Internal hydrocephalus is the most common malformation of the central nervous system in dogs, which can be successfully treated using ventriculoperitoneal shunting (VPS) (1). Since the introduction of the technique to veterinary neurosurgery in the 1960s (2), a number of studies have been conducted that described structural characteristics of the hydrocephalic brain indicating surgical therapy (3), documented changes in intraventricular pressures (4), VPS associated complications (5, 6), long-term survival (7), and success in terms of improvement of clinical signs (8). Although this information helps to optimize neurosurgical treatment, variation in surgical care and decision-making remains common in veterinary medicine. Refining patient selection and tailoring appropriate treatment to each individual animal is a further important step in improving treatment. In that respect, grading of ventricular distension has important implications for the appropriate surgical technique (9), the occurrence of overshunting related complications (6) and, potentially, on the prediction of clinical effectiveness of shunting. In order to compare results of surgical interventions performed in multiple institutions, a clear classification of structural changes of hydrocephalic brains would be necessary. Currently, ventricular distention is mostly subjectively graded without any standards that define the different grades (1, 10, 11). An objective and easy grading system is necessary that classifies ventricular distension with high reproducibility. It should not require high-level expertise or the need of time-consuming image postprocessing. In this study, we propose a grading system for the classification of ventricular distension in a large cohort of dogs and assess its interobserver variability.

This is the first investigation in veterinary medicine regarding a classification scheme describing 5 grades of ventricular distension in dogs. The investigated system showed a low interrater variability between the neurologist in training and the board-certified neurologist. However, grades 2 and 3 were difficult to differentiate in some images, resulting in a low reliability of these grades (Table 2). The reason for this inconsistency remains unclear. In order to keep the grading simple, the dimensions of the hippocampus and temporal horn were visually assessed. It is possible that variability would decrease if measurement tools were used to quantify these brain structures. The agreement between the board-certified neurologist and the general practitioner was only moderate among all grades, except grade 5 (Table 2). The general practitioner explained that his limited knowledge of brain morphology made it difficult for him to find the proper image plains and even localize structures like the hippocampus or corpus callosum. Therefore, a proper grading of the ventricular system seems to need at least some experience of the observer regarding brain morphology and MRI interpretation.

Surgical treatment for internal hydrocephalus continues to develop. The number of involved ventricles, as well as the amount of ventricular distension is relevant for the outcome of VPS as these findings are associated with a higher risk of hemispheric collapse and subdural hemorrhage (6). One important goal for the future treatment of hydrocephalus would be the development of a success score that allows prediction of which animal will benefit from VPS or not and which animal is likely to develop complications. The presented grading system has the potential of being useful for the definition of one important factor in such scoring: the ventricular dimensions.

Our experience with grading of ventricular dimensions of animals referred to our clinic or from radiologic assessment in our own hospital has shown that the assessment of ventricle dimensions is highly subjective in general. There is a special bias in the assessment of ventricular dimensions of brachycephalic dogs compared to mesocephalic dogs. Enlarged ventricular dimensions often get lower subjective grades in brachycephalic dogs, probably because small brachycephalic dogs frequently show some degree of ventricular enlargement, referred to as “ventriculomegaly.” Ventriculomegaly was for a long time assessed as clinically irrelevant (3, 12, 13). However, it was shown that white matter atrophy and a reduced cerebral blood flow occurs in brachycephalic dogs having “breed specific” enlarged lateral ventricles, which implies abnormal CSF drainage from the ventricular system and negative effects of ventricular distension on cerebral white matter (14, 15). These findings imply that the term “ventriculomegaly” should be avoided and the situation be assessed as ventricular enlargement of lower grades. It is currently under evaluation if white matter atrophy and reduced white matter perfusion are associated with cognitive problems in dogs as in humans (16–18). This grading scheme helps to make the assessment of ventricular enlargement more objective among different breeds and to categorize ventricular size by using standardized anatomical structures. Although brain structures differ between brachycephalic and mesocephalic breeds, the planes chosen for evaluation (P1 and P2) are less affected by the subsequent shortening and rotation of brachycephalic head shapes than very rostral (e.g., olfactory bulb) or very caudal anatomical structures (e.g., craniocervical junction). However ventricular enlargement is a common finding of brachycephalic dogs and result of the skull shape variation (19).

