Brainstem hyperintensity in patients with vestibular schwannoma is associated with labyrinth signal on magnetic resonance imaging but not vestibulocochlear tests

Abstract

Objectives

Focal hyperintensity in the dorsal brainstem (HDB) has been described in large cerebellopontine angle tumours and is thought to represent vestibular nuclei degeneration, but its functional significance has not been thoroughly investigated. Our aim was to analyse its relationship to imaging characteristics of the tumour and inner-ear structures and to vestibulocochlear functional tests.

Methods

We retrospectively reviewed 54 patients with a histological diagnosis of vestibular schwannoma (VS). Magnetic resonance imaging tumour characteristics (size, cystic composition and distance from the cochlear aperture), signal intensity ratio of the cochlea and vestibule in fluid-attenuated inversion recovery (FLAIR) and fast imaging employing steady-state acquisition (FIESTA)/fast spin-echo imaging with variable flip angles (CUBE) and vestibulocochlear function tests (audiometry, auditory brainstem response (ABR) and video head impulse testing (vHIT)) were obtained. Statistical analyses were performed to evaluate their relation to focal HDB.

Results

Focal HDB was found in 22% of VS. It was significantly associated with large (p < 0.001) and cystic (p = 0.004) tumours and also with tumours located further from the cochlear aperture (p = 0.039). The signal intensity ratio of the cochlea on FLAIR was higher in patients with HDB (p < 0.014), but this difference was not observed in FIESTA/CUBE (p = 0.981). Audiometry, ABR and vHIT results did not significantly differ in patients with HDB, but ABR results were worse in patients with higher cochlear signal intensity on FLAIR sequences (p = 0.026).

Conclusions

Focal HDB in patients with VS was associated with increased signal intensity ratio of the cochlea on FLAIR in patients with VS but not directly to the results of vestibulocochlear function tests.

Introduction

Vestibular schwannoma (VS) is the most common tumour of the cerebellopontine angle.¹ It is a benign neoplasm originating from the proliferation of Schwann cells in the vestibular branch of the vestibulocochlear nerve near the myelin transition zone adjacent to the vestibular ganglion.^2,3 Magnetic resonance imaging (MRI) provides excellent tumour characterisation and is the modality of choice for diagnosis and preoperative planning.^4,5 Previous studies have described MRI findings associated with VS: (a) a small area of hyperintensity in the dorsal brainstem (HDB), particularly in large VS, and thought to arise from degeneration of the vestibular nuclei;^6,7 and (b) changes in the signal intensity of ipsilateral inner-ear structures, mainly explained by an increased concentration of protein in the perilymphatic space.^8–10

The functional significance of HDB has not been thoroughly investigated. Therefore, we sought to analyse its relationship with vestibulocochlear function testing and signal alteration of the labyrinth on fast imaging employing steady-state acquisition (FIESTA)/fast spin-echo imaging with variable flip angles (CUBE) and fluid-attenuated inversion recovery (FLAIR) MRI sequences.^11–13

Methods

From January 2015 to January 2019, 61 patients with previously untreated unilateral sporadic VS proven by histopathological examination underwent surgical treatment at our tertiary neurosurgical referral centre. Patients with intralabyrinthine extension of the tumour (one patient) and inadequate preoperative MRI (six patients) were excluded. Fifty-four patients were finally included for analysis. Institutional Review Board approval was obtained. Written informed consent was not required because of the retrospective nature of the study and the anonymisation of patient data.

