Cerebellopontine Angle Epidermoids: Comparative Results of Microscopic and Endoscopic Excision Using the Retromastoid Approach”

Abstract

Background  Densely packed neurovascular structures, often times inseparable capsular adhesions and sometimes a multicompartmental tumor extension, make surgical excision of cerebellopontine angle epidermoids (CPEs) a challenging task. A simultaneous or an exclusive endoscopic visualization has added a new dimension to the classical microscopic approaches to these tumors recently.

Methods  Eighty-six patients (age: 31.6 ± 11.7 years, M:F = 1:1) were included. Nineteen patients (22.1%) had a multicompartmental tumor. Tumor extension was classified into five subtypes. Sixty-two patients underwent a pure microscopic approach (72%) out of which 10 patients (16%) underwent an endoscope-assisted surgery (11.6%) and 24 patients (28%) underwent an endoscope-controlled excision. Surgical outcomes were retrospectively analyzed.

Results  Headache (53.4%), hearing loss (46.5%), and trigeminal neuralgia (41.8%) were the leading symptoms. Interestingly, 21% of the patients had at least one preexisting cranial nerve deficit. Endoscopic assistance helped in removing an unseen tumor lobule in 3 of 10 patients (30%). Pure endoscopic approach significantly reduced the hospital stay from 9.2 to 7.3 days ( p  = 0.012), and had a statistically insignificant yet a clearly noticeable lesser incidence of subtotal tumor excision (0 vs. 10%, p  = 0.18) with comparable cranial nerve deficits but with a higher postoperative cerebrospinal fluid (CSF) leak rate (29% vs. 4.8%, p  = 0.004).

Conclusion  Endoscope assistance in CPE surgery is a useful addition to conventional microscopic retromastoid approach. Pure endoscopic excision in CPE is feasible, associated with a lesser duration of hospital stay, better extent of excision in selected cases, and it has a comparable cranial nerve morbidity profile albeit with a higher rate of CSF leak.

Introduction

Intracranial epidermoids are rare, congenital, inclusion tumors of the central nervous system (CNS). They comprise approximately 1.5% of all intracranial tumors. ¹ Cerebellopontine angle (CPA) represents the most frequent site for these tumors, accounting for nearly 7% of all tumors in this location. ² Unlike epidermoids elsewhere in the brain, CPA epidermoids (CPEs), along with suprasellar epidermoid, develop from the embryologically derived totipotent cells. ^{1 2}

Due to a combination of clinically silent insinuation of the cisternal spaces, often encircling different cranial nerves/vessels of the CPA and a low grade inflammatory reaction in the areas adjoining the tumor capsule, CPE often strangulate the cranial nerves and inseparably adhere to the neurovascular structures. Thus, surgical excision of these tumors does entail a certain degree of risks to these structures. ³

The key to surgical success lies in the selection of an appropriate surgical approach and adequate visualization of the tumor capsule. Several approaches have been described in the literature and these largely depend on the tumor extension and the surgeon's preference. ^{4 5} Recently, several publications have reported on the feasibility and efficacy of endoscopic tumor excision, either as a sole source of visualization or as an adjunct to the traditional microscopic techniques. The issue of the extent of excision is also unresolved in intracranial epidermoid surgery in general. ^{4 5} Despite several recent publications on the role of endoscopic excision of these tumors, comparative studies evaluating the classical microscopic approach and purely endoscopic approaches are lacking.

Retrosigmoid approach represents a simple and common technique employed by neurosurgeons for various lesions of the CPA. Herein, we detail our experience of surgical excision of these tumors using the retrosigmoid approach and the scope of endoscopy in conjunction with conventional microscopic retromastoid approach CPE in the contemporary era.

Methods

Patient Particulars

We performed a retrospective review (January, 2010–June, 2019) of a prospectively maintained database of histopathologically proven, posterior fossa epidermoids epicentered around seventh–eighth nerve complex, with or without an extension into the adjacent compartments. We selected 98 such patients out of which, 86 patients had undergone a retrosigmoid approach for their tumor excision and hence included in this study. The remaining 12 patients operated using a different surgical approach were excluded. Hospital case records, outpatient files, and the hospital information system were the primary sources of data collection. Institutional ethical board approval was obtained to review the medical records and neuroimaging studies of these patients. Patient consent was obtained at the time of surgery for the use of imaging and clinical data for academic purposes.

