Spinal meningiomas in dogs: Description of 8 cases including a novel radiological and histopathological presentation

Abstract

Clinical, imaging, and histological features of 8 canine spinal meningiomas, including a cervical cystic meningioma with imaging and intraoperative features of an arachnoid cyst, are described. All meningiomas were histologically classified and graded following the international World Health Organization human classification for tumors. Six meningiomas were located in the cervical spinal cord. Myelography showed intradural/ extramedullary lesions in 3/4 cases. Magnetic resonance imaging revealed hyperintense intradural/extramedullary masses on pre-contrast T1-weighted and T2-weighted images with homogeneous contrast enhancement in 7/8 cases. One dog had a cerebrospinal fluid-filled subarachnoid cavity dorsal to the cervical spinal cord. A spinal arachnoid cyst was diagnosed on imaging, but the histopathological study of the resected tissue revealed a grade I meningothelial cystic meningioma. There were no differences in outcome associated with tumor grade and surgical treatment (6/8). Cystic meningioma should be considered in the differential diagnosis of intraspinal cystic lesions, and biopsy is necessary for definitive diagnosis.

Introduction

Meningiomas are the most common primary spinal cord tumors in dogs (1–4). They are usually benign, slow-growing neoplasms that arise from the arachnoid cap cells and arachnoid granulations of the meninges and cause a chronic progressive myelopathy and spinal pain due to spinal cord compression (5–10). While the thoracic spine is the most common location of spinal meningiomas in humans (80%), canine spinal meningiomas are more frequent in the cervical spinal cord (68%) (3,6,7,10–12). Definitive diagnosis of spinal meningioma requires biopsy and histopathological examination, but a tentative diagnosis is often possible based on tumor characteristics using advanced imaging techniques. Myelography usually demonstrates a discrete, intradural/extramedullary mass (1,2,7,10). The most frequent magnetic resonance (MR) imaging finding is a contrast-enhancing intradural/extramedullary mass with a broad-based dural attachment and variable signal intensity on pre-contrast T1-weighted (T1W) and T2-weighted (T2W) images (4,9,10,13). Cerebrospinal fluid (CSF) analysis results may be normal, or show a mild to moderate increase in protein concentration, with or without pleocytosis (1,7,10). Meningiomas in humans and domestic animals share striking similarities. The World Health Organization (WHO) classification for human central nervous system tumors currently contains 15 distinct meningeal variants, divided into 3 grades (14). Tumor grade has both predictive and prognostic value for human intracranial meningiomas, and it is an important diagnostic criterion (15). The most recent classification of meningiomas in domestic animals describes 9 histological patterns that parallel some of the variants included in the human classification system (16). However, several reports and textbooks describe various histological subtypes of canine meningiomas, including meningeal variants that are not described in the domestic animal classification (5,7,8,12,17,18). Tumor grading following the WHO human classification criteria has only been performed in a few veterinary studies in an attempt to provide a more comprehensive list of histological subtypes of canine and feline meningiomas, and to try to establish a relationship between tumor grade and prognosis (10,11,18,19). However, no apparent correlation between tumor grade and long-term outcome has been found (10,18).

Cysts associated with meningiomas are uncommon in humans. They occur in 1% to 10% of all intracranial meningiomas, and are very rare in the spinal cord (20–24). In the canine species, cystic intracranial meningiomas have been described, but there are no reports of spinal cystic meningiomas (18,25,26).

The aims of this study, therefore, were to i) describe the clinical, imaging, and histological features of 8 canine spinal meningiomas; ii) evaluate the long-term outcome associated with tumor grade and surgical intervention as sole treatment; and iii) describe a novel spinal cystic meningioma presentation with distinguishing imaging features.

Materials and methods

Medical records from dogs with a histopathological diagnosis of spinal cord meningioma seen at the Neurology/Neurosurgery Service of the Veterinary Teaching Hospital of the Autonomous University of Barcelona between 2006 and 2011 were reviewed. Signalment, presenting clinical signs, physical and neurological examinations, clinicopathological data, surgery reports, and available images were reviewed. In addition, results of CSF analysis (when obtained before myelography) were also reviewed.

