Lack of association between inner ear FLAIR suppression on MRI and laterality of clinical signs in feline idiopathic vestibular syndrome

Abstract

Objectives

The underlying aetiology of feline idiopathic vestibular syndrome (IVS) remains unclear. In humans, specific MRI sequences are used to demonstrate differences in endolymph composition between unaffected and affected inner ears. A previous study in dogs with IVS reported that in 93% of cases, the clinically affected side exhibited a lower degree of fluid-attenuated inversion recovery (FLAIR) suppression, which may indicate altered endolymph composition, potentially due to increased protein concentration. The objective of this study was to determine whether similar changes can be observed in cats with IVS.

Methods

Medical records from three referral hospitals were reviewed. Cats were included if they had a diagnosis of IVS, clear lateralisation of clinical signs and had undergone MRI of the brain. Regions of interest (ROIs) were manually outlined to define the inner ear on T2-weighted and FLAIR images. The FLAIR suppression ratio for each ear was assessed by calculating the average signal intensity across both imaging sequences. Any unilateral reduction in suppression was compared with the direction of clinical signs. ROI size was also measured to evaluate potential swelling or collapse of inner ear structures.

Results

A total of 25 cats met the inclusion criteria. No correlation was found between the clinically affected side and the degree of FLAIR suppression. In 11/25 cats, the identified abnormal side did not correspond to the lateralisation of clinical signs, whereas in 14/25 cats it did. ROI measurements likewise showed no consistent association with the clinically affected side.

Conclusions and relevance

In the present cohort, these findings indicate that cats with IVS did not show a consistent association between clinical lateralisation and MRI-detectable endolymph changes, supporting the possibility of a different underlying pathophysiology compared with dogs.

Plain language summary

No link between MRI inner ear changes and the side of signs in cats with idiopathic vestibular syndrome

The cause of idiopathic vestibular syndrome (IVS) in cats is still unknown. This condition makes cats suddenly lose their balance, tilt their heads or walk in circles. In humans, certain MRI techniques can show fluid changes inside the inner ear on the affected side. A previous study in dogs with IVS found that in most cases, the inner ear on the side showing signs had a different MRI appearance, suggesting a change in the fluid composition of the ear. In this study, we looked for similar changes in cats with IVS. We reviewed medical records from three referral hospitals and included cats with clinical signs relating to one side of the body that had undergone MRI scanning. For each cat, we compared the inner ears on both sides using two MRI sequences (T2-weighted and fluid-attenuated inversion recovery). We measured how much signal was suppressed in each ear to see if one side showed less suppression, which could indicate a change in the inner ear fluid. A total of 25 cats met the study criteria. We found no consistent link between the side of the clinical signs and the MRI changes. In 11 cats, the MRI findings did not match the side of the signs, while in 14 cats, they did. We also measured the size of the inner ear structures, but this showed no clear pattern related to the affected side. These results suggest that IVS in cats may have a different underlying cause than in dogs.

Introduction

Vestibular syndrome (VS) is a common condition in cats resulting from dysfunction of any component of the vestibular system. ¹ Clinical signs include head tilt, pathological nystagmus, strabismus, gait abnormalities such as ataxia, and may also include nausea and emesis.^1,2 Clinical signs may arise from either central or peripheral causes. Central VS may be associated with dull mentation, cranial nerve deficits beyond those involving cranial nerves VII and VIII, which may also be affected in peripheral VS, as well as postural deficits; however, the absence of these findings does not rule out central involvement. ³ Central causes include vascular, infectious/inflammatory, traumatic, anomalous, metabolic/toxic and neoplastic aetiologies. Peripheral causes of VS in cats include otitis media/interna, middle ear polyps, neoplasia and idiopathic vestibular syndrome (IVS).^{4 –7}

