Diffuse meningeal oligodendrogliomatosis characterized by spinal intra-parenchymal nodules on magnetic resonance imaging in a dog

Abstract

Meningeal oligodendrogliomatosis is a relatively rare neoplasm in dogs. Ante-mortem diagnosis is difficult due to nonspecific neurologic signs overlapping other conditions. The only reported consistent feature is a high level of protein in the cerebrospinal fluid. Veterinary literature offers only 1 case report with magnetic resonance imaging (MRI) of canine spinal meningeal oligodendrogliomatosis in a single dog. In contrast to the predominant diffuse meningeal enhancement shown in that report, we present the case of a young female cane corso dog with marked nodular invasion of the spinal cord on MRI, confirmed by histopathology to be consistent with diffuse meningeal oligodendrogliomatosis.

Key clinical message:

Meningeal oligodendrogliomatosis should be a differential diagnosis when marked nodular invasion of the spinal cord is seen on MRI, both with and without meningeal enhancement.

Case description

An 18-month-old intact female cane corso dog, weighing 39 kg and adopted in Portugal at 4 mo of age, was presented to the Neurology Service of the Centre Vétérinaire DMV for locomotor disorders that had developed over 2 wk. She had previously been seen by her primary care veterinarian who noted right pelvic limb lameness, weakness of the limb, and development of blisters on the dorsal surface of the foot. Similar clinical signs occurred on the contralateral limb 3 d later. By day 6, she was non-ambulatory on her pelvic limbs. No signs of pain were present and anti-inflammatory [Dexamethasone 0.1 mg/kg body weight (BW), q24h] and anti-pain (Gabapentin 7.69 mg/kg BW, q12h) medications were initiated by the attending veterinarian. Mild possible improvement was noted by the owner after initiation of treatment.

At presentation, the dog was reportedly dysorexic and had lost 3.6 kg over the last week. Her physical examination was unremarkable. On neurologic examination, the dog was nonambulatory on her pelvic limbs with absence of voluntary movements. Other neurologic findings included absence of proprioception bilaterally in the pelvic limbs, withdrawal reflex decreased on the left pelvic limb and absent on the right, absence of patellar reflex bilaterally, nociception intact on the left pelvic limb but absent on the right, lumbar amyotrophy, and absence of spinal pain. Urinary incontinence was present. Owners were shown how to manually empty the urinary bladder to reduce the dog’s stress and discomfort. Fecal incontinence was not reported. Neurologic evaluation of thoracic limbs, cranial nerves, and mental status were within normal limits. An L4–S3 spinal cord lesion, worse on the right, was suspected. Differential diagnoses at the time included neoplasia, non-infectious inflammatory diseases (such as granulomatous meningomyelitis), infectious inflammatory disease, or intervertebral disc herniation.

Complete blood (cell) count (CBC) and biochemistry panels did not reveal any abnormalities. Thoracic radiography and abdominal ultrasound were declined by the owner. The owner accepted referral to the Centre Hospitalier Universitaire Vétérinaire of the Université de Montréal for thoracolumbar magnetic resonance imaging (MRI). Images of the thoracic and lumbar regions were acquired with a 1.5 Tesla magnet (GE Signa EchoSpeed HDx; GE Healthcare, Chicago, Illinois, USA) using a spine coil. The following sequences were acquired: sagittal, dorsal, and transverse T2-weighted (T2w) fast spin echo (FSE), sagittal T2w single-shot fast spin echo (T2-SSFSE or HASTE), sagittal and transverse fat-saturated T1-weighted (FS-T1w) FSE, T2w gradient recalled echo (GRE), and sagittal and transverse FS-T1w following IV injection of gadobenate dimeglumine (Multihance 529 mg/mL ®; Bracco Imaging Canada, Anjou, Quebec), 100 mg/kg BW twice or 0.1 mmol/kg BW twice.

