• Financed by grants from the LUPA project (FP7-HEALTH-201370).

  • This article first published online June 29, 2012. Some unnecessary reference entries were removed. A corrected version of this article first appears online July 26, 2012.

Address correspondence and reprint Eloisa Terzo. E-mail: eloisaterzo@yahoo.itThis article first published online June 29, 2012. Some unnecessary reference entries were removed. A corrected version of this article first appears online July 26, 2012.


Greyhound nonsuppurative meningoencephalitis is an idiopathic breed-associated fatal meningoencephalitis with lesions usually occurring within the rostral cerebrum. This disorder can only be confirmed by postmortem examination, with a diagnosis based upon the unique topography of inflammatory lesions. Our purpose was to describe the magnetic resonance (MR) imaging features of this disease. Four Greyhounds with confirmed Greyhound nonsuppurative meningoencephalitis were evaluated by MR imaging. Lesions predominantly affected the olfactory lobes and bulbs, frontal, and frontotemporal cortical gray matter, and caudate nuclei bilaterally. Fluid attenuation inversion recovery (FLAIR) and T2 weighted spin-echo (T2W) sequences were most useful to assess the nature, severity, extension, and topographic pattern of lesions. Lesions were predominantly T2-hyperintense and T1-isointense with minimal or absent contrast enhancement.


n dogs, idiopathic inflammatory disorders of the central nervous system (CNS) can be divided into a number of subtypes based largely upon histopathological features.

The most clinically relevant canine inflammatory CNS diseases are the following: granulomatous meningoencephalomyelitis, eosinophilic meningoencephalitis, and necrotizing encephalitides, including necrotizing meningoencephalitis, necrotizing leukoencephalitis, and necrotizing vasculitis (steroid responsive meningitis-arteritis).[1-7] Many of these disorders have strong breed associations.[8-13]

Greyhound nonsuppurative meningoencephalitis is a breed-associated disorder with a unique lesion distribution that allows it to be distinguished histopathologically from other inflammatory brain diseases.[14-16] Young Greyhounds, usually less than 12 months of age, develop acute or chronic, progressive, and nonspecific neurologic signs. These include altered mentation and/or behavior, proprioceptive deficit or other signs including head tilt, circling, ataxia, recumbency, and blindness. Although a presumptive diagnosis is often made following exclusion of infectious and other inflammatory disorders, a definitive diagnosis can only be made at postmortem examination.[15]

Histopathologically, affected brains are characterized by severe gliosis and gemistocytosis with mononuclear cell perivascular cuffing in the caudate nucleus and cortical gray matter of the cerebrum and in the periventricular gray matter of the rostral aspect of the brainstem. Milder lesions are found in the cerebellum, the caudal brainstem, and the cranial cervical spinal cord. Lymphocytic plasmacytic meningitis is also observed.[15, 16]

Magnetic resonance (MR) imaging of the brain is an important diagnostic tool for the investigation of canine neurologic disorders. The MR imaging features of many canine inflammatory brain diseases have been reported;[17-20] however, the features of Greyhound nonsuppurative meningoencephalitis have not been described. Our aim was to describe the MR imaging features of histopathologically confirmed Greyhound nonsuppurative meningoencephalitis.

Materials and methods

Four unrelated dogs were included in the study; one male and three females ranging in age from nine to 14 months. Historical features included obtundation (n = 2), altered behavior (n = 2), pacing (n = 1), and anorexia (n = 1). Clinical signs had been noted from two to four months prior to referral. In all dogs, at least one littermate had died or had been euthanized due to progressive neurological disease. In total, 16 littermates were affected, and Greyhound nonsuppurative meningoencephalitis was confirmed by postmortem examination in eight. At least one littermate of each dog had undergone euthanasia due to the progressive neurologic disease, and nonsuppurative meningoencephalitis had been confirmed histopathologically in the littermates of three of the four dogs. Neurological examination findings included altered behavior (n = 3), proprioceptive or postural reaction deficits (n = 3), and circling (n = 1). Hematology and biochemistry findings were unremarkable in all dogs. Cerebrospinal fluid analysis was performed in three dogs. Mild mononuclear/lymphocytic pleocytosis was reported in all three (nucleated cell count range 6–12 cells/ul (reference interval 0–5)) and protein concentration ranged from 22.8 to 30mg/dl (reference interval <25 mg/dl). All Toxoplasma, Neospora, and distemper tests were negative.