We divided the patient's symptomatology into: irritative symptoms (facial pain, hemifacial spasm [HFS], tinnitus) and compressive symptoms (hearing loss, ataxia, diplopia, motor weakness, hoarseness of voice, dysphagia). The predominant symptom was utilized to assign a patient to a particular group.

Neuroimaging and Classification

All patients underwent a preoperative cranial magnetic resonance imaging (MRI) including diffusion-weighted sequences. Based on the radiological extension on MRI, the tumors were subtyped from A to E ( Fig. 1 ). The clinical presentation and surgical outcomes were compared among the above stated subtypes to have an idea about the impact of radiological tumor extension.

Fig. 1.

Schematic diagram showing different types of cerebellopontine angle epidermoid. ( A ) Type A are the epidermoids restricted to the cerebellopontine angle without compression of the brainstem or cerebellum. ( B ) Type B epidermoid was in the cerebellopontine angle with brainstem or cerebellar compression, often with extension toward the foramen magnum. ( C ) Type C epidermoid extend to the contralateral side, engulfing the basilar artery along the prepontine cistern. ( D ) Type D epidermoid extended along the Meckel's cave reaching and partially occupying the tentorial incisura in the perimesencephalic region. ( E ) Type E epidermoids extend up to the mesial temporal lobe, having crossed the tentorium cerebelli.

According to the radiological extension, CPE without compression of the brainstem or cerebellum were labeled as type A tumors ( n  = 27, 31.4%); type B tumors occupied the CPA and compressed the brainstem and/or the cerebellar hemisphere in addition, often with extension toward the foramen magnum ( n  = 40, 46.5%). Type C tumors extended to the contralateral side, engulfing the basilar artery in the prepontine cistern ( n  = 4, 4.7%); CPE extending along the Meckel's cave, reaching and partially extending across the tentorial incisura in the perimesencephalic region were labeled as type D tumors ( n  = 10, 11.6%); and finally type E epidermoids were those that extended up to the mesial temporal lobe ( n  = 5, 5.8%), having crossed the tentorial incisure. We considered types A and B as “unicompartmental” ( n  = 67) CPEs while the types C, D, and E were considered together as “multicompartmental” CPEs ( n  = 19).

Surgical Management

All patients were operated using a retrosigmoid approach. It was either a microscopic ( n  = 62, 72%) or a purely endoscopic technique, called the endoscope-controlled tumor excision ( n  = 24, 28%). The approaches were performed under intraoperative facial and lower cranial nerve monitoring in our later cases (since 2017, n  = 35). All microsurgical excisions were performed using the classical techniques, the descriptions of which are beyond the scope of this article.

The endoscope-controlled surgeries were performed in a supine position with the head turned 60 to 70 degrees toward the contralateral side. A cotton pillow placed below the ipsilateral shoulder helped achieve an adequate head rotation without any inadvertent neck vein distortion. We used both 0- and 30-degree endoscopes (4 mm diameter, 18 cm length; Karl Storz, Germany) during tumor excision. We did not use an endoscope holder; instead a “four-hand technique” was utilized where the assistant held the endoscope while the surgeon used both the hands for a bimanual dissection/tumor removal. This allowed the surgeon to use an intermittent dynamic retraction on the cerebellum using the endoscope itself without the use of any brain retractors.

With regards to the capsule separation, meticulous microneurosurgical techniques were employed and all easily separable parts of the capsule were removed. Only when it was deemed that a microdissection of the capsule could jeopardize the neurovascular structure that a decision to leave behind the capsule was taken. In 10 patients in the microscopic excision category (16.1%), an endoscope was used to inspect the surgical cavity and to remove any overlooked residual tumor. Any additional neurovascular compression was addressed at the same time ( Fig. 2 ).

Fig. 2.

( A – F ) Intraoperative photographs showing endoscopic-assisted excision of cerebellopontine angle epidermoid. Through a 2-cm keyhole craniotomy, a multicompartmental epidermoid was excised with preservation of the seventh cranial nerve (straight white arrow shows the seventh cranial nerve; yellow straight arrow shows the sixth cranial nerve; and curve white arrows show the trigeminal nerve).