Sagittal T2-SSFSE images showed marked attenuation of both dorsal and ventral cerebrospinal fluid (CSF) signal throughout the thoracic (Figure 1A) and lumbar regions up to L5–L6 (Figure 1B). Normal thoracic and lumbar CSF signal on sagittal T2-SSFSE are provided for comparison (Figure 1C, D). On sagittal T2w images, the lumbar and thoracic spinal cord was diffusely heterogeneous with intermittent to complete attenuation of the CSF signal, respectively, as seen on the SSFSE images (Figure 2A, B). Attenuation of the CSF signal was attributed to an increase in the spinal cord to vertebral canal ratio. Indeed, this ratio was most elevated at the level of T4, T9 and L4 consisting of 0.75, 0.91, and 0.81, respectively. Reference mean and standard deviation values of these ratios for each site are reported as: 0.56 ± 0.03, 0.49 ± 0.05 and 0.51 ± 0.08, respectively (1). Several poorly circumscribed hyperintense areas were seen in the cranial to mid thoracic region bordered by a heterogenous parenchyma (Figure 2A). In addition, 3 isointense oval lesions, measuring 12 × 6.5 mm, 15 × 4.7 mm, and 12 × 5.6 mm, were observed in the spinal cord parenchyma at the levels of L4, L1 and T11–T12, respectively (Figure 2A). The 2 most caudal oval lesions were surrounded by a slightly hypointense halo, whereas the T11–T12 lesion was surrounded by a hyperintense halo (Figure 2A). Thoracic and lumbar intervertebral discs were unremarkable, being centrally hyperintense and of normal shape and size.

Figure 1.

Sagittal T2-weighted SSFSE images of the thoracic (A) and lumbar (B) spinal cord in an 18-month-old female intact cane corso dog with diffuse meningeal oligodendrogliomatosis. Representative sagittal T2-weighted SSFSE images of the thoracic (C) and lumbar (D) spinal cord of a normal dog are provided for comparison. The arrows outline the expected (A,B) or visible (C,D) location of the dorsal and ventral cerebrospinal fluid (CSF) signal, from the subarachnoid space. The CSF signal is markedly attenuated throughout the thoracic and lumbar regions, extending to L5–L6 caudally, in the affected dog. In contrast, a continuous double line signal representing the dorsal and ventral CSF signal, is seen in a normal dog. For comparison purposes, the levels (T11, L1, L4) at which the oval lesions are described on T2-weighted FSE and fat saturated T1-weighted FSE images are indicated.

Figure 2.

Sagittal (A and B) and transverse (C) T2-weighted FSE images of the thoracic and lumbar spinal cord in an 18-month-old female intact cane corso dog with diffuse meningeal oligodendrogliomatosis. A — The CSF signal is intermittently attenuated (arrowheads) by the enlarged and heterogeneous lumbar spinal cord. Three isointense oval lesions are present in the spinal cord parenchyma at the level of L4, L1, and T11–T12 (large arrows). Only the T11–T12 lesion demonstrates a hyperintense halo. B — Marked CSF signal attenuation is shown (arrowheads). Several poorly circumscribed hyperintense areas are seen in the thoracic cranial portion (small arrows). C — At T11, the spinal cord is heterogeneous with several hyperintense irregularly shaped nodulelike areas, which are slightly deforming the spinal cord (fine arrows).

On transverse T2w images, the lumbar and thoracic spinal cord was intermittently increased in size, as described, with the spinal cord occupying most of the vertebral canal throughout the middle portion of L3 and over the entire length of L4 extending caudally to the level of mid L5. The thoracic and lumbar spinal cord was diffusely heterogeneous and focally flattened (in the dorsoventral axis) at L1–L2. A mixture of poorly defined hyperintense and hypointense areas was seen throughout the thoracic and lumbar spinal cord, disrupting the normal dorsal and ventral gray matter horns (Figure 2C). Occasionally, distinction between the white and gray matter of the spinal cord was completely lost. Additionally, the spinal cord margins were periodically slightly elevated by the hyperintense areas, as illustrated at the level of T11 (Figure 2C). Cranially to T10–T11, the spinal cord was overall more hyperintense than in the remainder of the thoracic and lumbar regions. The 3 isointense oval structures described on sagittal images were observed on the transverse images. The T11–T12 lesion was more heterogeneous than the other 2 and was surrounded by a thin hyperintense halo, mainly dorsally. An oval and hypointense intramedullary structure was also present at the dorsal aspect of L3–L4 spinal cord, which was not detected on the sagittal images. Vessels and nerves located at the right ventrolateral aspect of the vertebral canal were increased in size compared to the contralateral side.

