Aggressive biological behavior of benign central nervous system neoplasms: The case of hemorrhagic subependymomas – An illustrated review

Abstract

Background:

Subependymomas (SEs) are low-frequency, histologically benign tumors. Their association with hemorrhagic phenomena is extremely unusual.

Methods:

This is a literature review of pathologically verified reports of hemorrhagic SE (hSE), analyzing their epidemiological, clinical, diagnostic, therapeutic, and outcome features. This review includes an original illustrating report.

Results:

A total of 30 cases were analyzed. All but three patients presented unique hSEs growing predominantly within the lateral ventricles (86.6%). The pattern of associated hemorrhage was, in decreasing order of frequency, intratumoral (55.2%), mixed, and extratumoral. Symptoms at presentation, mainly of rapid onset (75%), derived from high intracranial pressure, meningeal irritation, and/or focal neurological impairment. Superficial siderosis was an exceptional complication, presumably produced after repeated bleeding. Most hSEs were misdiagnosed either as a more aggressive neoplasm (84.6%) or as a vascular malformation (7.7%) because of their invariably atypical, heterogeneous radiological appearance. Twenty-nine patients underwent surgical treatment, reaching total resection rates close to 80%. Nevertheless, high figures of both permanent disability (25%) and mortality (10.3%) were recorded at the last follow-up.

Conclusion:

Bleeding increases both the severity of initial clinical symptoms and the complexity of the radiological differential diagnosis of SEs but does not seem to influence strongly the degree of surgical resection of these otherwise histologically benign neoplasms. Both the exact topography of the hSE within the ventricular system and the presence of intraparenchymal extension of the neoplasm and/or the associated bleeding could negatively affect the final clinical outcome.

INTRODUCTION

Subependymomas (SEs) are World Health Organization (WHO) grade 1 gliomas that usually grow near the ependymal lining of the ventricular system or spinal cord.[30] These are quite infrequent lesions since they represent up to 8% of ependymal tumors and <1% of intracranial neoplasms.[30] SEs generally follow a benign clinical course, only exceptionally demonstrating an aggressive biological behavior in terms of rapid growth and/or invasion of neighboring healthy tissue.[30,32] In this respect, association of SEs with hemorrhagic events (referred to hereafter as hemorrhagic SEs [hSEs]) represents a remarkably rare event, although it may lead to a life-threatening clinical scenario requiring rapid and accurate decision-making.[5]

The present study, intended to address the relationship of SEs with hemorrhagic phenomena, reviews the clinical features, diagnosis, and outcome of this infrequent form of presentation in those cases published to date in the scientific literature, including an original, unpublished report of a hSE diagnosed and treated with success at the authors’ institution.

MATERIALS AND METHODS

A systematic search in the PubMed search engine using the terms “subependymoma” and “hemorrhage” yielded 102 articles [Figure 1]. Among them, only those papers containing well-described cases of pathologically verified SEs associating with hemorrhage at presentation were selected for analysis. Additional cases were identified following a “snowballing” methodology, consisting of a meticulous review of the list of references from each selected paper. Finally, an original illustrative case was also included in this review.

Figure 1:

Flow diagram, elaborated following the preferred reporting items for systematic reviews and meta-analyses guidelines (www.prisma-statement.org), summarizing the methodology followed for the review of the scientific literature concerning cases of subependymomas associated with hemorrhage.

The following variables were extracted from each record for statistical analysis with the IBM software suite SPSS Statistics, version 23:

• Epidemiological: Previous diagnosis of SE, age, sex, risk factors for hemorrhage (i.e., arterial hypertension and oral anticoagulation therapy).

• Clinical: Chronology of symptoms at presentation, either acute (<3 days), subacute, or chronic (more than a month); clinical syndrome.

• Diagnostic: Quality of the preoperative imaging study, including computed tomography (CT), magnetic resonance imaging (MRI), and/or angiography findings; pattern of enhancement, consistency (either solid, cystic or mixed) location, and size or major axis of the mass (those tumors from reports lacking this specific information but with evidence of ventricular enlargement and brain distortion in the published images, probably surpassing 2.5–3 cm in their major diameter, were categorized as “large”); signs of parenchymal invasion and edema (either peritumoral or periventricular) and presence of hydrocephalus; type of bleeding (intratumoral, extratumoral or mixed); and suspicion radiological diagnosis.