Assessment of the Extent of Tumor Excision

The extent of tumor excision was determined by performing a MRI after 6 weeks of surgery. If no residual diffusion restricting tumor was found, we labeled it a gross or near total excision (the latter was documented when the operative notes indicated an intentionally left behind tumor capsule, but not seen on MRI). A subtotal excision meant a diffusion restricting tumor residue on the postoperative MRI. In this study, we utilized the maximum tumor dimension on a linear scale as the measure of the tumor size.

Outcome Assessment

The postoperative outcome was assessed at the time of discharge and again at last available follow-up. Follow-up parameters were: a relief in the neuralgic pain, HFS, as well as the development of postoperative deficits (new onset or worsened). Our postoperative assessment was based on a clinical patient examination including tests for hearing and other neurological functions, performed by a team member not involved in surgery. Only on a few occasions, a pure-tone audiometry was performed for confirming the clinical finding of hearing change, and on all occasions, the clinical findings correlated with the hearing thresholds as per the Gardner–Robertson scale. Asymptomatic tumor remnants were followed radiologically. True recurrences or increase in the size of the residual, when symptomatic, were subjected to resurgery. Data was analyzed in Statistical Package for Social Sciences (SPSS) version 23.0 (IBM, Chicago, Illinois, United States) and Medcalc software.

Results

Clinical Presentation

Table 1 shows the comparative symptomatology, radiological characteristics, and surgical outcomes among the five subtypes of CPE. There was an equal number of males and females in the study group and the gender ratio did not differ among the five radiological subtypes. The mean age of these patients was 31.5 years and patients with tumor types C and D were slightly older than the others. Interestingly, the clinical presentation was delayed in type D while a relatively shorter duration of symptoms characterized the type E tumors. The main presenting complaints were: headache ( n  = 46, 53.5%), hearing loss ( n  = 40, 46.5%), and facial pain ( n  = 35, 40.7%). As the Table 1 shows, hearing loss as a clinical presentation was comparatively higher in type B tumors. Type B tumors were also the most common group to have trigeminal neuralgic pain at presentation. Type A CPEs had the maximum propensity to cause headaches. Other less common symptoms and preoperative neurological deficits among the different tumor subtypes are shown in Table 1 . The fact that 12 patients (15%, 6 patients each in type A and B) had preoperative motor fifth nerve palsy and 3 patients (3.5%, 2 patients in type 2) had a preoperative lower cranial nerve palsy, indicates that the mass effect caused by CPE is often underrecognized. Twenty-one percent patients ( n  = 18) had at least one preoperative cranial nerve deficit at the time of clinical presentation. When we analyzed the clinical features as subgroups, we found that 33 patients (38.36%) presented with irritative symptoms while 53 patients had compressive symptoms (61.64%). We found that the irritative symptoms were more pronounced than the compressive symptoms in type A and type E tumors.

Table 1. Demographic profile and clinical features among various radiological subtypes of cerebellopontine angle epidermoid ( n  = 86) .