On sagittal and transverse pre-contrast FS-T1w images, the thoracic spinal cord was mildly heterogenous compared with the moderate heterogeneity of the lumbar region. The lumbar region showed multifocal hyperintense lesions near the central canal and 3 isointense oval lesions in the spinal cord (Figure 3A, B). Following contrast administration, moderate and marked meningeal enhancement was noted in the thoracic and lumbar regions, respectively (Figure 3C, D). The 3 oval lesions, isointense on pre-contrast images (Figure 3B) showed marked homogeneous enhancement and were relatively well-circumscribed on post-contrast images (Figure 3D). Occasionally, the spinal cord demonstrated greater enhancement, as seen at the level of T12–T13 and T13 (Figure 3D). Additionally, the GRE sequence showed focal intraparenchymal hemorrhage extending from L4 to the caudal aspect of L3 and dorsally to the cranial aspect of L3 (Figure 4).

Figure 3.

Sagittal fat saturated T1-weighted FSE images of the thoracic and lumbar spinal cord acquired pre-contrast (A, B) and post-contrast (C, D) administration in an 18-month-old female intact cane corso dog with diffuse meningeal oligodendrogliomatosis. A, B — The 3 spinal cord isointense oval lesions described on the sagittal T2-weighted images are visible (arrows), with the T11–T12 oval lesion surrounded by a subtle hyperintense halo. B — The lumbar spinal cord is heterogeneous with hyperintense thin focal lesions near the central canal (arrowheads). C, D — Meningeal enhancement is mild in the thoracic region and more marked in the lumbar region. The spinal cord oval lesions are relatively well-circumscribed and remarkably enhanced, especially for those located at the levels of L1 and L4. The T11–T12 lesion retains a hyperintense halo as seen in A and a mild degree of enhancement. D — At the level of T12–T13 and T13 (small arrows), the spinal cord enhancement is greater than in other areas. Of note, hyperintense signal from unsaturated adipose tissue is seen in the thoracic epaxial region on pre- and post-contrast images.

Figure 4.

Transverse T2-weighted gradient recalled echo (GRE) image of the lumbar spinal cord in an 18-month-old female intact cane corso dog with diffuse meningeal oligodendrogliomatosis. Irregularly marginated hypointense foci are present in the dorsal aspect of spinal cord at the level of L4, corresponding to intraparenchymal hemorrhage (arrowheads).

Based on MRI findings, a presumptive diagnosis of severe thoracolumbar meningomyelitis with focal intraparenchymal hemorrhage and extensive edema or inflammation was made. Differential diagnoses included granulomatous or infectious meningomyelitis (e.g., parasitic), a metastatic neoplastic process (from an unidentified primary tumor) or a primary neoplastic process (lymphoma or less likely nephroblastoma) of the spinal cord. At this point, a CSF tap would have been indicated following MRI acquisition. The presence and type of inflammatory cells, concentration of proteins, presence of infectious agents, or neoplastic cells may have helped to refine the diagnosis or at least guide towards an inflammatory rather than a neoplastic process or vice versa. Since the patient was admitted through the emergency service and a neurologist was not on site at the time of diagnostic imaging, CSF collection was not performed. Diagnostic imaging specialists do not routinely perform CSF punctures at our institution.

Following image acquisition, MRI findings were immediately communicated to the referring neurologist. Steroidal anti-inflammatory medication at an immunosuppressive dose (Dexamethasone; Vétoquinol, Lavaltrie, Quebec), 0.1 mg/kg BW, q12h was initiated the day of the MRI study and a single dose of chemotherapy was attempted (Lomustine; Mackesson Canada, Quebec), 45 mg PO. No improvement was noted. The dog continued to deteriorate and the owners elected to have the dog euthanized 1 mo later.

At necropsy, the animal was in poor body condition (1/5) with severe atrophy of the lumbar and hind limb muscles. Leptomeninges covering the brainstem were opaque with occasional foci of hemorrhage and fibrin while leptomeninges from the cervical to the sacral region of the spinal cord displayed extensive hemorrhage. Multiple ill-defined brownish areas of softening were present on cross sections of the thoracic and lumbar regions of the spinal cord parenchyma. No mass was palpated. The central canal of the cervical spinal cord was moderately dilated.