• Therapeutic: Treatment of hydrocephalus; surgical strategy for tumor removal (technique and approach); removal degree; complications of treatment.

• Outcome: Follow-up duration; identification of either recidive or progression of tumor remnants; effect of treatment on the initial clinical status of the patient (either improvement, worsening or unchanged); and final overall outcome according to the modified Rankin score (mRS).[3]

The availability of each variable is expressed in the Results section as “n;” this value is not specified, provided the specific variable was described in all cases included in this series.

RESULTS

Table 1 summarizes the main features of the 30 hSEs finally available for analysis,[1,2,4-6,9,10,12-15,17-19,21,23,25-27,32-34,36-38,40] which are described in detail in the following lines.

Table 1:

Summary of the main features of the 30 hSEs available for analysis in this review. A detailed description can be found in the text.

Epidemiology

The age of patients ranged from 15 to 81 years old (mean 50.4, median 51); most of them were males (63.3%). Four patients had a previous radiological diagnosis of the intraventricular tumor. Concerning medical records, six patients suffered from arterial hypertension and two of them also received treatment with oral anticoagulants.

Clinical features

In most patients, the clinical course (n = 28) was either acute (42.9%) or subacute (32.1%) and included symptoms derived from intracranial hypertension (36.7%), meningeal irritation (13.3%), focal neurological involvement (13.3%), superficial siderosis (3.3%), seizures (3.4%), or combinations of these syndromes see the table in Figure 2. Alternatively, 10% of patients presented with mild symptoms such as isolated headache or vertigo.

Figure 2:

(a) Schematic representation of a left posterolateral view of the ventricular system, illustrating the distribution of hemorrhagic subependymomas (SEs). Single tumors (90%) predominantly occupied the lateral ventricles (86.6%), with a practically symmetric distribution among the right and left cavities, while only one SE grew in the fourth ventricle; otherwise, three patients (10%) presented with multiple intraventricular masses. (b) Concerning the associated bleeding, intratumoral hemorrhage was the most common pattern observed, followed by mixed both intra and extratumoral bleeding. (c) Three-dimensional reconstruction of the ventricular system and eloquent periventricular structures, including the internal capsule (red), optic pathways (beige), and limbic system (blue), based in the preoperative MRI from the illustrative case report.. Observe the voluminous intraventricular mass occupying the right frontal horn that may be associated with intratumoral (black arrows), extratumoral (brown arrow), or both types of bleeding. The main clinical symptoms observed in the patients included in our series are summarized in the accompanying table (pertinent explanations in the text).

Diagnosis

Initial imaging studies (n = 28) included CT (17.9%), MRI (10.7%), or both techniques (71.4%). Figure 2 summarizes both the topographic features of the tumors and the pattern of associated bleeding. Tumor size (n = 14) ranged from 20 to 70 mm (mean 40.21, median 37.5). We were able to categorize the tumor as “large,” following the aforementioned criteria, in 15 cases. Most hSEs were solid (86.2%, n = 29), although four cases were associated with prominent cystic components. Parenchymal edema (n = 27) was present in 19, with either a peritumoral (37%) or a periventricular (33.3%) pattern. Most hSEs displayed heterogeneous contrast uptake (84%, n = 25), while four tumors did not show enhancement. Seventeen patients presented with radiological signs of hydrocephalus, predominantly obstructive, with only one report of communicating hydrocephalus caused by superficial siderosis. Whenever obtained, angiography (n = 10) invariably showed an avascular mass. Based on their imaging features (n = 13), most hSEs were either misdiagnosed as a higher-grade neoplasm (84.6%) – such as ependymoma, a central neurocytoma, or a metastasis – or confounded with a vascular lesion (7.7%), especially with a cavernoma.

Treatment

Twenty-nine patients underwent a surgical procedure; in most instances, a gross total resection of the tumor (79.3%) was achieved, followed by subtotal and partial removal (10.3% and 3.4%, respectively) of the mass, while biopsy was carried out only in one patient. Concerning the surgical approach, specified in 24 cases, lateral ventricle lesions were resected through either the transcortical (66.7%) or the transcallosal (29.2%) routes, while the sole case occupying the fourth ventricle was operated on through the telovelar approach. Microsurgical technique was preferred (89.3%) over endoscopy, the latter having been employed in only 3 cases of hSE. Parenchymal extension was described in 40.7% of cases during the surgical procedure (n = 27), despite which most authors found the hSEs easily distinguishable and dissectable from normal parenchyma.