Magnetic resonance imaging was performed using either a 1.5T unit (1.5T Siemens in dog 1; a 1.5T Gyroscan NT [Philips Medical Systems] in dogs 2 and 3) or a 0.4T unit (Aperto; Hitachi in dog 4). In all dogs, T1-weighted (T1W) sequences (pre- and postadministration of gadolinium) were used with a repetition time (TR) between 400 and 700 ms and an echo time (TE) of 10–15 ms; T2-weighted (T2W) sequences were obtained using a TR of 2000–4000 and TE of 100 ms. Fluid attenuation inversion recovery (FLAIR) sequences were obtained selecting a TR of 5000–6000 and TE of 100–125 ms, with inversion recovery time (TI) delay of 2000 ms. Images of the brain were acquired in transverse, sagittal, and dorsal planes. Postcontrast T1W images were acquired immediately after intravenous bolus administration (0.1mmol/Kg) of either gadodiamide1 or gadopentetate dimeglumine.2 The images of each dog were assessed by one radiologist (F.Mc.). The presence of any lesion, neurotopography and parenchymal distribution (gray matter, white matter, or both) were recorded. The margination of any lesions was classified as well-defined or ill-defined. The intensity of the magnetic resonance signal was subjectively evaluated as isointense, hyperintense, hypointense, or heterogeneous/mixed signal intensity relative to normal gray matter on T1W and T2W images and FLAIR sequences. Mass effect was evaluated on the basis of the presence of effacement of sulci, shift of midline structures, and dilation of the ventricular system due to a compression effect. Hydrocephalus was defined subjectively as dilation of the cerebral ventricles greater than that observed in healthy Greyhounds. The presence or absence of meningeal involvement was evaluated in T1W postcontrast and FLAIR images. Pre- and postcontrast T1W images were compared to detect contrast enhancement, and when present, the pattern of enhancement was noted. Contrast enhancement was described as mild, moderate, or severe (intensity similar to or greater than fat), with or without ring enhancement. The extent of white matter T2-hyperintensity, presumed to be vasogenic edema, was assessed subjectively on T2W and FLAIR images.

All dogs underwent euthanasia by the administration of intravenous barbiturate.3 Euthanasia was performed immediately following imaging in three dogs and after 17 days of unsuccessful treatment in the final dog. Postmortem examinations were performed in all dogs. Brain tissue was fixed in 10% formalin. Sections of olfactory lobes, frontal lobes, cerebellum, midbrain, pons, and medulla were paraffin embedded and 5-μm sections stained with hematoxylin and eosin suitable for histopathological evaluations.


The MR image findings are summarized by dog.

Dog 1: There were bilateral areas of T2-hyperintensity in gray matter and subcortical white matter of the olfactory lobes and ventral aspects of the frontotemporal lobes and in the caudate nuclei (Fig. 1A). There was mild cortical atrophy, focal dilation of the cavity of the olfactory bulbs, and abnormal periventricular hyperintensity was noted within the olfactory lobes (Fig. 1A). Changes were slightly more severe in the left cerebral hemisphere where there was a small (2–3 mm) area of cavitation in the subcortical white matter adjacent to the middle cerebral artery (Figs. 2A and B, and 4A and B). There were also small areas of T2-hyperintensity in the gray matter of the cingulate gyrus. Heterogeneous/mixed T2-hyperintensity extended from the thalamus ventrally and caudally into the rostral mesencephalon/pons (Fig. 4B). The mesencephalic changes were less severe than the cerebral changes. There was no abnormal parenchymal or meningeal enhancement (Fig. 3).

Figure 1.

