Cerebellar diseases that reduce the cerebellar volume can be classified into congenital and developmental diseases in dogs (3,4). Congenital cerebellar diseases include cerebellar aplasia, hypoplasia, and dysplasia (2-4). Cerebellar cortical degeneration, a developmental disease, is commonly called cerebellar abiotrophy (2-4,12). The onset of the clinical signs of cerebellar cortical degeneration varies from neonatal to adult, but the condition is progressive (3,4,11). In many reports, a diagnosis of cerebellar cortical degeneration in dogs was made based on magnetic resonance images (MRI). The cerebellar cortical degeneration on MRI is characterized by a reduction in the size of the cortical folia, widening of the fissures, increased conspicuity of the cerebrospinal fluid (CSF) between the cerebellar folia and enlargement of the fourth ventricle (12,17,19). On the other hand, this qualitative analysis is subjective and can occasionally be obscure. Several studies used quantitative analysis for objective evaluation, such as cerebellar volumetry, brainstem-to-cerebellar cross-sectional area ratio (BS:CBM ratio), and T2-signal intensity histograms (8,18). The BS:CBM ratio and T2-signal intensity histograms can be evaluated easily and quickly because they only use the mid-sagittal plane of the T2-weighted images and are less susceptible to breed and age because the size of the cerebellum shows a consistent ratio with other brain regions in normal dogs (8,18). Referring to D. Henke’ study, severely affected dogs had a mean relative CSF space of 26.7%; moderately affected dogs had a mean CSF space of 23.8%, and mildly affected dogs had a mean relative CSF space of 20.5% on the T2-signal intensity histograms (8).

This paper describes cerebellar cortical degeneration in three dogs based on MRI using qualitative and quantitative analyses.

A five-year-old intact male Beagle (Dog 1) presented with abnormal gait and tremors. The physical examination revealed hypermetry and wide-based stances for gait. A 13-year-old spayed female Shih-tzu (Dog 2) presented with a right head tilt and mild loss of balance for gait. A five-month-old intact male Bernese Mountain (Dog 3) presented with progressively worsening bilateral forelimb ataxia. On a physical examination, mild hypermetry in the gait was identified in Dog 3. The neurological examination showed no abnormal findings in the mental status, cranial nerve examination, and spinal reaction in all three dogs. The complete blood count and serum biochemistry were within the normal reference range in all dogs. The MRI (1.5 Tesla unit, Vantage ElanTM^TM, Canon Medical Systems, Japan) examinations were performed with the dogs in sternal recumbency. The MRI sequences were T1-weighted (TR/TE = 750/12, 560/9, and 750/12 on the transverse, sagittal, and dorsal planes, respectively), T2-weighted (TR/TE = 6143/108, 4071/84, and 6143/108 on the transverse, sagittal, and dorsal plane, respectively), and T2-FLAIR (TR/TE = 8000/120, TI = 2450 on the transverse plane). Additionally, Dogs 1 and 2 underwent susceptibility-weighted imaging, diffusion-weighted imaging and post-contrast T1-weighted imaging to rule out other brain disorders such as brain infarction or tumors. As a result, no other brain disorders were found in Dogs 1 and 2 on MRI.

All dogs showed widening fissures, increased CSF conspicuity between the cerebellar folia, and enlargement of the fourth ventricle in the T2-weighted images. These lesions identified on conventional MRI were severe, moderate, and mild in Dogs 1, 2, and 3, respectively (Fig. 1). The T2-weighted mid-sagittal images were used for BS:CBM ratio and T2-signal intensity histograms. Table 1 lists the results of the quantitative analyses. On the BS:CBM ratio, the separation between forebrain and brainstem was denoted using a straight line drawn from the most rostroventral aspect of the cerebellum to the mammillary bodies, and the CSF in the fourth ventricle was excluded (Fig. 2) (18). The values for Dogs 1, 2, and 3 were 86.4%, 80.2%, and 79.5%, respectively. The BS:CBM ratio in Dog 1 was higher than the cut-off value of 85.2%, while it was normal in Dogs 2 and 3 (18). The cerebellum and CSF area around the cerebellum were manually circumscribed for the T2-signal intensity histograms (8). The number of voxels can be calculated using ImageJ^® software (Wayne Rasband, National Institutes of Health, Bethesda, MD) (Fig. 3). The relative CSF space was calculated as the number of voxels within the CSF divided by the number of voxels within the cerebellum and CSF (8). The relative CSF spaces of Dogs 1, 2, and 3 were 28.1%, 24.3%, and 20.8%, respectively. All dogs showed higher results for the relative CSF space using the T2-signal intensity histograms than the cut-off value of 12.8% (8). The severity of the relative CSF space was severe, moderate, and mild in Dogs 1, 2, and 3, respectively. Dog 1 was euthanized due to worsening clinical symptoms, while Dogs 2 and 3 did not come for follow-up examination.

Table 1 . Result values of the brainstem to cerebellum cross-sectional area ratio and T2-signal intensity histograms.