Imaging of the spine was performed under general anesthesia. Survey radiographs of the vertebral column consisted of lateral and ventrodorsal views. Myelography was performed by intrathecal injection of 0.5 mL/kg body weight (BW) of iohexol (Omnipaque, 300 mg/mL; GE Healthcare, Princeton, New Jersey, USA) at L5–L6. Magnetic resonance imaging examinations were performed with a 0.2 Tesla permanent magnet (Vet-MR; Esaote, Genova, Italy), using the following acquisition parameters: T2W spin echo (SE) sequences, repetition time (TR) 2800 and echo time (TE) 80, and T1W SE sequences, TR 690/TE 26. All available imaging studies were reviewed and interpreted by a board-certified neurologist (SA) together with 2 senior neurology residents (RJL and CDF).

Myelographic and MR images were assessed to determine tumor location longitudinally along the neuraxis and transversely in relation to the dura mater and the spinal cord (extradural, intradural/extramedullary, or intramedullary). Magnetic resonance images were also evaluated for signal intensity relative to spinal cord gray matter, degree and homogeneity of contrast enhancement (CE), presence or absence of adjacent nervous tissue infiltration, peritumoral edema based on T2W images, and presence or absence of a dural tail as defined in a previous study (10). Plain spinal radiographs were reviewed to detect associated bony changes in the adjacent vertebrae.

A diagnosis of meningioma was confirmed after histopathological examination of tissue obtained by surgical excision or at necropsy. Tissue samples were formalin-fixed and processed by routine paraffin embedding. Sections were cut at 5 μm and stained with hematoxylin and eosin (H&E). All meningiomas were histologically examined, classified, and graded by a board-certified pathologist (MP) according to the criteria defined by the international WHO human classification for central nervous system (CNS) tumors (14). The chordoid subtype was included in the grade II group.

All available information about surgical procedures performed was also reviewed. Information regarding outcome was obtained from the records or by telephone conversation with owners and/or referring veterinarians. Survival was reported in days from diagnosis to death or euthanasia. Time to relapse, when applicable, was reported in days from diagnosis to first neurological signs of recurrence noticed by owners or referring veterinarians.

Results

Eight dogs were included in the study, 5 males and 3 females. Median age at presentation was 9.2 y (range: 3.5 to 11.4 y), and median body weight was 35 kg (range: 8 to 45 kg). There were 7 breeds included (Table 1).

Table 1.

Clinicopathological data and outcome for 8 dogs with spinal meningiomas

					Neurologic statusa
Dog	Age (years)	Breed	Tumor location	Histopathologic diagnosis (tumor grade)	Admission/duration of signs (days)	Best status post-treatment (day)	Treatment	Relapse (day)	Survival time (days)
#1	8	CS	T11–T12	Transitional (I)	Paraparesis, mild pain/(45)	Normal (120)	HL	Yes (210)b	230
#2	8.3	GR	C3–C4	Atypical (II)	Right hemiparesis, mild pain/(108)	Normal (45)	HL	Yes (450)b	540
#3	10.3	GSD	C1–C2	Atypical (II)	Right hemiparesis, ataxia/(120)	Mild right hemiparesis (90)	HL	Yes (500)	530
#4	11	GSD	T2–T3	Atypical (II)	Paraparesis, ataxia, moderate pain/(60)	Normal (38)	HDL	Yes (180)	300
#5	11.4	SH	C2	Chordoid (II)	Ataxia, moderate pain/(51)	N/A	Not treated	N/A	Euthanized
#6	3.5	AGD	C5–C6	Meningothelial cystic (I)	Tetraparesis, ataxia, moderate pain/(365)	Moderate ataxia (110)	DL	N/A	110c
#7	8.7	Cross	C1	Transitional (I)	Tetraparesis, severe pain/(290)	Normal (45)	HDL	Yes (365)d	600d
#8	9.8	BS	C2–C3	Microcystic (I)	Non-ambulatory tetraparesis/(210)	N/A	Not treated	N/A	Euthanized

^aAll dogs were ambulatory unless stated otherwise.