Shunting procedures were employed in 20.6% of patients (n = 29), mainly consisting of perioperative insertion of an external ventricular drain (17.2%). Only one patient required permanent shunting.

Pathology

All hSEs corresponded to WHO grade 1 neoplasms, given the absence of histological signs of aggressiveness such as necrosis, prominent mitoses, or microvascular proliferation. Tumor specimens presented microscopic signs of either recent (46.7%), old (3.3%), or both types of hemorrhage (36.7%), while 4 cases did not show signs of bleeding. Prominent arterial vascularization was identified in 17.9% of cases while cavernous areas, resembling cavernous angiomas, were observed in two patients (n = 28).

Outcome

Clinical incidences during patient care included surgical (34.6%), medical (3.8%), or both (3.8%) kinds of complications (n = 26) [Table 1]. The sole nonoperated patient followed a fatal clinical course caused by massive tumor hemorrhage. Regarding the prognosis of operated patients (n = 28), 85.7% improved, 10.7% worsened, and 3.6% remained in a similar clinical status when compared to that before surgery. Follow-up duration after surgery (n = 23) ranged from 21 days to 74 months (mean 17.83 months and median 12 months). There were no reports of tumor recurrence or progression of residual neoplasms. According to the mRS, most patients (64.2%) from the current series experienced a favorable clinical overall outcome, while another quarter remained disabled, mainly to a mild degree and requiring no assistance (17.9%) (n = 28). Finally, this series includes two additional deceases caused by postoperative massive bleeding of tumor remnants and by shunt dysfunction and pneumonia, respectively.

Illustrative case report

A 50-year-old male consulted the emergency department of his referral hospital with a history of progressive headache and spontaneous vomiting during the previous week. His medical records were relevant only for systemic arterial hypertension.

An emergent brain CT showed a ventricular mass [Figure 3a], so the patient was transferred to a tertiary hospital for neurosurgical evaluation. The physical examination revealed mild impairment of consciousness, with a Glasgow Coma Scale score of 14 (E3; V5; M6), and nuchal rigidity. The fundoscopic examination was normal. The brain MRI study confirmed the presence of a ventricular mass associated with signs of both intra and extratumoral hemorrhage [Figures 2c and 3b-e].

Figure 3:

Illustrative report: radiological features of the tumor. (a) Preoperative nonenhanced CT scan showing a heterogeneous mass, displaying intermixed high and low density, which occupies and expands the frontal horn and anterior half of the body of the right lateral ventricle. (b and c) Preoperative axial T1 (b) and T2 (c)-weighted MRI, showing the ovoid, polylobulated intraventricular mass (49 × 32 × 41 mm), displaying heterogeneous signal including hyper and isointense areas in both sequences, suggesting associated subacute hemorrhage within the lesion. Neither signs of associated parenchymal edema nor of active hydrocephalus can be identified. Observe the moderate expansion of the right ventricle, causing contralateral displacement of the septum pellucidum (white arrow in c). (d) Preoperative coronal slice of a gadolinium-enhanced T1-weighted sequence showing no significant contrast uptake by the lesion, which extends into the third ventricle through the foramen of Monro (white arrow). (e) Preoperative axial, T2 gradient-echo sequence showing hypointense nodules in the occipital horn (white arrows), corresponding to intraventricular hemorrhage. (f) Postoperative axial T1-weighted image, confirming that a complete resection of the tumor was achieved.

Surgical treatment, scheduled a week later, included resection of the neoplasm [Figures 3f and 4] and insertion of an external ventricular drainage at the end of the procedure. Both the microscopic and the immunostaining profile of the tumor were consistent with the diagnosis of SE-associated signs of tumoral bleeding [Figure 5].