IVS is common in cats and is typically described as an acute to peracute, non-painful peripheral vestibular disorder affecting cats of any age.^7,8 Maintaining fluid balance within the inner ear is a complex mechanism. Even minor changes in fluid composition – resulting from degeneration, excessive secretion, impaired absorption, shifts in osmotic pressure or changes in blood vessel permeability within the peri- or endolymphatic tissues – can cause considerable disturbances in balance or hearing, as described in human inner ear physiology. ⁹ The aetiology of IVS, however, remains unknown. Most cats show clinical improvement within a few days, with complete resolution of clinical signs within 2–4 weeks. However, in some instances, mild residual clinical signs, such as a slight head tilt or mild ataxia may persist. ³ The final diagnosis of IVS continues to rely on the exclusion of discernible diseases. ¹⁰ To rule out causes for VS, the literature recommends general clinical examination, neurological examination, complete blood count (CBC), serum biochemistry (SE), thyroxine serum concentration, otoscopy, blood pressure measurement, MRI and cerebrospinal fluid (CSF) examination.^6,8 Recent studies have challenged the assumption that MRI is always normal in dogs with IVS. In dogs and cats without vestibular signs, perilymph and, in practice, also endolymph, usually appears as low signal on fluid-attenuated inversion recovery (FLAIR), consistent with CSF‑nulling. ¹¹ Foth et al reported a consistent lack of FLAIR suppression of the inner ear on the clinically affected side in dogs with IVS. This MRI appearance has been interpreted as reflecting alterations in endolymph composition, such as increased protein content. Based on these findings, the authors questioned whether the term ‘idiopathic’ accurately reflects the underlying pathophysiology and proposed the terminology ‘benign peripheral vestibular syndrome’. ¹²

The aim of the present study was to assess whether similar FLAIR abnormalities are present in cats with IVS.

Materials and methods

Medical records from three veterinary referral clinics, collected between July 2009 and July 2020, were reviewed. The collected data included signalment, medical history, including onset and progression of clinical signs, as well as findings from neurological examinations, with an emphasis on the lateralisation of clinical signs. In addition, results of further diagnostic testing (including CBC, SE, otoscopy, blood pressure measurement, MRI of the head and CSF examination) and administered medications were recorded. The diagnosis of IVS was established in cases where all diagnostic findings were unremarkable and no underlying cause for the vestibular signs could be identified. ⁸

MRI was performed using either a 1.5 T magnet (Intera; Philips Medical Systems and MAGNETOM; Siemens Healthineers) or a 3 T magnet (Achieva dStream; Philips). Although specific MRI protocols varied among cats, all examinations included, at minimum, T2-weighted (T2W) and T2W fluid-attenuated inversion recovery (T2-FLAIR) sequences with a slice thickness of 3–4 mm, encompassing the level of the inner ear. The typical imaging parameters for the T2W sequences were repetition time (TR) of 3000–5500 ms and echo time (TE) of 80–110 ms. For T2-FLAIR sequences, TR was in the range of 6000–10,000 ms, TE was 120–140 ms and inversion time was 2000–2600 ms. Because side-to-side differences are typically not perceptible on visual inspection of routine images, all assessments were performed quantitatively. All required measurements were performed by a final year neurology resident at the European College of Veterinary Neurology, blinded to signalment, clinical presentation, lateralisation and final diagnosis, with formal training in MRI interpretation. Both inner ears were identified in the transverse T2W images. Their outlines were defined as the region of interest (ROI), using the image processing program ImageJ (Wayne Rasband; National Institute of Health), and the mean grey value was measured. These outlines were then transferred to the corresponding T2-FLAIR image to ensure the identification of the same anatomical area, allowing the calculation of the mean grey value of this sequence, as previously described. ¹² The FLAIR suppression was quantified for each ROI by computing the ratio of the difference in signal intensity between the T2W and T2-FLAIR images. The pathological side was expected to have a higher protein content, resulting in incomplete FLAIR suppression (ie, higher T2-FLAIR signal intensity) and therefore a lower suppression ratio (see formula below). The side of the inner ear with the lower ratio of suppression was defined as the ‘affected side’, whereas the contralateral side was considered the ‘unaffected side’. The identified affected side was subsequently compared with the lateralisation of clinical signs. The FLAIR suppression ratio for each ROI was calculated by dividing the difference in signal intensity between T2W and T2-FLAIR images on the same side by the signal intensity of the T2W image (see formula below). In addition, the ROI size was measured on the T2W images to allow comparison between the clinically affected and unaffected side.

$$FLAIRsuppression=\frac{(T2W-T2FLAIR)}{T2W}$$

All statistical analyses were performed using GraphPad Prism 10 (GraphPad Software). Data were assessed for significance using the non-parametric Mann–Whitney U-test, with results presented as box plots indicating median, interquartile range (IQR) and minimum/maximum values. Two-sided tests were applied throughout, with P <0.05 considered indicative of statistical significance.