Clinical parameter		Classification of cerebellopontine angle epidermoid					Total ( n , % of total)
		Unicompartmental		Multicompartmental
		Type A ( n  = 27)	Type B ( n  = 40)	Type C ( n  = 4)	Type D ( n  = 10)	Type E ( n  = 5)	86
Age (y)		29.6 ± 11.6	30.5 ± 10.8	36.7 ± 15.1	34.2 ± 11.6	28.2 ± 9.8	31.6 ± 11.5
Gender (M:F)		13:14	20:20	3:1	4:6	3:2	1:1
Duration of symptoms (mo)		18.7 ± 17.9	19.3 ± 22.4	54.5 ± 46.4	24.3 ± 30.2	8 ± 6.3	24.9 ±  24.64
Leading clinical profiles (in descending order of frequency, expressed as a fraction of total numbers in each group)
Headache		17 (63%)	20 (50%)	2 (50%)	4 (40%)	3(6.5%)	46 (53.4%)
Hearing loss		6 (22.2%)	28 (70.0%)	1 (25%)	5 (50%)	0	40 (46.5%)
Cerebellar signs		6 (22.2%)	25 (62.5%)	−	5 (50%)	−	36 (41.8%)
Facial pain		10 (37%)	19 (47.5%)	1 (25%)	3 (30%)	2(5.7%)	35 (40.7%)
Preoperative facial paresis		3 (11.1%)	7 (17.5%)	3 (75%)	4 (40%)	1(5.5%)	18 (20.9%)
Preoperative sensory trigeminal deficit		3 (11.%)	10 (25%)	−	4 (40%)	−	17 (19.7%)
Vision loss		−	9 (22.5%)	2 (50%)	1 (10%)	−	12 (13.9%)
Trigeminal motor weakness		6 (22.2%)	6 (15%)	−	−	−	12 (13.9%)
Hemiparesis		−	1 (2.5%)	1 (25%)	3 (30%)	−	5 (5.8%)
Hemifacial spasm		−	3 (7.5%)	−	1 (10%)	−	4 (4.7%)
Lower cranial nerve impairment		−	1 (2.5%)	−	2 (20%)	−	3 (3.4%)
Radiology	Hydrocephalus	−	12 (80.0%)	−	3 (30%)	−	15 (17.4%)
	Maximum tumor size	3.4 ± 0.7	4.5 ± 1.2	3.6 ± 1.4	4.1 ± 2.2	5.4 ± 2.9	4.2 ±  1.68
Irritative/compressive symptoms ratio		16/11	9/31	0/4	3/7	5/0	33/53
Extent of excision
Total		11 (40.7%)	21 (52.5%)	3 (75%)	5 (50%)	0	40 (46.5%)
Near-total		16 (59.3%)	18 (45%)	1 (25%)	4 (40%)	1 (20%)	40 (46.5%)
Subtotal		0	1 (2.5%)	0	1 (10%)	4 (80%)	6 (6.9%)
Postoperative changes in irritative symptoms (as percentage of total numbers of complete improvement)
Improvement in facial spasm		−	1	−	−	−	1 (25%)
Complete improvement in facial pain		7 (28%)	15 (60%)	1 (4%)	2 (8%)	−	25 (71.4%)
Major postoperative complications
Postoperative hearing loss		1 (14.3%)	5 (71.4%)	−	1 (14.3%)	−	7 (8.1%)
Seventh nerve palsy		1 (25.0%)	2 (50.0%)	−	1 (25.0%)	−	4 (4.7%)
Facial numbness		1 (33.3%)	1 (33.3%)	−	1 (33.3%)	0	3 (3.4%)
LCN palsy		−	1 (50.0%)	−	1 (50.0%)	−	2 (2.3%)
Death		−	1	−	−	−	1 (1.1%)
Mean follow-up (mo)		48 ± 24.1 (2 to 120)

Abbreviation: LCN, lower cranial nerve.

Surgical Results

Out of 86 primary surgical procedures performed in these patients, the approach was a purely endoscopic one in 24 (28%) patients while the remaining 62 (72%) patients underwent a microscopic retrosigmoid approach. The differences in patient profiles and surgical results between microscopic and pure endoscopic approaches are shown in Table 2 . There were no significant differences between endoscopic and microscopic excisions with regards to the patient age ( p  = 0.77) and radiologic subtypes of tumors, although the purely endoscopic approaches were employed for only infratentorial tumors (type A [ n  = 11], type B [ n  = 11], and one patient each for type C and D tumors).

Table 2. Comparison of endoscopic versus microscopic surgery cohorts in our series.