Dorsal plane T2-weighted images from dog 1 (A), dog 2 (B), and dog 3 (C). Note the similarity of lesion distribution in all three dogs with bilateral, almost symmetric increased signal within the olfactory lobes involving both gray and white matter. In dog 1 (A) there are also bilateral areas of poorly marginated T2-hyperintensity in the internal capsule and subcortical white matter of the frontotemporal lobes. (R-right)

Figure 2.

Transverse plane FLAIR images of the cerebrum of dog 1 (A and B). There is a symmetric T2-hyperintensity in the region of caudate nuclei and ventral parts of the frontotemporal lobes (A). The lesions are predominately within the gray matter and appear relatively sharply marginated. Some heterogeneous T2-hyperintensities involve the region of basal nuclei and adjacent white matter. Further caudally the lesions are asymmetric, poorly marginated and involved the ventral left thalamus, piriform lobes, and subcortical white matter ventrally (B). (R-right)

Figure 3.

Transverse plane T1-weighted postcontrast image of the cerebrum of dog 1 at the same level as Fig. 2. Note there is no abnormal contrast enhancement of the meninges or brain parenchyma. (R-right)

Figure 4.

Transverse plane T2-weighted (A) and FLAIR images (B) at the same level of dog 1. Areas of heterogeneous T2-hyperintensity involve the region of basal nuclei and adjacent white matter. Within the subcortical white matter of the left temporal lobe adjacent to the left middle cerebral artery there is a small area of cavitation. (R-right).

Dog 2: There were multifocal symmetric lesions in the brain that were predominantly ventral. There was mild symmetric cortical atrophy in the rostral cerebrum, most marked adjacent to the right and left olfactory lobes. There was mild increased T2-hyperintensity within the superficial cortical gray matter of the rostroventral frontal lobes and ventral aspect of the olfactory bulbs. There was a linear T2-hyperintensity within superficial gray and subcortical white matter adjacent and rostral to the middle cerebral arteries (Fig. 1B). There was no mass effect. There was no dilation of the ventricular system or abnormal contrast enhancement.

Dog 3: There were multifocal, mainly asymmetric lesions. There was mild ventricular dilation and cortical atrophy with the atrophy being most severe around the olfactory bulbs. There was severe bilateral, almost symmetric, T2-hyperintensity within both olfactory bulbs (Fig. 1C). The gray matter changes were more severe within the left olfactory bulb. There was linear T2-hyperintensity within the deeper layers of the superficial cortical gray matter of the ventral aspect of the left frontal lobe, extending caudally to the gray matter and subcortical white matter adjacent to the left middle cerebral artery. The rostroventral aspects of the caudate nuclei had slightly pinpoint heterogeneous areas with subtle increased T2-signal (Fig. 1C). The FLAIR images were characterized by diffuse hyperintensity in the cortical gray matter of the left olfactory-frontal lobes and slightly heterogeneous hyperintensity in the superficial cortical gray matter in the occipital and parietal lobes and adjacent to the right middle cerebral artery. Dorsal to both caudate nuclei and adjacent to the lateral ventricles there was a mild focal T2-hyperintensity. There was no mass effect. There was mild, linear meningeal enhancement surrounding the olfactory bulbs.

Dog 4: There were mild multifocal, asymmetric, and ill-defined T2-hyperintense lesions, affecting mostly the rostroventral cerebral cortex. Changes were of greatest severity in the region of the olfactory bulbs, which had dilated cavities. No abnormal contrast enhancement was present.

The brain was normal grossly in all dogs. Similar to the previous reports, histopathologically, there was patchy nodular gliosis, gemistocytosis, perivascular cuffing and prominent vascularity within cerebrocortical gray matter as well as periventricular tissues in the regions of the caudate nuclei. Since the original descriptions of the lesions[15] are more extensive sectioning-enabled identification of greater involvement of the rostral portions of the cerebrum; and, in particular, the olfactory bulbs and lobes with prominent gliosis and perivascular cuffing with occasional tissue rarefaction. Cerebral white matter was also occasionally involved, and foci of gliosis were noted within the midbrain. The medulla and pons displayed mild to moderate gliosis in one dog, but were unaffected in the remaining dogs. Gliosis of the white matter of the cerebellum was apparent in three dogs. Although not observed in the original study,[15] this has been observed occasionally in the subsequent dogs (unpublished data). Meningeal mononuclear (lymphocyte and plasma cell) infiltrations were prominent over the olfactory lobes in all dogs, and over the frontal lobe and midbrain in two dogs.