Results

A total of 25 cats met the inclusion criteria and were included in the study (Table 1). The median age was 79 months (range 15–196; n = 24). Age information for one cat was not available. Of the cats, 13 (52%) were male and 12 (48%) were female; all cats were neutered. The affected cat breeds included 17 (68%) domestic shorthairs, four (16%) British Shorthair, one (4%) Russian Blue, one (4%) Siamese, one (4%) domestic longhair and one (4%) Birman. This breed and sex distribution is consistent with previous studies investigating feline peripheral VS of different aetiologies. ¹³ Clinical signs included head tilt in 25 (100%) cats, with 14 cats exhibiting a head tilt to the left and 11 cats showing a head tilt to the right side, ataxia or falling in 23 (92%) cats and nystagmus in 17 (68%) cats. In addition, facial nerve paresis was observed in four cats, with two of those four cats also having Horner’s syndrome. Clinical signs were acute in 17 (68%) cats, peracute in four (16%) cats and not documented in the remaining four (16%) cats. All cats had normal mentation, showed no postural deficits and demonstrated clinical improvement over time. In all cats, both inner ears could be detected to define the ROI, and a unilateral decrease in the suppression ratio was observed in every patient. The side with the lower ratio of suppression was defined as the ‘affected side’, whereas the contralateral side was considered the ‘unaffected side’. The identified ‘affected side’ was subsequently compared with the lateralisation of clinical signs. In the concordant example shown in Figure 1, the clinically affected side exhibited a higher suppression ratio than the contralateral ear; full cohort-level values and concordance are provided in Table S1 in the supplementary material. The median ratio of suppression of the affected side was 0.8329 (IQR 0.6721–0.8907) (Figure 2), whereas the unaffected side had a median of 0.8423 (IQR 0.7418–0.9101; P = 0.2859).

Table 1.

Signalment and ratio of fluid-attenuated inversion recovery (FLAIR) suppression in the inner ear in cats with idiopathic vestibular syndrome (IVS)

	IVS (n = 25)
Age (months)	Median 79 (IQR 71), range 15–196
Sex	13 (52%) male (100% castrated), 12 (48%) female (100% spayed)
Breed	17 (68%) domestic shorthair
	4 (16%) British Shorthair
	1 (4%) Russian Blue
	1 (4%) Siamese
	1 (4%) domestic longhair
	1 (4%) Birman
Ratio of FLAIR suppression	Affected side 0.8329 (IQR 0.6721–0.8907)
	Unaffected side 0.8423 (IQR 0.7418–0.9101)

IQR = interquartile range

Figure 1.

Transverse MRI images at the level of the inner ear in a cat with idiopathic vestibular syndrome: (a) T2-weighted (T2W) and (b) fluid-attenuated inversion recovery (FLAIR). On the T2W image (a), bilateral regions of interest (ROIs) encompassing the vestibular labyrinth were delineated and co-registered to the corresponding FLAIR image (b). Mean signal intensity (SI) within each ROI was measured on both sequences, and the FLAIR suppression ratio was calculated as SI FLAIR/SI T2. In this example (cat 18: clinical head tilt left), suppression ratios were within the cohort range and ROI asymmetry did not concord with clinical lateralisation. This figure illustrates the ROI placement and suppression ratio methodology and is not intended to depict a pathognomonic side-to-side difference. Accordingly, this case demonstrates methodological comparison between quantitative suppression ratios and clinical lateralisation, whereas visual inspection alone is not intended to indicate a diagnostic side-to-side difference. Note that residual FLAIR hyperintensity (higher suppression ratio) may arise from incomplete fluid suppression and/or increased protein content; with the present protocol, abnormal suppression was not detected

Figure 2.