^bReoperated without subsequent improvement.

^cLost to follow-up.

^dReoperated with recovery to normal neurological function, alive at time of writing.

CS — cocker spaniel, GR — golden retriever, GSD — German shepherd, SH — Siberian husky, AGD — Argentine Great Dane, Cross — crossbreed, BS — Brittany spaniel. HL — hemilaminectomy, HDL — hemi-dorsal laminectomy, DL — cervical dorsal laminectomy. N/A — not applicable.

The median time between onset of clinical signs and presentation was 114 d (range: 45 d to 1 y). All dogs were admitted for evaluation of chronic progressive ataxia and/or paresis, and neurologic examinations were consistent with a C1–C5 myelopathy in 4 dogs, a C6–T2 myelopathy in 2 dogs, and a T3–L3 myelopathy in 2 dogs. Two dogs showed overt signs of lateralization, such as hemiparesis and propioceptive deficits in the limbs of the affected side. Spinal pain was present in 5 dogs (Table 1), and ranged from mild to severe. Two dogs had received oral corticosteroids, and 2 other dogs had received oral non-steroidal anti-inflammatory drugs (NSAIDs) before admission. Of these, only 1 dog receiving corticosteroids had improved neurologically for 2 mo, but deteriorated afterwards.

Analysis of CSF, collected at L5–L6 before myelography, was performed in 4 dogs. There was increased protein concentration in dogs #2 and #6 (1.34 g/L and 2.61 g/L, respectively; reference value: < 0.45 g/L) with a normal total nucleated cell count (TNCC) (reference range: < 5 cells/μL). The other 2 dogs had unremarkable CSF analysis results.

Spinal radiographs were obtained in 6 dogs and revealed widening of the vertebral canal at the tumor site in dog #1. Myelography (4/8 dogs) demonstrated intradural/extramedullary mass lesions in dogs #1 and #2, and an extradural lesion in dog #4. In dog #6, myelography showed a dilation of the dorsal subarachnoid space over the C5–C6 vertebrae that resembled a spinal arachnoid cyst (Figure 1A). The subarachnoid filling defects corresponded to tumor location along the spinal cord, which was confirmed at surgery or necropsy in all dogs.

Figure 1.

A — Lateral myelographic view of the cervical spine of dog #6. Note the focal accumulation of contrast medium in the dorsal subarachnoid space (arrows), over the C5–C6 vertebrae, resembling a cervical arachnoid cyst. B — Midsagittal T2-weighted (T2W) magnetic resonance (MR) image showing a dorsal enlargement of the subarachnoid space at C5–C6 (arrows), isointense to cerebrospinal fluid, and immediately caudal intramedullary hyperintensity consistent with edema (arrowhead).

All dogs had MR imaging (MRI) of the spinal cord. A complete series consisted of transverse, sagittal, and dorsal T2W images, and transverse, sagittal, and dorsal pre-contrast and post-contrast T1W images. In dog #6, an additional transverse fluid attenuated inversion recovery (FLAIR) sequence was obtained. Slice thickness ranged from 4 to 4.5 mm. Gadolinium dimeglumine (Magnevist; Bayer Schering Pharma AG, Berlin, Germany) was the paramagnetic contrast medium used at a dose of 0.1 mmol/kg BW intravenously.