Definitive diagnosis of inflammatory brain disease requires histopathologic examination, so the term meningitis/meningoencephalitis of unknown etiology has been proposed when histopathologic evaluation is lacking. Meningoencephalitis of unknown etiology may be divided into two broad groups, granulomatous meningoencephalitis and necrotizing encephalitis/leucoencephalitis,[18] with granulomatous meningoencephalitis being subdivided further into three forms based on the morphology and clinical findings; the disseminated, the focal, and the ocular forms.[21-23] The disseminated form is most common with image findings reflecting a multifocal distribution. Lesions are typically T2-hyperintense and poorly marginated with a predilection for white matter although gray matter involvement occurs.[23] Necrotizing encephalitis has been divided into two types, necrotizing meninogencephalitis and necrotizing leucoencephalitis, but there is a overlap in clinical and neuropathologic findings and the encompassing term necrotizing encephalitis has been suggested to describe these diseases.[18] Necrotizing encephalitis occurs in a number of small breed dogs and was described historically as breed specific encephalitides, such as pug encephalitis.[8, 11, 19][20]

Greyhound nonsuppurative meningoencephalitis, which is of unknown etiology and may affect multiple dogs in a litter, is characterized by infiltrative inflammatory lesions. The distribution of lesions is predominately in gray matter with the cerebral cortex and caudate nuclei most severely affected. In the current study, imaging abnormalities were present in the olfactory and frontal lobes (four out of four dogs) and caudate nuclei (two out of four dogs). These changes are consistent with the histologic lesions.[15]

All dogs in this study had histological evidence of meningitis, but imaging evidence of meningitis was present in only one dog. A similar disconnect has been found in granulomatous meningoencephalitis in which meningitis was not seen on MR images despite histological evidence of meningeal disease.[23] Therefore, contrast enhancement of intracranial lesions does not reflect histopathologic findings consistently, and histologic features of lesions cannot be predicted solely by contrast enhancement patterns.[24]

The distribution of image changes was similar in all four dogs, with a predilection for the rostroventral parts of the cerebrum, especially the olfactory lobes. The consistent involvement of the olfactory lobe is different from other nonsuppurative meningoencephalitides in which no olfactory involvement has been reported, suggesting unique topographic predilection (Figs. 1A, B, and C). The olfactory pathways are a part of the rhinencephalon, primarily related to olfaction and commence in the highly vascularized olfactory mucosa, where bundles of neurons form the olfactory nerves and terminate in the olfactory bulb through the cribriform plate.[26] While these pathways may be affected by inflammatory processes induced by viral, fungal and bacterial agents, and neoplastic disorders originating in the nasal cavity, no agents were identified and the cause for this unique lesion distribution remains unknown.[27] Although obvious olfactory lesions were noted on MR images in three of the four Greyhounds and confirmed histopathologically, none of these patients had evidence of nasal disease.

On MR imaging, the distribution of the lesions and relative sparing of internal capsule, absence of significant cavitation, minimal or no contrast enhancement, and no mass effect are important features to distinguish Greyhound nonsuppurative meningoencephalitis from other meningoencephalitides of unknown etiology.