Comparison of the affected and unaffected side of the inner ear of feline patients with idiopathic vestibular syndrome. Boxes represent the interquartile range (IQR), horizontal lines indicate the median and whiskers denote the data range. Diamonds mark the mean values. The ratio of fluid-attenuated inversion recovery (FLAIR) suppression was calculated as a ratio of signal intensity for each region of interest and side as follows:

$$FLAIRsuppression=\frac{\left(T2W-T2FLAIR\right)}{T2W}$$

The median ratio of suppression of the affected side was 0.8329 (IQR 0.6721–0.8907), whereas the unaffected side had a median of 0.8423 (IQR 0.7418–0.9101; P = 0.2859)

Evaluation of the determined affected side, in consideration of the lateralisation of clinical signs such as head tilt, resulted in a 56% agreement between clinical and MRI findings. In 11 (44%) cases, the direction of head tilt did not match the MRI-identified affected side, whereas in 14 (56%) cases, they did match. When comparing the ROI size between sides, a larger area corresponded with the clinically affected side in 15/25 (60%) cats, whereas in the remaining 10 (40%) cats this was not the case. No clinically relevant swelling of the inner ear structures was identified on qualitative MRI assessment.

Discussion

In this retrospective multicentre study, the inner ear structures were clearly identifiable on transverse T2W images and were used to compare both ears for potential intralabyrinthine abnormalities using FLAIR images. FLAIR imaging is considered more sensitive to subtle changes in the peri- and endolymph of the inner ear than conventional T1-weighted or T2W sequences, particularly for detecting elevated protein concentrations in fluids.^11,14,15 On T2W images, both perilymph and endolymph are typically hyperintense. Because perilymph is CSF‑like in composition, normal perilymph is expected to be suppressed on FLAIR MRI; endolymph, although also fluid, has a distinct ionic composition and typically appears low signal on FLAIR under standard parameters. We note that FLAIR hyperintensity of the aforementioned fluids suggests altered protein content rather than normal physiology. Consequently, reduced FLAIR signal suppression of the inner ear is expected in the presence of increased protein or cellular content within the inner ear fluids. To ensure consistent anatomical comparison, ROIs defined on T2W images were transferred to the corresponding T2-FLAIR images. By defining an ‘unaffected side’, each cat served as its own internal control. In contrast to a previously published study of dogs with IVS, ¹² in which 93% of cases showed concordance between reduced FLAIR suppression and the clinically affected side, the degree of FLAIR suppression in the present study did not correlate with the lateralisation of clinical signs in cats with IVS. This finding may indicate species-related differences in underlying disease mechanisms or suggest that feline IVS represents a heterogeneous group of disorders rather than a single disease entity.

In the present cohort, the radiologically defined ‘affected side’ correlated with the lateralisation of clinical signs in only 56% of cases. In addition, the difference between suppression ratios of the radiologically affected and unaffected sides was considerably smaller in cats than previously reported in dogs. ¹² Although dogs showed a median difference of 0.0462 between sides, the difference in cats was only 0.0094. In cats, the observed between-side differences were numerically smaller than in dogs, reducing the dynamic range of the asymmetry metric and, together with the limited sample size, may have contributed to the absence of a significant association between imaging findings and clinical lateralisation. Beyond signal suppression, we quantified the ROI of the vestibular labyrinth to assess potential asymmetry.

In 60% of cats, the clinically affected side corresponded to a larger ROI; in the remaining cats, no such correspondence was observed. These findings indicate that overt swelling or collapse of the labyrinth cannot be consistently demonstrated using conventional MRI. Importantly, focal collapse of individual vestibular structures (eg, ampullae) would not necessarily manifest as a measurable change in the composite ROI given the inclusion of adjacent compartments, partial‑volume effects and sequence resolution. Subtle structural changes below the resolution of standard imaging may therefore still contribute to feline IVS. In human medicine, vestibular atelectasis (VA) is characterised by collapse of endolymphatic compartments – typically involving the utricle and at least two ampullae while sparing the saccule and cochlea – and may present paroxysmally or gradually.^16,17 In our cats, the absence of abnormal FLAIR suppression, together with ROI measurements that did not consistently correspond to the clinically affected side and the acute, self‑limiting signs, supports VA as a plausible differential diagnosis, although direct imaging evidence of ampullary collapse was not obtained. Mechanistically, collapse of the membranous labyrinth may reduce normal perilymphatic separation from the stapes footplate, allowing sound‑ or pressure‑induced stapes motion to abnormally stimulate vestibular receptors.^18,19 A similar mechanism could be hypothesised in cats.

