Abstract

An 8-year-old intact male Lagotto Romagnolo was presented with forebrain signs. Neuroanatomic localization was diffuse prosencephalon. MRI revealed diffuse, bilateral, and symmetric T2 and FLAIR hyperintensities in the parieto-occipital white matter and corpus callosum. No mass effect or contrast enhancement was noted. Analysis of cerebrospinal fluid revealed normal protein content and mild mononuclear pleocytosis. Atypical cells were not identified. 15 days later because of the worsening of clinical condition the patient was euthanized upon owner’s request. Neuropathological investigations were consistent with gliomatosis cerebri (GC). Such an unusual imaging pattern appeared similar to some cases of human GC and to a previous reported case in a dog, suggesting a possible repeatable imaging findings for this rare brain neoplasm. GC should be included in the MRI differentials for diffuse bilateral white matter signal changes and specific MRI findings described in this report may help in reaching a presumptive diagnosis of this tumor.

1. Case Presentation

An 8-year-old intact male Lagotto Romagnolo was evaluated for a 1-month history of abnormal mental status. Clinical and neurological examination findings included depression, decreased postural reactions in all four limbs, especially on the right side, and decreased menace response in both eyes. During clinical examination the dog had seizures. The medical case was considered to be consistent with a diffuse forebrain lesion. Because of the age of the patient and the absence of hyperthermia neoplasm was considered the most probable cause of neurological signs while inflammation was considered to be less likely. Complete blood count, serum chemistry, and urinalysis results were within normal limits.

MRI of the brain was performed under general anesthesia using a 0.25 Tesla permanent magnet (ESAOTE VET-MR GRANDE, Esaote, Genoa, Italy). MRI sequences protocol included Fast Spin Echo T2-weighted images acquired in sagittal and transverse plane, a FLAIR acquired in the transverse plane, and Spin Echo T1-weighted images acquired in the transverse plane before and after intravenous administration of paramagnetic contrast medium (Magnegita—gadopentatedimeglumine 500 mmol/mL—insight agents; 0.15 mmol/kg BW). MRI showed extensive, bilateral, and symmetric T2 and FLAIR hyperintensity of the parieto-occipital periventricular and subcortical white matter. The occipital lobes appeared severely involved. The lesions were isointense on T1-weighted images and did not enhance after contrast medium administration (Figures 1 and 2). On sagittal T2-weighted images the aboral part of the corpus callosum lacked homogeneous signal intensity and had ill-defined margins (Figure 3). Mass effect or loss of anatomical architecture was not present. Based on the imaging findings toxic-metabolic or degenerative disorders involving white matter (leukoencephalopathy) were considered the main differential diagnoses, while an inflammatory process, either immunomediated or infectious, and a diffuse infiltrative neoplasm were believed less likely.

Analysis of CSF collected from cerebellomedullary cistern revealed normal protein content (20 mg/dL; range, <25 mg/dL) and mild mononuclear pleocytosis (30 cells/μL; range, <5 cells/μL). Atypical cells were not identified. The dog was treated with phenobarbital (5 mg/kg q 12 hr). Pending on the PCR tests on the CSF sample for canine distemper virus, Toxoplasma gondii, Neospora caninum, Ehrlichia canis, and Rickettsia spp., clindamycin (15 mg/kg q 12 hr) was added on. All the required PCR investigations turned out to be negative. After 15 days the dog was reevaluated. Because of the worsening of clinical condition the dog was euthanized upon owner’s request.

Soon after death the dog was submitted to necropsy that did not reveal significant gross visceral lesions. At neuropathological examination performed on formalin-fixed coronal brain sections, cortical white matter of centrum semiovale was markedly expanded to the detriment of cortical gray matter. No more gross brain lesions were observed.

At histologic examination a severe diffuse infiltration of neoplastic cells was observed affecting centrum semiovale, periventricular white matter, corpus callosum, septum pellucidum, fimbria fornix, and parahippocampal gyrus. The neoplastic cells had round to elongated nuclei with coarsely stippled chromatin and indistinct cytoplasm borders. They were haphazardly arranged in a finely fibrillary neuropile (Figure 4(a)). Focally the neoplastic invasion was associated with a diffuse vacuolization of white matter and a coexisting foam cell infiltration. A mild neoplastic infiltration of the fourth ventricle floor was also observed associated with necrosis of the medial vestibular nucleus. Immunolabeling on selected paraffin-embedded brain sections was performed with avidin-biotin peroxidase complex staining for GFAP (GFAP, rabbit polyclonal antibody, 1 : 500, Dako, Carpenteria, CA, USA). It showed a diffuse marked GFAP-immunoreaction of the neoplastic cells (Figure 4(b)). The histological findings were consistent with GC of astrocytic type.

2. Discussion

GC is a primary wide and diffuse infiltration of the CNS by neoplastic glial cells [1]. It has been reported in humans, with approximately 300 described cases [26], and very rarely in dogs, cats, and goats [715]. In humans two general forms of GC are recognized. Type I GC is characterized by neoplastic infiltration with good architecture preservation and without a grossly visible mass. Type II GC consists in a neoplastic infiltration accompanied by a mass lesion with ill-defined margins [1, 2, 4, 5, 7, 16]. Affected patients show clinical signs reflecting tumor localization. In humans GC is primarily localized in the telencephalon with mono- or bilateral involvement of different brain lobes. However, more areas may also be affected such as basal nuclei, thalamus, hypothalamus, corpus callosum, cranial nerves, cerebellum, brainstem, and spinal cord [17, 18]. Infiltration of the brain parenchyma typically occurs along the myelinated fibres producing extensive demyelination of the affected areas [19]. This condition gives a histological pattern, which strictly correlates with the hemispheric white matter signal changes seen on MRI [1820]. In human medicine, MRI features of histologically confirmed GC include extensive, symmetrical, or asymmetrical and often poorly delineated T2-hyperintense lesions, which tend to be T1-iso to T1-hypointense with variable contrast enhancement [2, 16, 21, 22]. However, these findings are considered nonspecific for GC so that the differential diagnosis includes immunomediated or virus induced white matter diseases, leukodystrophy, and other brain tumors such as primary brain lymphoma or glioblastoma multiforme [18]. In people, the involvement of commissural structures such as the corpus callosum has been unequivocally related to GC allowing differentiating this neoplasm by a demyelinating disease [20, 21].