A focal intradural/extramedullary mass causing severe spinal cord compression was seen on the MR images of 7 dogs. On T2W images, 6 tumors were moderately hyperintense, and 1 was markedly hyperintense to spinal cord gray matter. Mild to moderate peritumoral hyperintensity consistent with edema was seen in the T2W images of 2 dogs (Figure 1B). On pre-contrast T1W images, 7 tumors were mildly to markedly hyperintense (Figure 2A). These tumors were well-delineated and enhanced uniformly after contrast administration (Figure 2B), except in the case of dog #1, which had an intradural/extramedullary lesion that enhanced homogeneously, and also an adjacent intramedullary CE lesion consistent with spinal cord infiltration. A dural tail was seen in 3 dogs and extension of the mass through the right C3–C4 intervertebral foramen was observed in dog #2 (Figure 2B). On the dorsal T1W images of 3 dogs, there was an enlargement of the subarachnoid space adjacent to the mass that had an appearance similar to that of a myelographic “golf tee” sign (Figures 2A, B). In dog #4, despite the extradural appearance of the lesion on myelography, MRI revealed an intradural/extramedullary mass that was confirmed at surgery. In dog 6, MRI showed a CSF-filled arachnoid cavity, hyperintense on T2W images and hypointense on T1W and FLAIR images (Figures 1B and 3). No CE was observed on T1W images after gadolinium administration. These imaging findings were consistent with an arachnoid cyst.

Figure 2.

A — Dorsal MR, pre-contrast T1-weighted (T1W) image of the cervical spine of dog #2. Note a round, well-defined and markedly hyperintense mass on the right side over the C3–C4 vertebrae. The subarachnoid space expands cranially and caudally (arrowheads) to accommodate the mass, giving the image of a “golf tee” sign, similar to that described for intradural/extramedullary myelographic lesions. B — Dorsal post-contrast T1W image showing marked contrast enhancement and extension of the mass through the right C3–C4 intervertebral foramen.

Figure 3.

Transverse MR images at the level of the C5–C6 intervertebral disc space of dog #6 show a cerebrospinal fluid-filled subarachnoid cavity dorsal to the spinal cord, hypointense on T1W (A), hyperintense on T2W image (B) and fluid-attenuated inversion recovery (FLAIR) (C) image (arrows).

Histopathological examination of tumors from surgical biopsies (6 dogs), or necropsy specimens (2 dogs) identified 5 types of meningioma (Table 1). Four tumors were classified as grade I (2 transitional, 1 microcystic, and 1 meningothelial), and 4 tumors as grade II (3 atypical and 1 chordoid). Grade I and grade II meningiomas were equally distributed along the spinal cord, 3 tumors were located in the cervical region and 1 in the thoracic spine within each grade.

Transitional meningiomas (dogs #1 and #7) had a mixture of fusiform cells arranged in intermingled bundles and concentric whorls of epithelioid cells. The meningioma found at necropsy in dog #8 was predominantly composed of tumor cells with clear and vacuolated cytoplasm, and was therefore classified as microcystic. In dog #6, the histopathological study of surgically obtained biopsy specimens revealed a cyst wall formed by a thick layer of collagen fibers and an internal lining of leptomeningeal cell proliferation (Figure 4A). This lining was made of a dense and homogeneous cell population that proliferated in a rich vascular stroma. The neoplastic cells were distributed in a meningothelial pattern, and had elongated to oval nuclei, lax chromatin, moderate to abundant cytoplasm, and indistinct cytoplasmic borders (Figure 4B). Some syncytia of concentric leptomeningeal cell proliferations were also found (Figure 4C). Therefore, the definitive diagnosis was meningothelial cystic meningioma (Figures 4A–C).

Figure 4.

A — Internal lining of the cyst wall of dog #6. Note the high cellularity. H&E stain. Bar = 200 μm. B — Higher magnification of the area corresponding to (*) on (A). Note the leptomeningeal cells proliferating in the wall of the cyst in a meningothelial pattern and their elongated to ovoid nuclei, lax chromatin, and indistinct cytoplasmic borders. H&E stain. Bar = 10 μm. C — Higher magnification of the area corresponding to (&) on (A) showing a focus of whorl-like leptomeningeal cell proliferation. D — Chordoid, atypical (grade II) meningioma formed by chords of epithelial-like cells in a basophilic matrix, and showing marked cellular pleomorphism, overt anisokaryosis, prominent nucleoli and some multinucleated cells mixed with pyknotic cells. H&E stain. Bar = 25 μm.