Category	Dog 1	Dog 2	Dog 3	Cut-off
Breed	Beagle	Shih-tzu	Bernese Mountain
Age	5 years	13 years	5 months
B.W (kg)	7.6	5.8	20
CSA (mm2)
Cerebellum	286	248	317
Brainstem	247	199	252
BS:CBM ratio (%)a	86.4	80.2	79.5	85.2
Voxels (n)
Cerebellum	6008	5219	6145
CSF	2343	1679	1610
Cerebellum + CSF	8351	6898	7755
CSF: cerebellum + CSF (%)b	28.1	24.3	20.8	12.8

^aThe cut-off value for the brainstem to cerebellum cross-sectional area ratio was referenced from (18). ^bThe cut-off value for relative CSF space using T2-signal intensity histograms was referenced from (8)..

Figure 1. Cerebellum on the T2-weighted image of the mid-sagittal plane (A-D), the transverse plane of the caudal lobe (E-H), and the dorsal plane (I-L). A normal cerebellum was shown in A, E, and I. The dogs with cerebellar cortical degeneration showed increased conspicuity of CSF between the cerebellar folia, in Dog 1 (B, F, and J), Dog 2 (C, G, and K), and Dog 3 (D, H, and L).

Figure 2. T2-weighted images of the mid- sagittal plane in a normal dog (A) and Dog 1 (B). Manually traced outline of the cerebellum (white arrow) and brainstem (hollow arrow) to calculate BS:CBM ratio.

Figure 3. (A) Relative CSF space area for the T2-signal intensity histograms within a circumscribed cerebellum area. (B) The number of voxels in the CSF space area was calculated using ImageJ^® software.

Cerebellar cortical degeneration has been reported in the Gordon Setter (5,15), Old English Sheepdogs (16), Brittany (9), and American Staffordshire Terriers (1,7,12-14). The dogs in the present study show a variable onset of clinical signs at five months, five years, and 13 years. Despite being only five months old, Dog 2 showed progressive worsening of the clinical signs, and the cerebellar vermis and hemisphere had a normal shape on MRI, making it difficult to diagnose it as a congenital disease, such as cerebellar hypoplasia or dysplasia. The cerebellar cortical degeneration results in decreased cerebellar volume, which can be detected by a reduction in the size of the cortical folia, widening of the fissures, and increased conspicuity of CSF between the cerebellar folia on MRI (12,17,19). When the cerebellar volume is decreased, as in the cerebellar cortical degeneration, the surrounding cisterns, such as the superior cerebellar cistern and the fourth ventricle, undergo a compensatory enlargement (10). When there is uncertainty in visual analysis, quantitative evaluation can be useful. Several documents used quantitative analysis, such as the cerebellar volumetry, BS:CBM ratio, and T2-signal intensity histograms (8,18). Cerebellar volumetry produces a three-dimensional and accurate value by measuring the overall cerebellar volume but is a time-consuming and laboring procedure when measuring the cross-sectional areas in multiple brain slices (18) or lack of reference range for various breeds and ages. The BS:CBM ratio and T2-signal intensity histograms were faster, easier, and less susceptible to breed and age (8,18). Dogs 2 and 3, which show normal values in the BS:CBM ratio, were found to have moderate and mild cerebellar atrophy, respectively, on the T2-signal intensity histograms. Therefore, even if a normal value is observed in the BS:CBM ratio, if cerebellar atrophy is suspected, it is necessary to use T2-signal histograms in combination. The following reasons can be speculated for this. First, cerebellar cortical degeneration may be present with brainstem atrophy because olivary nuclei provide input to the cerebellum through climbing fibers through the caudal cerebellar peduncle (6,18). If concurrent brainstem atrophy is present, it can decrease the sensitivity of the BS:CBM ratio (18). On the other hand, relative CSF space using T2-signal intensity histograms can evaluate concurrent brainstem atrophy independently. Second, if cerebellar atrophy is more prominent in the cerebellar hemisphere than the cerebellar vermis, the sensitivity of the BS:CBM ratio, which is evaluated only based on the sectional area on mid-sagittal image, could decrease more than the sensitivity of T2-signal intensity histograms, which also evaluate the voxels of the surrounding CSF. Third, the number of cases in the referenced and present studies was small. The relative CSF space using T2-signal intensity histograms is predicted to be possible to detect in the early stage of the lesion, mild diseases, or when the BS:CBM ratio is not significant. Other diseases that cause enlargement of the fourth ventricle and quadrigeminal cistern need to be excluded when diagnosing cerebellar cortical degeneration using T2-signal intensity histograms to avoid misdiagnoses.

Cerebellar cortical degeneration can occur at any age in dogs. The BS:CBM ratio and relative CSF space obtained from T2-signal intensity histograms on MRI can be used for a rapid and easy quantitative evaluation to diagnose cerebellar cortical degeneration. The relative CSF space using the T2-signal intensity histograms could be helpful if cerebellar cortical degeneration is strongly suspected, even if it is not significant on the BS:CBM ratio.

The authors have no conflicting interests.