To date 12 cases of GC have been reported in dogs with multiple distributions of the lesions throughout the brain and spinal cord, while MRI patterns have been described in five cases [7, 9, 11, 12, 2224]. Three dogs had telencephalic involvement [9, 11, 13] but only one of them showed bilateral and symmetric white matter T2 and FLAIR hyperintensities without any apparent mass effect or anatomical distortion [11]. Based on the MRI features in that case the lesions were misdiagnosed as a leukoencephalopathy of unknown origin. However information about the involvement of the corpus callosum was not available in the referred study [11].

In this report we describe a similar imaging pattern with bilateral and symmetric involvement of dorsal parieto-occipital white matter and corpus callosum, that showed inhomogeneous T2-hyperintensity at level of the splenium. Interestingly in people the involvement of corpus callosum has been reported in 8 out of 9 cases of cerebrum GC [17]. This is why this finding is nowadays considered very suggestive for GC [20, 21, 25].

Generally, bilateral and symmetric involvement of specific anatomic areas of CNS, without relevant mass effect and with anatomic preservation, is typical of metabolic-toxic and degenerative diseases [2628]. On the contrary, primary CNS tumors usually appear as space-occupying lesions with mass effect and peritumoral edema of variable grade [2628]. This peculiar distribution pattern of GC clearly subverts these rules and makes the MRI-based ante mortem diagnosis difficult.

Similar and sometimes overlapping imaging findings have been described in dogs with different brain diseases [13, 2830]. In NLE, MRI findings are characterized by bilateral but not symmetric T2 and FLAIR hyperintensities in the telencephalic white matter [13, 29]. Moreover anatomical distortion may be present. Also in necrotizing leukoencephalitis mononuclear pleocytosis is generally found at cerebrospinal fluid analysis [13, 29]. In our case mild mononuclear pleocytosis detected at CSF analysis was consistent with previous cases of GC in dogs in which normal or mild increased CSF cell count has been found [7, 12, 22]. Canine distemper meningoencephalitis has been reported in an adult dog showing diffuse bilaterally and symmetric T2 hyperintensity of the subcortical parietal and frontal white matter [30]. In that case the predominately white matter distribution of signal intensity changes was consistent with the extensive white matter involvement occurring in the chronic forms [30]. As in this previous report, in our case the negativity of the PCR analyses made chronic distemper encephalitis less likely; any way it did not allow us to rule out distemper definitely. Cerebral edema is generally T1-hypointense, T2-hyperintense and it does not enhance after contrast medium administration [28]. Moreover, edema tends to diffuse along white matter and may outline the corona radiata providing a typical spiked pattern and uniform signal changes [28]. However, extensive cerebral edema is generally associated with a primary brain lesion either vascular, inflammatory, or neoplastic [28]. None of these lesions were evident in our case. To the authors’ knowledge none of the toxic-metabolic and degenerative diseases described in dogs show an imaging pattern of diffuse bilateral and symmetric T2 and FLAIR hyperintensities in the white matter. On the contrary, bilateral but selective focal involvement of different brain areas has been observed in thiamine deficiency [31, 32], hepatic encephalopathy [28, 33], osmotic myelinolysis [28], and L2-hydroxyglutaric aciduria (L2-HGA) [34]. A degenerative disease showing MRI patter comparable to that observed in this case is the adult cheetah leukoencephalopathy [35]. This is a diffuse and bilateral white matter disease of unknown etiology described in 1999 in a group of cheetahs, characterized by axonal and myelin degeneration and reactive astrocytosis [35]. However, there are no reports on a similar disease in dogs.

In this study the areas affected by neoplastic infiltration did not express contrast enhancement suggesting blood-brain integrity [36] as described for low-grade gliomas [2628] and GC in humans [16, 37] as in dogs [9, 11, 22, 24].

In conclusion we describe a distinct MRI pattern of GC in a dog in which bilateral and symmetric T2 and FLAIR white matter hyperintensities were found in the parieto-occipital lobes and corpus callosum. The lesions were isointense on T1W images without contrast enhancement and mass effect. Our imaging findings, unusual for brain tumor, were confirmed as repeatable imaging pattern in canine GC. The most likely differential diagnoses based on MRI signal changes and anatomical preservation seem to be chronic canine distemper meningoencephalitis and NLE. To the authors’ opinion GC should be considered in the list of differential diagnosis in a dog having a so peculiar MRI pattern.

Abbreviations

GC:Gliomatosis cerebri
MRI:Magnetic resonance imaging
FLAIR:Fluid attenuated inversion recovery
CSF:Cerebrospinal fluid
PCR:Polymerase chain reaction
GFAP:Glial fibrillary acidic protein
CNS:Central nervous system
NLE:Necrotizing leukoencephalitis.

Conflict of Interests

The authors declare that there is no conflict of interests regarding the publication of this paper.