The meningioma found at necropsy in dog #5 was classified as chordoid because it consisted of chords or nests of epithelial-like cells (Figure 4D) and infiltrated the spinal cord in some areas. The remaining 3 tumors (dogs #2, #3, and #4) were diagnosed as atypical meningiomas. Histological features of these were increased cellularity, multiple and large primary necrotic foci, cellular pleomorphism, overt anisokaryosis, and a mitotic index that ranged from 8 to 15 mitoses/10 high-power fields.

Six dogs underwent cytoreductive surgical procedures within a week after diagnosis (Table 1). The surgical procedures included: thoracolumbar or cervical hemilaminectomy (3 dogs), cervical or thoracic hemi-dorsal laminectomy (2 dogs), and cervical dorsal laminectomy (1 dog), and they were always performed by the same neurosurgeon (SA). Durotomy was performed in all surgeries. In addition, partial durectomies were performed because of presumed tumor growth from the inner aspect of the dura in 5 cases and in order to remove visible subarachnoid adhesions in the cystic case. A dorsal rhizotomy was performed in dog #2 because of the intimate association between the mass and the right C3 dorsal nerve root (Figure 2B). Two dogs were euthanized immediately after diagnosis. Surgical and necropsy findings confirmed the presence of intradural/ extramedullary lesions in all dogs.

After surgical resection, there was marked to complete improvement of neurological signs in all dogs (Table 1). In the grade I tumor group, 2 dogs with transitional meningiomas and 1 dog with meningothelial cystic meningioma underwent surgery. Of the 2 dogs with transitional meningiomas, dog #7 is still alive 20 mo after the first surgery: this dog had a relapse of clinical signs 1 year after surgery and tumor regrowth was apparent on MRI. A second surgery was performed and the dog is clinically normal at the time of writing. For the other transitional meningioma (dog #1), only partial resection of the mass was possible, as it appeared tightly attached to the spinal cord. This dog survived for 7.6 mo until clinical signs relapsed due to tumor regrowth, which was confirmed on MRI. A second cytoreductive surgery was attempted but the dog was euthanized a few days later due to persistent paraplegia and absent nociception in the pelvic limbs. The dog with a meningothelial cystic meningioma was moderately ataxic 4 mo after surgery, at which time it was lost to follow-up.

Three dogs with atypical grade II meningiomas were surgically treated. All dogs suffered relapses of neurological signs at some point after the initial surgery (Table 1). Dog #4 survived for 10 mo and died due to unrelated reasons, and dog #3 was euthanized 17.7 mo after surgery because of neurological deterioration. Dog #2 survived 18 mo, was reoperated for tumor recurrence confirmed by MRI, and euthanized a few weeks later due to lack of improvement.

Discussion

Primary spinal cord neoplasia is uncommon in dogs (3). Interestingly, meningiomas may account for up to 65% of canine primary spinal cord tumors (10). Although meningiomas have been widely described in veterinary medicine, the cases reported in the present study introduce a novel radiological and histopathological presentation of a spinal cystic meningioma and also provide additional information about outcome in graded tumors undergoing surgical treatment (1–4,7–13,17).

Signalment, clinical signs, CSF analysis, spinal radiographs and myelography results of the dogs in this study were similar to those from previous reports (1,2,7–10,17). Myelography accurately identified tumor location along the spinal cord, and demonstrated intradural/extramedullary lesions in 3/4 cases.

Magnetic resonance images demonstrated intradural/ extramedullary masses in 7 cases, including 1 that appeared to be extradural on myelography. Magnetic resonance imaging was more accurate than myelography to define the location of the tumors relative to the dura mater. T2W images revealed hyperintense lesions in all cases, as commonly reported (9,10,13). All solid mass lesions showed some degree of homogeneous hyperintensity on pre-contrast T1W images, as described in 2 previous studies, and homogeneous CE (10,13). Pre-contrast hyperintensity on T1W images may be related to differences in MRI protocol, as suspected in 1 of the previously published studies (10), or it may be related to tumor calcification, although no significant areas of calcification were observed in the histopathology of the present cases (27). The T1W values of different tissues increase at different rates as the magnetic field is increased, so inherent T1W signal may differ depending on magnetic field strength (4). However, in previous studies using high field, or both high and low field magnets, pre-contrast T1W MR signal intensity of meningiomas showed high variability independently of magnet field strength (4,9,10,13). Tumor histological subtype does not seem to be related to pre-contrast T1W intensity either, as the same histological subtype and grade may show different T1W signal intensities using the same magnet and MR imaging protocol (9). In the cases reported herein, different tumor subtypes and grades showed the same intensity pattern. Therefore, the reasons for the variable signal intensity of meningiomas on pre-contrast T1W MR images, and the reason for the pre-contrast T1W hyperintensity of the meningiomas in this study, remain uncertain.