Parameters	Pure endoscopic approach ( n  = 24)	Microscopic approach ( n  = 62)	p -Value
Hospital stay (d)	7.3 ± 2.0 [range 4–14]	9.2 ± 3.4 [range 3–21]	0.012
Age (y)	31.27 ± 10.9 [range 16–62]	30.5 ± 11.42 [range 15–57]	0.77
Types
Type A	11 (45.8%)	16 (25.8%)	0.07
Type B	11 (45.8%)	29 (46.8%)	0.94
Type C	1 (4.2%)	3 (4.8%)	0.43
Type D	1 (4.2%)	9 (14.5%)	0.27
Type E	0	5 (8.1%)	−
Preoperative symptoms
Hearing impairment	10 (41.7%)	30 (48.4%)	0.58
Facial pain	12 (50%)	23 (37.1%)	0.28
Diplopia	0	6 (9.8%)	−
Hoarseness of voice	1 (4.2%)	2 (3.2%)	0.65
Hemiparesis	0	5 (8%)	−
Facial palsy	1 (4.2%)	17 (27.4%)	0.02
Impaired lower cranial nerves	1 (4.2%)	8 (13%)	0.43
Type of symptoms
Irritative	12 (50%)	21 (33.9%)	0.17
Compressive	12 (50%)	41 (66.1%)	0.17
Excision
Total/Near	24 (100%)	56 (90.3%)	0.18
Partial	0	6 (9.8%)	−
Postoperative complications
Overall permanent complications	4 (16.7%)	8 (12.9%)	0.65
Transient CSF leak	7 (29.1%)	3 (4.8%)	0.004
Transient ataxia	2 (8.3%)	8 (13%)	0.12
New-onset fifth nerve paresis	1 (4.2%)	1 (1.6%)	1.0
Postoperative hearing loss	2 (8.3%)	5 (8%)	0.33
Permanent postoperative seventh nerve palsy	1 (4.2%)	3 (4.8%)	0.44
New-onset LCN palsy	1 (4.2%)	1 (1.6%)	1.0
New-onset motor weakness	0	2 (3.2%)	−
Postoperative death	0	1 (1.6%)	−

Abbreviations: CSF, cerebrospinal fluid; LCN, lower cranial nerve.

Note: Bold faced values indicate statistically significant values.

Barring trigeminal neuralgic pain, all other symptoms and signs were more marked in the microscopic approach group, out of which preoperative facial paresis was significantly more frequent in the microscopic group ( p  = 0.02). Although not statistically significant, the microscopic surgery group had more of compressive symptoms than irritative symptoms (66% vs. 50%, p  = 0.17).

Use of an endoscopic visualization at the terminal part of a pure microscopic approach ( n  = 10) helped identify and remove an additional tumor lobule in 3 patients (30%). In the remaining patients, it only confirmed that the extent of excision as complete. In two patients (type A), an additional trigeminal nerve neurovascular conflict was noted during surgery and addressed simultaneously. Both patients showed an additional trigeminal nerve compression by a displaced superior cerebellar artery (the tumor had pushed artery toward the nerve).

The extent of tumor excision in our series was as follows: gross total excision in 40 (46.5%) patients, near-total excision in 40 (46.5%) patients, and subtotal excision in 6 (7%, type B, n  = 1; type D, n  = 1; type E, n  = 4). Interestingly, all subtotal excisions were in the microscopic group but the number of total/near-total excisions did not differ between the groups, probably because of unequal number of patients in the two groups ( p  = 0.18).

Postoperative complications were noted in 25 patients (29%), out of which 12 patients had permanent postoperative deficit (14%). There was one operation-related mortality (1.1%). This patient had a postoperative cerebrospinal fluid (CSF) leak leading to meningitis and sepsis following a gross total excision of type B tumor using a microscopic retrosigmoid approach. Persistent House and Brackmann (H & B) grade 4 facial palsy was present in four patients (4.7%) all of whom underwent a protective lateral tarsorrhaphy. In addition, two patients (2.3%) had persistent new-onset lower cranial nerve dysfunction, requiring a tracheostomy to protect the lower airway and assisted feeding (new-onset in one, worsened in another). Two patients each had a persistent new-onset hemiparesis (2.3% each) after surgery and persistent new-onset postoperative trigeminal motor weakness. Five patients developed fifth nerve sensory deficit of new onset (5.8%). Postoperative hearing loss was observed in 7 patients (8.1%) and all were persistent. The transient complications included: transient facial palsy in 11 patients ( n  = 8, H & B grade 2; n  = 3, H & B grade 3), transient ataxia ( n  = 10), and CSF leak/wound bulge ( n  = 10). All transient complications recovered at the time of discharge. CSF leak complications were managed by resuturing of the wound ( n  = 8) or by temporary lumbar drainage for 5 days ( n  = 2).

The duration of hospital stay was significantly shorter in the endoscopy group ( p  = 0.012). Overall, permanent neurological complications were seen in 4 patients (16.7%) in the endoscopy group and 8 patients in the microscopy group (12.9%) ( p  = 0.65). Though transient, CSF leak was comparatively higher in the endoscopic excision group (29.1% vs. 4.8%, p  = 0.004). While the rest of the postoperative complications were comparable between the two groups, new-onset postoperative motor weakness and postoperative death were exclusively noted in the microscopic group.