Figure 4:

Illustrative report: sequential photographic details of the surgical procedure. (a) An interhemispheric approach was carried out through a right fronto-parietal parasagittal craniotomy. The right pericallosal artery (transparent arrow) was displaced toward the left and a right far lateral, 1.5 cm callosotomy was performed, followed by partial evacuation of ventricular, xanthochromic cerebrospinal fluid. (b) Upon accessing the right lateral ventricle, a yellow-grayish friable, well-circumscribed mass was found. Prominent vessels (asterisk) arising from the lateral and medial ependymal walls could be identified on its surface; they were coagulated and sectioned. (c) The neoplasm was intimately adherent to the anterior portion of the septum pellucidum (transparent arrow), suggesting a possible origin in this structure. (d) The remaining surface of the mass was bluntly dissected from the ventricular walls, reducing the component herniated into the third ventricle through the right foramen of Monro (transparent arrow). Prominent tumoral adhesions were also identified in the anterior course of the thalamostriate vein (asterisk), requiring sharp dissection. (e) An en bloc tumor removal was finally achieved. (f) Surgical field after resection of the tumor, showing a permeable foramen of Monro (asterisk).

Figure 5:

Illustrative report: Pathological features of the tumor. (a) Low cellular density neoplasm, with abundant fibrillary material (white arrow); monomorphic rounded nuclei are arranged in chords and nests (black arrows), intermixed with numerous microcystic areas (transparent arrows) with basophile material of myxoid appearance. Neither mitosis nor necrosis were observed. (H-E, 10×). (b) Intratumoral vessels (white arrow) presented hyalinized walls and a perivascular, mononuclear inflammatory infiltrate (white arrow) (H-E, 4×). (c) Associated areas of acute and subacute bleeding (white arrows) (H-E, 20×). (d) Foci of intratumoral hemorrhage (white arrows) (H-E, 10×). (e) Low cellular proliferation index (2%) (Ki 67, 10×). (f). Dense, diffuse cellular GFAP immunostaining (10×); the tumoral cells demonstrated negative immunostaining to synaptophysin, Neu-N, neurofilament and EMA (not shown). H-E: Hematoxylin and eosin staining, EMA: Epithelial membrane antigen

The postoperative period was uneventful except for a hypertensive crisis, managed medically with success. Removal of the external drain was accomplished 72 h after surgery, with the only incidence of a spontaneously resolving pseudomeningocele. The patient experienced a progressive recovery from his initial clinical state, which allowed his discharge from the hospital 6 days after surgery. The postoperative MRI study confirmed the absence of residual neoplasm [Figure 3f]. At the last follow-up visit, 1 year later, the patient remained asymptomatic so he had resumed his previous usual activities, and there were no signs of tumor recurrence in the MRI obtained at that time.

DISCUSSION

Background

After their designation as a distinct pathological entity by Scheinker in 1945, SEs were largely considered indolent tumors that mainly constituted an incidental finding during autopsy.[31] These neoplasms were indeed relegated mainly as a source of interest for the neuropathologists, who focused on speculation as to its cell of origin until Scheithuauer et al. published the first large series of symptomatic SEs in 1978, establishing the clinical relevance of these tumors and documenting the high operative mortality rate associated with their excision.[32]

Nowadays, SEs are cataloged as ependymal, slowly, or null-growing benign neoplasms, which develop with preference within preexisting anatomical cavities, including the fourth ventricle (50–60%), the lateral ventricles (30–40%) and, less commonly, the third ventricle, septum pellucidum, and spinal cord.[28,30] Because of such biological and topographical features, patients frequently remain asymptomatic, rendering the true incidence of SEs difficult to determine. In a revision of 1,000 serial routine necropsies and 1,000 serial surgical specimens of intracranial neoplasms, 0.4% corresponded to asymptomatic SEs and 0.7% to symptomatic ones, with a statistical estimation in the general population at 0.01–0.79% and 0.19–1.21%, respectively, being more frequent in middle-aged and elderly men.[22,28,30,31] Symptomatic presentations, which occur in approximately one third of SEs, are usually secondary to either obstruction of cerebrospinal fluid (CSF) pathways or mass effect due to slow but progressive growth after a variable quiescent period.[20,29,30] Although extremely rare, with only 30 cases reported in the scientific literature to date according to the current review, a hemorrhage can complicate the natural history of SEs, turning these histologically benign neoplasms ostensibly symptomatic [Table 1].