The brain and the entire spinal cord as well as samples of the lungs, heart, liver, kidneys, spleen, stomach, small intestine, bone marrow, adrenal glands, pancreas, thyroid glands, and muscles were fixed in 10% buffered formalin and routinely processed for histopathology. Microscopically, the subarachnoid space of the leptomeninges from the brainstem to the sacral spinal cord was severely expanded by a population of neoplastic cells (Figure 5A). Multifocally throughout the spinal cord, the neoplastic cells invaded the white and gray matter (Figure 5A). Leptomeninges covering the brain were also affected, but to a lesser degree. Neoplastic cells formed loose sheets supported by a thin fibrovascular stroma. Cells were small with little cytoplasm and a round to ovoid often hyperchromatic nucleus (Figure 5B). Anisocytosis and anisocaryosis were mild and the mitotic index was low [0 to 3 mitoses per 10 consecutive hpf (400×)]. Most of the neoplastic cells were strongly positive for Olig2 on immunohistochemistry (Animal Health Diagnostic Center, Cornell University, Ithaca, New York, USA) (Figure 5C). Other lesions included a vacuolar hepatopathy (most likely due to the corticosteroid therapy) and a severe atrophy of skeletal myofibers. Considering the absence of a distinct mass, these findings were consistent with a diagnosis of diffuse meningeal oligodendrogliomatosis.

Figure 5.

Leptomeninges covering the spinal cord in an 18-month-old female intact cane corso with diffuse meningeal oligodendrogliomatosis. A — The subarachnoid space was severely expanded by sheets of neoplastic cells that also invaded the spinal cord parenchyma (between arrows) (scale bar = 200 μm. B — Neoplastic cells were small with little cytoplasm and a round to ovoid, often hyperchromatic nucleus (scale bar = 50 μm). C — Immunohistochemical analysis revealed that most cells displayed intense nuclear staining for Olig2 (scale bar = 50 μm). A, B — Hematoxylin-eosin-saffron-phloxine stain.

Discussion

This report presents the case of a young female intact cane corso dog with diffuse meningeal oligodendrogliomatosis for which MRI lesions have not been reported in the literature. To our knowledge, this is the first canine case of diffuse meningeal oligodendrogliomatosis without a primary mass and with MRI evidence of invasion of the spinal cord by multifocal oval structures, with specific imaging characteristics on T2w and FS-T1w. These oval lesions corresponded to clusters of neoplastic cells on histopathology. These clusters shared MRI characteristics similar to those described for gliomas including being hyperintense on T2w, isointense on T1w, and enhancing on post-contrast T1w sequences (2–7). Interestingly, the oval lesion located at T11–T12 had a hyperintense halo on T2w, interpreted posteriori as representing CSF in the central canal surrounding the lesion.

Diffuse meningeal oligodendrogliomatosis is a rare neoplasm in dogs and cats (3,8–10). This low incidence is also found with humans, in which children are more often affected (11–12). To our knowledge, a single case report of a 15-year-old cat with histopathologic confirmation of meningeal oligodendrogliomatosis, but without description of MRI features, has been published (13). This neoplastic disorder is characterized by lesions apparently spreading via CSF in the ventricular system, central canal, and subarachnoid space (2,6,11,15) and rarely occurs by direct invasion of the leptomeninges (15). Diffuse meningeal oligodendrogliomatosis may or may not be associated with a primary intraparenchymal oligodendroglioma (3,4,8,9,14,16). In the present case, no solid masses were found at necropsy in the brain or spinal cord, while multiple areas of softening were observed on transverse section of the formalin-fixed spinal cord. These areas corresponded to intra-parenchymal neoplastic foci and were interpreted as secondary invasion from leptomeninges and not multiple oligodendrogliomas.

The clinical diagnosis of diffuse meningeal oligodendrogliomatosis without a primary mass is difficult since clinical signs overlap with other neurologic disorders and its rare occurrence is why it was not included as a differential diagnosis at the time of clinical presentation. In our case, no CSF analysis was performed. Unfortunately, CSF analysis can be deceiving with this condition since it is often negative for the presence of neoplastic cells (9–12,14). Histopathologic findings from meningeal biopsies may be equivocal (9) and fail to provide histopathologic diagnosis. Definitive diagnosis is currently achieved on postmortem histopathologic examination.