Evans’ index or the VBI was previously used to grade ventricular enlargement and as an indirect evaluation method for ventricular system distention in neurosurgery in human medicine (20). However, the Evans’ index describes a linear model of continuous ventricular enlargement, but it does not resemble a classification system as measurements are performed routinely at one anatomical point. A classification, as used in human medicine, concentrates on more parts of the ventricular system and is able to evaluate the whole ventricular space and its distention (20, 21). As a matter of fact, the interobserver variability of the Evans’ index was never confirmed (21, 22). In this study, we observed a considerable overlap of VBI in relation to the different grades. The VBI was calculated and correlated with the grading system by the certified neurologist. The mean VBI for grade 0 was 0.30 and 0.91 for grade 5 (Table 2). Although our results showed that the higher VBI, the greater the ventricular dilatation, the same VBI was often measured in three different grades, which makes it a rather unreliable parameter for an unequivocal grading system. The reason for this range is unclear. Interobserver differences for calculating Evans’ index were also found in studies on hydrocephalic humans (19). It was found to vary greatly depending on the location and angle of the slice of the CT or MRI at which the frontal horns and maximal inner skull diameters are measured (21, 22). In order to determine VBI, it is necessary to set CSF and brain parenchyma boundaries, which are not always clear cut (21, 22). Given this imprecision, VBI seems rather inappropriate to precisely grade ventricular distension. But as there is a considerable overlap between our graduation and VBI, further studies are needed to test both variables for their correlation with clinical and outcome measures.

One limitation of this study is the presence of only one observer in each of the 3 different experience groups. This is enough to assess the feasibility of the above described grading system, but a correlation between more observers with the same experience level would be helpful to evaluate the reliability of this new grading system among different experience levels. Another possibility would be an interobserver assessment of the neurologist in training and the general practitioner after some training in MRI interpretation and brain anatomy. In consequence interobserver agreement could improve, if this is not the case definitions of specific grades might be specified to increase the agreement. Further validation by intraobserver evaluation to provide evidence of consistency, more interobserver measurements at different experience levels and whether training will improve reliability should be performed on a larger cohort should be performed. Another limitation is that the correlation between the degree of the ventricular enlargement and the clinical symptoms or other MRI findings like periventricular edema or deformation of the interthalamic adhesion has not been evaluated, therefore we cannot conclude whether the grading system has good clinical dependence and relevance. The decision about surgical VPS implantation should be considered based on the individual clinical symptoms, progression of clinical signs and concurrent diseases. The risk of ventricular collapse is very high in case of very advanced dilatation of the ventricles, which reflects a grade 5 in this study (6). In case with grade 1–2 ventricular enlargement, the ventricular enlargement might be without clinical importance (8), but further studies are warranted. However further validation of VBI and this grading system are needed to guarantee a reliable assessment of ventricular enlargement and correlation to clinic and outcome measurements.

The findings of the study allow us to conclude that the system described in this study is a functional method for an objective grading of the ventricular system in dogs but requires basic neuroanatomical knowledge. Although the present grading system might have some inaccuracy in the lower grades, it represents an improvement compared to former methods. Further evaluation of this grading scheme with regard to inter- and intra-observer agreement and to clinical and outcome measurements are needed.

The original contributions presented in the study are included in the article/supplementary material, further inquiries can be directed to the corresponding author.

Ethical approval was not required for the study involving animals in accordance with the local legislation and institutional requirements because only retrospective data from MRI archives were examined, which are not subject to ethical approval according to the German Animal Protection Act.

AC: Data curation, Formal analysis, Methodology, Writing – original draft, Writing – review & editing. MS: Supervision, Writing – review & editing. AO: Writing – review & editing. SH: Data curation, Methodology, Writing – review & editing. KB: Data curation, Methodology, Software, Writing – review & editing. DF: Data curation, Methodology, Supervision, Writing – review & editing.