Follow-Up

At a mean follow-up of 48 (range 12–120) months, there were two recurrences (both were increase in the size of residual tumor) while the other 4 patients with residual disease showed a clinico-radiologically stable disease. Both the recurrences were late (5 and 6.5 years after surgery, respectively) and both patients had undergone a subtotal excision using a microscopic retrosigmoid approach. At recurrence, a reexploration of the retromastoid approach and near-total excision was achieved in both patients using a pure endoscopic approach, the histopathology being the same as the primary tumor. Both patients made uneventful recovery without any new-onset neurological deficits. Of 35 patients with trigeminal neuralgia (TN), 25 (71.4%) had complete pain relief, 3 patients had partial pain relief (3.5%), while the pain remained unchanged in the remaining 7 patients (8.1%). Out of the 4 patients with HFS, 1 (25%) had a complete relief while the remaining 3 patients had only a partial relief after surgery.

Discussion

Intracranial epidermoids are uncommon tumors (∼1.5% of all CNS tumors). ^{6 7} Embryologically, these tumors arise from migrated dorsal midline ectodermal cells except in the CPA and the suprasellar region, where a proliferation of the multipotent embryogenic cell rests or lateral displacement of ectodermal cells by the developing otic vesicles lead to these tumors. ⁸ CPA cistern is the most common site (nearly 60–70%) for intracranial epidermoids followed by the fourth ventricle (15–18%) and middle cranial fossa region (15%). ⁸ Other reported sites are optic chiasma and diploe of the skull. ⁹

Despite being a common site, CPEs often entangle multiple cranial nerves, the posterior fossa vasculature, and by dint of their ability to insinuate into various natural crevices, develop multicompartmental extensions, often across the midline and the tentorial hiatus. We saw that one-fifth of our patients had tumor extensions into more than one compartment (22.1%). As Samii et al noted, ¹¹ there are multiple subcompartments even within one CPA, thereby making CPE actually more challenging than their seemingly indolent nature may suggest.

Epidermoids are often slow growing tumors and remain undetected for many years. In this time duration, due to persistent low grade chemical inflammation, the capsule tends to adhere to the neurovascular structures. Therefore, as these tumors extend, their size increases, the distance of the deeper part of the tumor from the surgical entry site increases, more number of cranial nerves get entangled, and the capsule tends to become more adherent to these structures.

The affected patients are often in the fourth decades of life, despite the condition being congenital in nature. A slight male preponderance is the rule with these tumors. ^{8 9 10} Our experience also showed that the affected patients are in their fourth decades of life and there is usually a delay in the clinical presentation (24 months in our series), consistent with indolent growth of these neoplasms. We also noted that the type C tumors, wherein the tumor occupies the prepontine space, presented later than other subtypes. We believe the late presentation could be due to a relatively lesser distortion of the cranial nerves of the CPA unlike types A and B.

Surgical Approaches

Several surgical approaches have been described for CPEs. But by far the most common approach for these tumors is the retrosigmoid suboccipital approach, used exclusively in our series. Yasargil et al reported that large epidermoids generally create sufficient space of their own and thus facilitate tumor resection through a retrosigmoid approach. ⁶ We also noted that it was also possible to mobilize small tumor lobules that extended across the tentorium, as the incisura is generally widened by the tumor growth. The dilemma regarding the surgical approach arises when the tumor has a prepontine component or has contralateral extension or when a significant part extends supratentorially. Several approaches have been described for these specific subsets of epidermoid. ^{6 7 8 9 10} These are all essentially various skull base approaches like the anterior transpetrous approach, orbitozygomatic approach, and combined anterior and posterior transpetrosal and endoscopic endonasal approaches. ^{11 12 13 14} The reason why some authors advocate these approaches is that a standard retrosigmoid approach may not allow sufficient visualization and room to carry out microsurgical dissection of the capsule, particularly if the CPA portion of the tumor is smaller. Four of the six subtotal excisions took place in type E tumors in our series, highlighting this very point. A wider working space and direct visualization of the ventral brainstem has been reported to allow more extensive surgical resection and preservation of the neurovascular structures through this approach. ^{9 10 11 12 13 14} Recently, endonasal endoscopic approaches have been reported in prepontine epidermoid. However, the approach may not be ideal for CPE with prepontine cistern extension. ^{4 15 16} Our belief is that the natural space created by the tumor and endoscope assistance to visualize the blind spots may obviate the need for these traditional skull base approaches in most such cases and the majority of these tumors can be taken out using the retrosigmoid approach. An endoscopic assessment at the end of the procedure appears a must particularly in multicompartmental cases.