A “golf tee” sign on T2W images, similar to that found on myelography, has been described to help localize intradural/ extramedullary lesions in previous studies (4,9). Here we describe for the first time the “golf tee” appearance of 3 spinal meningiomas on T1W images before and after contrast administration, probably caused by marked differences in signal between the hyperintense CE masses and the hypointense CSF.

The thoracic spine is the most common location of spinal meningiomas in humans, whereas canine spinal meningiomas are more commonly located in the cervical and lumbar regions (1,3,6–13,17). Reports of canine meningiomas in the thoracic spinal canal are scarce (3,10). In a recent study, grade I meningiomas showed a strong predilection for the cervical spinal cord, whereas grade II tumors were more likely to be found in the thoracolumbar region (10). In our series, although there is a small number of cases, grade I and grade II tumors were equally distributed along the neuraxis, with 3 and 1 meningioma of each grade present in the cervical and thoracic spine, respectively.

Most human spinal meningiomas are grade I, and there is a correlation between higher grade tumors (II and III) and worse prognosis (28,29). In contrast, grade I and grade II spinal meningiomas in dogs occur with similar frequency (10,19) and there is no obvious correlation between outcome and tumor grade (10). Survival times after cytoreductive surgery in this case series were in agreement with those previously reported (1,3,7–10). No apparent differences in outcome were observed related to surgery technique, which was chosen depending on tumor location to optimize exposure and resection of neoplastic tissue. A worse outcome than expected was obtained in a grade I meningioma, but this was likely because of incomplete tumor resection due to its tight attachment to the spinal cord, as histopathological analysis of the mass revealed a benign transitional meningioma. No obvious association between outcome and tumor grade was seen in this study, although the results could be biased because of the small number of cases. Of the few grade III canine spinal meningiomas reported in the literature, only 1 received surgical treatment and had a poor outcome (10,11,19). However, a study involving a larger number of tumors would be necessary in order to establish a relationship between tumor grade and prognosis.

This series included an unusual meningioma that had myelographic and MRI features consistent with those previously described for spinal arachnoid cysts (30). However, histopathological examination of the surgically resected tissue demonstrated a grade I meningothelial cystic meningioma. In humans, cysts within intracranial meningiomas are seldomly seen, and they are even more rare in spinal meningiomas (20–24). Cysts are thought to develop secondary to hemorrhage, ischemic necrosis, degenerative changes, and/or fluid production by the tumor itself (20,21,23,24). On MRI, cysts associated with meningiomas are usually hypointense on T1W and hyperintense on T2W images. The degree of cyst wall CE is variable, and lack of enhancement does not indicate absence of tumor cells within the wall (20,21,23). Noticeably, the meningothelial subtype is the most frequent histological variant of cystic meningioma in humans (21,22). Histologically confirmed intracranial meningiomas and ssociated cysts have been described in dogs, but, to the authors’ knowledge, the case reported herein represents the first description of a spinal cystic meningioma in a dog (18,25,26).

In conclusion, most spinal meningiomas in this study were hyperintense on T1W MRimages, and some tumors displayed a “golf tee” sign. There were no apparent differences in outcome associated with tumor grade after surgical resection; however, further studies are needed to correlate tumor histology with biological tumor behavior and long-term prognosis. Finally, cystic meningioma should be considered in the differential diagnosis of spinal cord cystic lesions, and histopathological evaluation is essential for a definitive diagnosis.