incomplete endoscopic resection (defined as either STR or PR) of 2.91 (confidence interval, 2.09–4.05), which achieved a p  < 0.05. CSF leak was reported in four patients (10.0%). Moreover, there was one report of a pontine stroke in a patient with chordoma of the clivus. There were five recurrences, with a median time to recurrence of 20.5 months. Outcome at follow-up was reported for 30.0% of these patients. At an average of 40.2 months (median 36 months), 75.0% of patients were free of disease, two patients were dead of disease (16.7% of reported outcomes), and one patient was alive with disease.

Discussion

Major Pathologies and Recurrence

The most common benign pathologies treated with ESBS were JNA, craniopharyngioma, and pituitary adenoma at 41.6, 15.5, and 9.1%, respectively. Malignant pathology constituted only ∼5.9% of the 574 patients treated; of these chondrosarcoma and unspecified sarcoma were most common at 1.9 (11 cases) and 0.7% (4 cases), respectively. These major pathologies are discussed below with comparison to current consensus rates of recurrence. It should be noted that consensus rates often include adult data as pediatric population-based data alone is lacking in the literature.

Although classified as benign, JNAs are locally aggressive vascular neoplasms most commonly found in adolescent males. ¹ Common presenting symptoms include unilateral epistaxis and nasal obstruction, which were both in the top four most common signs and symptoms found in our analysis with 16.5 and 14.0% of cases, respectively. JNAs are best treated by surgical resection which traditionally have been through the open approach. ¹ With the introduction of preoperative embolization and reduced intraoperative blood loss, ESBS has become an acceptable treatment modality today. ^{7 8} The endoscopic approach with the concomitant use of preoperative embolization decreases the need for intraoperative transfusion due to bleeding and therefore increases visualization, possibly allowing better tumor resection and ultimately improved recurrence-free survival. ⁹ Chivukula et al reported in their 21 patient JNA series a 76.2% GTR rate with the endonasal endoscopic surgery approach, with near total or STR in the remaining 23.8%. The STR was due to residual tumor in the cavernous sinus or proximity to the carotid artery with the residual often simply being observed without evidence of further regrowth. ⁹ Of the 239 patients that presented with JNA in our analysis, only 128, or 53.5%, of patients were treated preoperatively with embolization. Even with this relatively low rate of embolization, GTR was still 79.0% (132 patients) with recurrence in 8.7% or 21 patients. Consensus recurrence rates for JNA are between 8 and 36%, variation being attributed to sample size. ^{8 10} Therefore, we can see that ESBS is a viable method of treatment on par with nonendoscopic methods.

The literature on craniopharyngioma suggests a total resection without recurrence rate of 18 to 76% depending on tumor size and surgical approach. ^{11 12 13} Craniopharyngiomas are ideally treated with radical resection due to their high rate of recurrence. However, this is often difficult because of the tumor's deep location and close relationship with important neurovascular structures, raising the risk of neurological and endocrine side effects. In a previous study, a GTR rate of 48.3% was found by the open approach versus 66.9% by an endoscopic approach. ¹⁴ Patel et al in 2017 showed in a brief review of the literature that the endoscopic transsphenoidal approach garnered an average GTR rate of 63.5%. ¹⁵ In our analysis from the 89 patients with craniopharyngioma, GTR was achieved in 71.9% of patients with recurrence being reported in 21 patients, or 23.6%. Additionally, of the 24 patients with follow-up data, 83.3% were disease free at an average follow-up of 23 months. With rates of recurrence reported in the range of 0 to 50% in the literature when GTR is attained and 30 to 100% for all other cases, the calculated rate of recurrence in our analysis is within current standards. ^{16 17}

Pituitary adenomas were the next most common pathology found in our study with 52 patients accounting for 9.1% of the cases. Reported total resection rates in the literature range from 40 to 89%. ^{18 19 20 21} Previously reported recurrence-free survival rates range from 48 to 97% at 5 years and 14 to 56% at 10 years. ^{22 23 24 25 26 27} Chang et al also showed that GTR resulted in far lower rates of recurrence than STR alone, 11 versus 45% at 10 years. ²⁸ Our series showed a GTR of 84.6% and only 3 (5.8%) reported recurrences. These positive results of tumor resection and low recurrence demonstrate the utility of the endoscopic approach for pituitary adenomas. It should be noted, however, that follow-up data was lacking with 75.0% of cases having no follow-up data and average follow-up being only 16.5 months. Therefore, additional reporting of follow-up is necessary in future studies.