Endoscope-Controlled versus Endoscope-Assisted Surgery and Their Impact

Before the introduction of surgical microscopes, the operative mortality of CPE ranged from 20 to 57%. ^{6 9 11} The figures have drastically improved ever since. However, there are a few drawbacks of a pure microscopic approach that include an inability to visualize the nooks and corners leading to residual tumors in some “blind areas,” inability to ensure the completeness of the extirpation of tumor capsule, need for a larger bony opening, and a cerebellar retraction. ^{14 15} All six subtotal excision in our series were operated using a microscopic approach. Moreover, among 10 patients where the microscopic surgery was assisted by an endoscopic inspection, 30% patients were found to have a missed tumor lobule. Therefore, conventional microscopic approach does have a risk of subtotal excision, particularly in multicompartmental tumors. These difficulties were well recognized by previous surgeons which led to the recent popularity of endoscopes in such tumor excisions, often in conjunction with the microscope (endoscope-assisted surgery) or less commonly as a sole method of tumor visualization (endoscope-controlled surgery). The quality of illumination provided by endoscopes is far better than the commonly used dental mirrors by many surgeons. ¹⁷ Endoscopic assistance in a microscopic surgery is perhaps the most common application of endoscopy in CPEs. In a series of 50 patients with different tumors in the CPA (that included 4 patients with CPE), Abolfotoh et al utilized an endoscope assistance in different skull base approaches. ⁵ They noted that an endoscope assistance in surgery for CPA offered three key advantages: (1) extension of the surgical field into additional intracranial compartments; (2) better visualization and resection of residual tumor not adequately visualized with the microscope, that is, “blind areas”; and (3) improving the prediction of the extent of resection, particularly for tumors like CPE where the residual capsule may not always be seen on postoperative imaging. In the report by Schroeder et al, four of the eight patients underwent an endoscope-assisted surgery while the remaining four were endoscope controlled. ¹⁸ Endoscope assistance allowed them to detect a residual tumor lobule in one patient that was not seen in the line of vision of the microscope. In all of their endoscope-controlled cases, angled scopes allowed them to excise all four tumors without any additional need for enlarging the craniotomy. They treated not only the CPEs, but also those extending to the prepontine cistern, Meckel's cave, and the mesiotemporal structures. They did not report any complication that could be attributed to the use of an endoscope. Peng et al reported their experience of endoscope-controlled excision of CPE. ¹⁵ They reported a complete excision of the capsule in five out of their six patients with only one patient developing a permanent cranial nerve dysfunction. Their patients could be discharged within 3 to 5 days of surgery. All their patients had tumor limited to the CPA and all approaches were through the retrosigmoid corridor. We observed that the addition of endoscopic scrutiny of the postoperative cavity led to improvement in the extent of excision in 3 of our 10 patients (10%). We believe this could be a boon in multicompartmental epidermoids and can obviate the need for extensive or multistage approaches. We found that a pure endoscopy-controlled surgery had a shorter hospital stay (7.2 vs. 9.3 days, p  = 0.012) and a statistically insignificant, yet a lesser likelihood of subtotal tumor excision (0% vs. 9.8% in microscopic approach, p  = 0.18). We generally preferred endoscopy-controlled approach in tumors confined to the CPA (types A and B), most of whom had an irritative symptomatology and minimal preoperative deficits. It could have reflected our initial years of experience wherein the lesser difficult cases are taken up for the approach.