Pattern of hemorrhage

SEs may associate either pure intratumoral, pure extratumoral, or a mixed pattern of hemorrhage. The pathogenic mechanisms that may lead to any of these forms of presentation remain unclear, although several hypotheses have been proposed.[39] As a premise, it is worth noting that the high tumor volume at the time of diagnosis – when compared to their nonhemorrhagic counterparts – seems to represent the most important risk factor for all three patterns of hemorrhagic presentation. That said, intratumoral hemorrhage could be related to an abnormal, neoplastic vascularization, a concept that may include increased vascular fragility, the presence of endothelial proliferation with subsequent inadequate blood supply, and/or the degeneration of the vessels within the tumor.[39] In the present series, although a prominent arterial network could be identified in almost 20% of SEs, the vast majority were found to be practically avascular. Otherwise, two specimens displayed cavernomatous vessels, potentially prone to repeated bleeding as occurs in the case of cavernous angiomas.[2,21] Another hypothetical mechanism entails the invasion of normal surrounding vessels by the neoplasm,[39] which could be the case of those hSEs that displayed a seemingly more aggressive growth with periventricular tissue extension [Table 1].

Otherwise, pure extratumoral bleedings could be considered the consequence of prominent tumor growth, as observed in the series herein analyzed, leading to compression and/or tearing of otherwise normal ependymal or subependymal veins.[33] Alternatively, this type of bleeding might also be considered an independent event, unrelated to the coexistent ventricular SE.

Finally, other variables not derived from the microscopic features of the tumor itself but from systemic morbidities such as abnormal coagulation, hypertension, or the coexistence of the neoplasm with other external insults (radiation therapy, head injury) may also participate in the pathogenesis of tumoral hemorrhage.[5,39] Nevertheless, such kind of risk factors, quite infrequent in the current series, seem to have played a facilitating role instead of being sufficient causes in the special case of hSEs [Table 1].

Clinical presentation

The pathophysiology of hSEs was highly dependent on both the pattern of bleeding and tumor volume at diagnosis [Figure 2c]. Moreover, the amount and rapidity of tumor volume increase, because of unique bleeding or repeated hemorrhages, also seemed to influence the velocity of symptom onset. Altered consciousness was invariably described in those patients exhibiting acute hydrocephalus, a condition caused either by acute/subacute intratumoral hemorrhage provoking a sudden increase in lateral ventricle hSEs volume so obstructing the foramen of Monro or by a ventricular hemorrhage, whether from a pure extratumoral or a mixed-pattern bleeding. Whenever a subarachnoid hemorrhage was present, meningism predominated in the clinical picture.

Focal neurological signs were varied [Table 1 and Figure 2c] depending on tumor location and were caused by direct compression/invasion of nearby parenchymal structures, by vasogenic peritumoral edema, or even by direct tissue destruction caused by the bleeding.[5] Finally, those hSEs associated with low volume repeated bleedings had a more protracted presentation with chronic headache, progressive-onset focal neurological symptoms, adult chronic hydrocephalus-like symptoms, or superficial siderosis, a very uncommon syndrome caused by recurrent hemorrhages into the ventricular cavities and/or the subarachnoid space.[11] Interrupting the progression of such an irreversible and disabling condition can only be guaranteed by precocious identification and control of the source of bleeding.[11]

Radiological diagnosis

Even though numerous studies discussing the CT and MRI characteristics of SEs have been published in the literature, an accurate diagnosis is often not possible and can pose a challenge even to the skilled neuroradiologist in the special case of hSEs.[20,32,34,35] In contrast with hSEs, confined to the intracranial compartment and predominantly growing in the lateral ventricles [Figure 2], typical SEs most frequently develop in the fourth ventricle, followed by the lateral ventricles, and distantly by the third ventricle, spine, and other more infrequent locations.[35] Neuroimaging has been reported to consistently show SEs as well as demarcated, nodular tumors, with a major diameter of 1–2 cm.[20,32,34,35] On CT, they usually appear as well-defined, hypo- or isodense masses that frequently show small internal cystic components and display little to no contrast enhancement.[20,35] MRI characteristically shows a hypo- to isointense lesion on T1WI and hyperintense on T2WI with scarce or no contrast uptake.[34-36] Otherwise, the presence of hemorrhagic changes in association with the SE, besides the higher tumor size at diagnosis and the heterogeneous density/signal invariably reported on both CT and MRI in the current series, definitely contributes to increase the complexity of the differential diagnosis. Hence, as parenchymal edema, prominent cystic changes, heterogeneous contrast enhancement, and/or multiple lesions may also be observed, the hSEs herein analyzed were often misdiagnosed as either more aggressive neoplasms or vascular lesions.