Within literature included in this analysis, chordoma and chondrosarcoma were often not distinguished, therefore these two populations were combined in our calculations. These tumors are known to be challenging to resect because of their tendency to involve important structures such as the cavernous sinus and internal carotid artery. ^{29 30} Therefore, it is not a surprise that reported rates of GTR in the literature are relatively low at 4.7 to 58%. ²⁹ Of the 40 cases representing 6.9% of patients, data on resection were available for only 26 patients. GTR was achieved in 34.6% of these patients, the majority having STR at 53.8%. Recurrence was reported in five patients, which is a rate of 12.5% that is lower than reported in the literature of 16 to 68%. ^{29 31} These findings are notable, especially when considering that chordomas are more aggressive in the pediatric population. ⁹ However, as with previously discussed data, follow-up information was limited and only available for 30.0%, which should temper the greatly reduced recurrence rate found in this analysis. Nevertheless, at an average follow-up of 40.2 months, 75.0% of patients within this group were disease-free.

Complications

The most common complications after surgical resection of tumor were endocrinopathies and CSF leaks at 20.7 and 17.3%, respectively. Craniopharyngiomas, pituitary adenomas, and chordomas/chondrosarcomas had the highest number of these complications. Craniopharyngiomas and pituitary adenomas had the highest rate of endocrinopathies, 14.6 and 11.5%, respectively. Among the craniopharyngioma group, diabetes insipidus represented 76.9% of the endocrinopathies. This is consistent with traditional teaching and the tumors' anatomic relationship with structures centrally related to the hypothalamic–pituitary–adrenal axis. ⁹ Among the major pathologies, CSF leak was most common in chordomas, with a rate of 10.0%. This finding also matches with current literature on CSF leaks from ESBS which has found clival chordomas to have an especially high rate of leak. ³²

The finding in our analysis accounts for only four cases within the chordoma/chondrosarcoma group. Also of note is the rate of meningitis of which there was only one reported case, which is less than 0.5% of the population. This is improved from the literature reported rate of 1.8% by Kono et al. ³³ Overall rate of infection was found to be 8.7%, which is also less than previous estimates of ∼12%. ⁹

Limitations

To our understanding, this study constitutes the largest analysis of ESBS in the pediatric population to date. Due to the scant literature on the subject, there are several limitations to this analysis. Previous studies on the safety and efficacy of ESBS in the pediatric population have been limited to case reports, case series, and retrospective cohorts with small sample sizes, which reduce the statistical power of the analysis. Additionally, the literature on ESBS is geared toward the adult population, and if data on children are present, it is often mixed together with the adults. Therefore, comparative analysis is difficult to conduct within the pediatric population and we are forced to compare with the adult population as well. We understand this is a major limitation, but until the body of literature on ESBS in our population of interest is further expanded, this is the best evidence available. Another major limitation is the heterogeneity present within the pathologies treated with ESBS in the pediatric population. It is difficult to make generalized statements for a technique when it is used for such a variety of diseases with varying anatomy and risks for complication. However, to get a better understanding of the utility of ESBS in the pediatric population, multiple pathologies had to be included. We have made an effort to limit overreaching statements and have provided proper justification where necessary.

Conclusion

ESBS provides a minimally invasive means to treat both anatomically and pathologically challenging tumors. This analysis demonstrates comparable outcomes in pediatric patients undergoing endoscopic resection of skull base tumors to historical figures noted for open procedures. Although this study provides an initial analysis on safety and outcomes in the pediatric population, future prospective trials with adequate sample sizes are necessary to delineate the optimal use of the different modalities for each specific skull base pathology.