Asymmetric, multifocal T2-hyperintensities are typical of inflammatory CNS disease and of the noninfectious encephalitides, Greyhound nonsuppurative meningoencephalitis and necrotizing meningoencephalitis share similar MR imaging characteristics. However, a diagnosis of necrotizing meningoencephalitis would not be considered in Greyhounds as it appears confined to small breed dogs, such as the pug.[20] Greyhound nonsuppurative meningoencephalitis with predominantly gray matter lesions may resemble disseminated granulomatous meningoencephalitis on MR imaging.[2, 23, 25, 28] However, the subcortical white matter involvement and the distribution primarily within the forebrain are not typical features of granulomatous meningoencephalitis, and this is an important difference. The Greyhound nonsuppurative meningoencephalitis patients had multifocal T2-hyperintensities within the cerebrum affecting gray and white matter. Disseminated granulomatous meningoencephalitis has a white matter predominance and variable gray matter involvement and lesions are usually poorly marginated with variable vasogenic edema and contrast enhancement. Brainstem involvement is common in granulomatous meningoencephalitis, but mild changes were only seen in two dogs with Greyhound nonsuppurative meningoencephalitis.

The lesion distribution differs in necrotizing meningoencephalitis and necrotizing leucoencephalitis, with necrotizing leucoencephalitis often sparing the cerebral cortex and meninges and predominately affecting the white matter whereas necrotizing meningoencephalitis commonly affecting the cerebral cortex, subcortical white matter, and meninges.[17, 18] The MR imaging findings reflect the histopathologic distribution with multifocal, asymmetric lesions affecting the prosencephalon. With necrotizing meningoencephalitis, the lesions involve the gray and white matter and are typically T2-hyperintense and T1-isointense to T1-hypointense and may have slight contrast enhancement.[19, 25] Loss of gray and white matter demarcation has been reported in necrotizing meningoencephalitis.[19] In necrotizing leucoencephalitis, there are often areas of necrosis and cavitation. The areas of cavitation may be larger in necrotizing meningoencephalitis and are related to the duration and severity of the disease. The MR imaging features of greyhound nonsuppurative meningoencephalitis differed from necrotizing leucoencephalitis, with involvement of the cortical gray matter in all dogs and relative sparing of the white matter. Lesions in necrotizing leucoencephalitis are the most severe in the occipital and parietal lobes, which may also help to differentiate necrotizing leucoencephalitis from Greyhound nonsuppurative meningoencephalitis.[17, 23] Cavitation due to necrosis is commonly seen in necrotizing encephalitis and appears as areas of fluid signal within the brain parenchyma. In the current study, only the most severely affected Greyhound had evidence of cavitation with a small area of fluid signal within the subcortical white matter.

Fluid attenuation inversion recovery sequences have a high sensitivity for the detection of lesions in dogs with inflammatory CSF.[17, 30] In all four dogs with Greyhound nonsuppurative meningoencephalitis, the lesions were best visualized on FLAIR images (Figs. 2A and B, and 4B).

Cortical atrophy may be secondary to inflammatory mediators, as for example in Maremma Shepherd or Rottweiler dogs;[30, 31] however, neither eosinophilic involvement nor inflammatory processes are present in Greyhounds affected by Greyhound nonsuppurative meningoencephalitis. Cortical brain atrophy, commonly seen in humans with inflammatory brain lesions independent of the cause, is a nonspecific response to the irreversible neurologic injury.[32-34] This suggests that as brain tissue degenerates, atrophy may be observed.

In summary, magnetic resonance imaging findings largely reflect the unique histopathologic distribution of lesions observed in Greyhound nonsuppurative meningoencephalitis. The intense inflammatory process affecting the olfactory lobes appears unique to this disorder. The combination of breed, age of onset of clinical signs and pattern of MR imaging abnormalities is highly suggestive of Greyhound nonsuppurative meningoencephalitis, and MR imaging can be very useful to make a presumptive ante-mortem diagnosis in affected dogs.


The authors would like to acknowledge the staff of Davies Specialist Diagnostic Imaging Unit, particularly Francisco Llabrés-Díaz for imaging data provided for dog 4. This study was financed by grants from the LUPA project (FP7-HEALTH-201370).

  1. 1

    OMNISCAN®, Nycomed Imaging, Oslo, Norway.

  2. 2

    Magnevist®, Bayer Schering Pharma AG, Germany.

  3. 3

    Euthatal® May & Baker, Dagenham, England.