Treatment

Gross total removal represents the optimal treatment for hSEs. This assertion is supported by the fact that two of the fatalities recorded in the series herein analyzed were the consequence of either a lack of surgical treatment or postoperative massive bleeding of tumor remnants, respectively.

A detailed discussion of the surgical approaches to the lateral ventricles is out of the scope of this work.[8] As a general rule, the operative route was tailored depending on the location of the mass, as can be appreciated in Table 1. Specifically, the far lateral transcallosal approach, illustrated in Figure 4, represents a safe surgical corridor that allows the complete removal of those SEs involving the frontal horns.[16]

Regarding surgery for hSEs, some specifications must be highlighted. While non-hSEs may only show a narrow pedicle of attachment to their site of origin,[30,35] their hemorrhagic counterparts are largely more voluminous and may present wider areas of adhesion to the ventricular walls and/or extend into the periventricular parenchyma [Table 1]. Moreover, when operating on a hemorrhagic ventricular mass, and especially on those associated with either extratumoral or mixed bleedings, it is advisable to keep in mind that the presence of hemorrhagic CSF and/or hemosiderin deposits may preclude the identification of the anatomical references of the ventricles, although said deposits may facilitate a safer dissection plane.[11] Nevertheless, these differential characteristics did not seem to have a strong influence on the degree of removal in the series herein studied, as gross total resection was achieved in nearly 80% of patients [Table 1].

Concerning the management of hSEs causing acute hydrocephalus, the results of this review suggest that both prompt tumor removal and the temporal use of an external ventricular drain are effective alternatives to resolve this issue in a life-threatening clinical scenario.[5] Both attitudes are also associated with low rates of patients requiring permanent shunting [Table 1].

Prognosis

As the WHO grade 1 neoplasms, non-hSEs are usually associated with a favorable prognosis.[30] Nevertheless, the series herein studied does not completely fit this premise [Table 1]. The sole patient who was not operated on finally died, and another who underwent a subtotal removal of the neoplasm also followed a fatal clinical course as a direct consequence of a massive tumoral hemorrhage. Postoperative transitory complications included cognitive impairment and pseudomeningocele. Among permanent sequels, altered consciousness and focal neurological symptoms (hemianopsia, memory impairment), caused by either tumor bleeding, tumor extension into brain parenchyma, or surgery, were recorded. Considering the impact of the neurological sequelae in the mRS, up to 25% of patients in the series herein analyzed presented some degree of disability at the last follow-up visit [Table 1], reflecting that hemorrhage complicates both the natural history and therapeutic management of these otherwise benign ventricular neoplasms.

CONCLUSION

SEs can exceptionally debut with a hemorrhagic event. Depending on the pattern of bleeding, neurological symptoms may range from chronic headache to a life-threatening condition. The presence of hemorrhage contributes to increase the complexity of the radiological differential diagnosis of SEs, in favor of more aggressive neoplasms or vascular entities. Surgery represents the treatment of choice for hSEs. The hemorrhage does not seem to strongly influence the degree of tumoral removal. However, the final clinical outcome could be negatively influenced both by the exact topography of the neoplasm within the ventricular system, a factor that also determines the choice of surgical approach and the inherent risk of iatrogenic injury, and by the presence of extension of the neoplasm and/or hemorrhage in the surrounding, eloquent brain parenchyma.

Ethical approval:

Institutional review board approval is not required.

Declaration of patient consent:

The authors certify that they have obtained all appropriate patient consent.

Financial support and sponsorship:

Nil.

Conflicts of interest:

There are no conflicts of interest.

Use of artificial intelligence (AI)-assisted technology for manuscript preparation:

The authors confirm that there was no use of artificial intelligence (AI)-assisted technology for assisting in the writing or editing of the manuscript and no images were manipulated using AI.

Disclaimer

The views and opinions expressed in this article are those of the authors and do not necessarily reflect the official policy or position of the Journal or its management. The information contained in this article should not be considered to be medical advice; patients should consult their own physicians for advice as to their specific medical needs.