There are, however, certain disadvantages of an endoscope in these tumors and include heat damage to the cranial nerves, poor depth perception, difficulty in controlling hemorrhage, sword fighting of instruments, and most importantly, the structures proximal to the scope where there can be injuries from passage of instruments. With increasing experience and the use of four-hand technique, the depth perception no longer remains a major drawback for us, as the dynamic scope movement provides a good depth perception. Nevertheless, transient CSF leak-related complications were higher in endoscopic controlled excision as compared with the microscopic excisions in our series ( p  = 0.004). We have used short incision and keyhole craniotomy in most of our cases and this could have given rise to high CSF leaks in our series.

Surgical Results in CPE: Literature Review

Aboud et al recently reported their experience of aggressive surgical resection of 34 giant intracranial epidermoids, out which 18 patients had CPA. ¹⁹ They noted that both de novo and recurrent tumors derived benefit from an aggressive surgery. They however noted a higher incidence of permanent cranial nerve complications in recurrent tumors and also a shorter time to recurrence in the recurrent tumors in successive recurrences. ¹⁹ They ascribed the same to a thicker and more adherent tumor capsule in the recurrent tumors. Thus, surgical difficulty in recurrent tumors, increased chances of chemical meningitis following subtotal excision, and risks of malignant degeneration of residual tumors were put forward as reasons to be surgically more aggressive in these tumors. ^{20 21 22} They also analyzed the literature and found that differences in the mortality/morbidity profile between series with moderate surgical aggression (percentage of total excision in the series being 28–60%) were not significantly lesser than aggressive surgical resections (total removal > 62%). On the other hand, Schroeder et al as well as many others advocate a conservative approach, in that only the loosely adherent capsule was removed, leaving behind the dense ones. ^{17 18} They noted two recurrences in their series, after 15 and 19 years, and both the recurrences were addressed surgically without any additional surgical difficulty or complications. They stressed on the importance of neurological preservation and not to pursue aggressive surgical removal of the capsule. Our approach of removal of all loosely separable tumor capsules led to a satisfactory resolution of the irritative symptoms in the majority of the patients. Kobata et al reported on TN and HFS secondary to CPE and recommended a thorough decompression of the nerve not only from the tumor but also any vessel in contact with the nerve. ²³ The complication profiles of our patients were comparable with the literature. A relatively higher incidence of postoperative cranial nerve deficits has been reported in CPE surgery. ^{2 6 7 8 9 17 20 23} Interestingly, we observed that a large number of these patients already have cranial nerve deficits at the time of clinical presentation. Contrary to an otherwise favorable profile of endoscopic approaches in these tumors, a higher incidence of a postoperative CSF leak in our series was a concern and needs further investigations.

There were several limitations of our study. Apart from its retrospective nature, the two surgical groups had inequal number of patients for comparison. Moreover, due to a lower number of endoscopy-assisted cases, we did not include these cases separately in the comparison. The pure endoscopic approaches were limited to simpler cases with confined tumors, having mainly irritative symptoms without much preoperative deficits, unlike the microscopic group where the tumors were more extensive with preexisting cranial nerve deficits thereby precluding a head-to-head comparison of the approaches in similar kind of patients. It probably also led to a better or at least a comparative outcomes between the two groups in our study. Moreover, we did not record the operative times in many of our cases, therefore it could not be compared between the two groups. Lastly, a relatively short follow-up is a major limitation in the assessment of recurrences in tumors like intracranial epidermoids. Despite these limitations, we believe, endoscopic surgery for CPE is feasible in selected cases while microscopic surgery, often supplemented by endoscopic insection, still has its role in CPE surgery.

Conclusion

Nearly one-fifth of CPE have multicompartmental extensions. Despite a reported indolent clinical profile, CPE patients often have preoperative cranial nerve dysfunctions due to neurovascular strangulation and chronic inflammatory processes. Retrosigmoid approach provides an optimal access to these tumors but microscopic visualization may lead to missing out on some tumor lobules in multicompartmental extensions. Endoscopy assistance may be useful not only in reducing the chances of “accidental residuals” but also in confirmation of the actual extent of excision. Pure endoscopic excision in CPE is feasible, associated with a lesser duration of hospital stay, a better extent of tumor capsule excision, and has similar cranial nerve morbidity but a higher rate of transient CSF leak. Recurrences are delayed and clinical judgment is necessary in balancing risks and benefits of an aggressive tumor capsule excision.

Funding Statement

Funding None.