Common MRI features in dogs and humans include marked meningeal enhancement following contrast on T1w sequences (3,4,8,9,11). Among 2 case reports of canine diffuse meningeal oligodendrogliomatosis, only 2 breeds were represented including a Staffordshire terrier (3) and 2 boxers, only 1 of which had an MRI examination (14). In both cases with MRI imaging, studies showed thickened hyperintense meninges on T2w that were hypointense on pre-contrast T1w in the cervical region for the boxer (14) and isointense in the thoraco-lumbar spine for the Staffordshire terrier (3). In both dogs, the meningeal lesions caused compression of the spinal cord, described as severe for the Staffordshire terrier and moderate compression of the cervical spinal cord for the boxer. None of these cases reported spinal cord parenchymal anomalies on MRI or histopathologic examination. To our knowledge, no sudden death associated with this condition has been described in dogs. In contrast, in literature on humans 1 case of sudden death attributed to diffuse leptomeningeal oligodendrogliomatosis has been reported (15). The 37-year-old woman reportedly had very few symptoms, and the diagnosis was achieved postmortem, despite thorough antemortem diagnostics including MRI, computed tomography, and CSF analysis. The autopsy revealed diffuse leptomeningeal oligodendrogliomatosis involving the brain and spinal cord with an intramedullary lumbosacral infiltration. The cause of death was presumably attributed to the leptomeningeal disease causing a slowly progressive rise in intracranial pressure from tumor cells preventing resorption of CSF.

As previously reported in dogs with diffuse meningeal oligodendrogliomatosis, MRI of our patient showed on postcontrast T1w marked enhancement of the meninges and the central canal, probably corresponding to invasion and dissemination of neoplastic cells. Indeed, the dissemination of neoplastic cells by CSF (“CSF drop metastasis”) is known in humans and was described in veterinary medicine in a female English bulldog which had a mass in the left lateral ventricle; a series of 3 MRIs revealed the progression and appearance of metastatic nodules along the leptomeninges and central canal by CSF (2). Based on the microscopic appearance of the spinal cord, the most likely explanation for the nodular structures is the invasion of the leptomeninges in the spinal cord forming nodular-like structures.

Due to its rarity, diffuse meningeal oligodendrogliomatosis may fail to be included in the differential diagnosis of veterinary patients demonstrating diffuse meningeal enhancement on MRI (3,14). Bacterial or cryptococcal meningitis, plasmacytic meningitis, granulomatous meningitis or meningoencephalitis, neoplastic processes (lymphoproliferative tumors and histiocytic sarcoma), and inflammation secondary to internal otitis (4) are more commonly included on the list of differential diagnoses associated with diffuse meningeal enhancement. In human medicine, the differential diagnosis also includes sarcoidosis (15) and subarachnoid hemorrhages (10).

Of note, MRI acquisition and postmortem examination were performed 1 mo apart. It is possible that the lesions found on MRI had progressed over time and involved a greater percentage of the spinal cord at death compared to the imaging study. Nevertheless, the MRI lesions correlated well with those observed on postmortem examination. A few limitations to our case report include the absence of MRI of the cervical spine or brain and the lack of performance of a CSF puncture. First, the neurologic examination findings oriented to a lesion localized in the L4–S3 segments of the spinal cord. Perhaps, visualization of diffuse meningeal enhancement affecting the brain and entire spinal cord would have prompted inclusion of diffuse meningeal oligodendrogliomatosis as a potential differential diagnosis at the time. Finally, in retrospect, it seems unlikely that a cytological and quantitative analysis of CSF, although clinically indicated, would have led to a diagnosis of diffuse meningeal oligodendrogliomatosis based on literature.

In conclusion, we describe for the first time in veterinary medicine, the case of a dog with diffuse meningeal oligodendrogliomatosis characterized by intramedullary spinal cord abnormalities on MRI in addition to the expected meningeal enhancement on T1w post-contrast images. Based on our findings, it is important to include diffuse meningeal oligodendrogliomatosis on the list of differential diagnoses in dogs with diffuse meningeal enhancement with or without abnormalities of the spinal cord, such as nodules, on MRI. Furthermore, in dogs with meningeal enhancement and intramedullary nodules, MRI acquisition of the entire central nervous system (i.e., brain and complete spine) may be warranted to aid the antemortem diagnosis since it is not uncommon for affected dogs to have both brain and spinal cord lesions. Of note, biopsies and CSF cytology are often inconclusive and diffuse meningeal oligodendrogliomatosis may not be excluded based on negative findings.