Several other peripheral vestibular disorders described in human medicine may be considered as potential comparators for feline IVS, including benign paroxysmal positional vertigo (BPPV), Ménière’s disease and vestibular neuritis. BPPV is a mechanical disorder caused by dislodged otoconia within the semi-circular canals and is characterised by brief, recurrent episodes of vertigo triggered by head movements. ²⁰ Ménière’s disease, in contrast, is a chronic inner ear disorder marked by recurrent vertigo episodes, fluctuating hearing loss and tinnitus, and is associated with endolymphatic hydrops, reflecting an abnormal expansion of the endolymphatic space that represents a key feature of the disease and is detectable on specialised contrast-enhanced MRI sequences.^{21 –23} The waxing and waning or recurrent nature of both BPPV and Ménière’s disease appears less consistent with the typically acute and self-limiting clinical course observed in feline IVS. Vestibular neuritis, however, is characterised in humans by a sudden onset of prolonged vertigo accompanied by nausea and vomiting, with clinical signs often persisting for several days. ²⁴ This disease course most closely parallels the presentation of feline IVS and therefore represents a particularly relevant human comparator. Vestibular neuritis in humans is thought to result from an acute viral or post-viral inflammatory process affecting the vestibular branch of the eighth cranial nerve, although vascular and immunological mechanisms have also been proposed. Although herpes simplex viruses have been detected in the vestibular ganglion and labyrinth of affected human patients, their role remains incompletely understood.^{25 –28} In cats, herpesvirus infection is common. ²⁹ However, feline herpesvirus-1 has been identified in the vestibular ganglion of a subset of cats without vestibular signs, making a direct causal association with IVS unlikely. ³⁰ The age distribution in the present cohort indicates that IVS in cats is not confined to older or geriatric individuals, consistent with recent studies demonstrating that cats of various ages may be affected. ¹³

This study has several limitations. Because of its retrospective and multicentre design, minor variability in MRI protocols could not entirely be avoided, including differences in scanner models and sequence parameters. Differences in slice thickness and positioning may have further limited comparability, as identical slice orientation could not be ensured in every cat. Although all cats underwent both T2W and FLAIR imaging, subtle differences in acquisition may have affected suppression ratios. All measurements were performed manually using a standardised protocol, and although consistency was maintained by a single trained evaluator, some degree of measurement variability is inherent in ROI-based analysis. Furthermore, the relatively small sample size may have limited the statistical power to detect subtle differences between sides; however, the largely overlapping values suggest that even larger cohorts might not necessarily have yielded clearer separation. Finally, the small size of feline inner ear structures may limit the spatial resolution of conventional MRI and partly explain differences compared with canine studies. Accordingly, higher resolution inner ear-specific imaging sequences or histopathological correlation would be required to confirm vestibular atelectasis. Advanced imaging techniques, such as delayed contrast-enhanced three-dimensional FLAIR, were not available in this cohort and may have revealed subtle structural changes undetectable with conventional sequences. To illustrate the degree of patient-level variability, individual ROI asymmetry plots are provided in the supplementary material.

Conclusions

This study suggests that IVS in cats may be mechanistically distinct from the corresponding condition in dogs. The observed lack of correlation between imaging and clinical findings, the additional evaluation of ROI size showing no consistent asymmetry, the limited difference in suppression ratios and the wide age range of affected cats argue against a degenerative or inflammatory origin. Instead, structural or mechanically induced dysfunctions such as vestibular atelectasis may be involved. Future prospective studies using high-resolution and contrast-enhanced imaging protocols are warranted to further elucidate the underlying mechanisms and to assess the potential diagnostic role of advanced MRI in feline VS.