MRI technique and evaluation

All patients underwent complete brain and inner-ear MRI protocols performed with a 1.5 Tesla scanner (Signa HDxt; General Electric) using an eight-channel head coil or a 3.0 Tesla scanner (Discovery MR750w; General Electric) using a 28-channel head and neck coil. In the 1.5 Tesla scanner, three-dimensional (3D)-FIESTA sequences were obtained with the following parameters: TR = 4.900 ms; TE = 1.800 ms; flip angle = 60°; FOV = 190 mm; matrix = 320 × 224; section thickness = 0.8 mm; pixel size = 0.37 × 0.37 mm; and acquisition time = three minutes and 40 seconds. Three-dimensional-FLAIR sequence parameters in the 1.5 Tesla scanner were as follows: TR = 6500 ms; TE = 130 ms; TI = 1960 ms; flip angle = 90°; FOV = 250 mm; matrix = 224 × 224; section thickness = 1.0 mm; pixel size 0.48 × 0.48 mm; and acquisition time = seven minutes and 20 seconds. In the 3.0 Tesla scanner, 3D-CUBE T2-weighted sequences were obtained with the following parameters: TR = 3000 ms; TE = 414 ms; flip angle = 90°; FOV = 120 mm; matrix = 320 × 320; section thickness = 0.4 mm; pixel size = 0.35 × 0.35; and acquisition time = six minutes and 34 seconds. Parameters for 3D-FLAIR sequences in the 3.0 Tesla scanner were as follows: TR = 5002 ms; TE = 138 ms; TI = 1440 ms; flip angle = 90°; FOV = 250 mm; matrix = 256 × 224; section thickness = 1.2 mm; pixel size 0.50 × 0.50 mm; and acquisition time = three minutes and 53 seconds.

Axial maximum tumour diameter and distance from the tumour to the cochlear aperture were measured by one neuroradiologist. We also classified tumours by size using the Hannover classification: T1, completely intracanalicular; T2, both intracanalicular and small cisternal extension; T3, tumour fills the cerebellopontine angle cistern (with or without brainstem contact); and T4, tumour displaces the brainstem (with or without fourth ventricle displacement).¹⁴ VS were also morphologically classified as solid (tumour shows homogeneous contrast enhancement), heterogeneous (tumour has areas without contrast enhancement but no significant alteration on T2-weighted images) and cystic (tumour has hyperintense cavities within the tumour on T2-weighted images without contrast enhancement of the cavity on gadolinium-enhanced images.¹⁵ Finally, signal intensity measurements were performed on FIESTA/CUBE and FLAIR sequences using elliptical regions of interest (ROIs) in the cochleae, vestibules and cerebellar white matter bilaterally, as previously described by Ishikawa et al.⁹ The ratio of the signal intensity of the cochlea (SIRc) and vestibule (SIRv) to that of the cerebellar white matter was calculated for each side (Figure 1). Finally, the senior author supervised all measurements and ROI placements, and dubious cases were resolved by consensus.

Figure 1.

Measurements and positions of ROIs on an axial FIESTA MR image: focal hyperintensity in the dorsal brainstem (red arrow), size (purple line), cochlear aperture distance (green line), cochlea signal intensity (orange circle), vestibule signal intensity (blue circle) and cerebellar signal intensity (yellow circle). SIRc is indicated by orange/yellow circles; SIRv is indicated by blue/yellow circles. ROI: region of interest; FIESTA: fast imaging employing steady-state acquisition; MR: magnetic resonance; SIRc: signal intensity of the cochlea; SIRv: signal intensity of the vestibule.

Vestibulocochlear function tests

Video head impulse testing (vHIT) was performed by an experienced neuro-otologist using ICS Impulse video goggles and Vestlab 7.1 software (Otometrics) for data analysis, according to a previously described protocol.¹⁶ Briefly, the test consists of a system that senses head movement and recognises eye movement to quantify the vestibulo-ocular reflex (VOR). The examiner rotates the patient’s head several times on various planes to evaluate the lateral semi-circular, superior semi-circular and posterior semi-circular canals on both sides. VOR gain is expressed as a percentage. In order to compare vestibular function between patients, we considered the average VOR gain of the three semi-circular canals on the affected side, not just those that were abnormal.¹⁷

Hearing function was assessed by pure-tone audiogram and speech discrimination scores following the recommendations of the American Academy of Otolaryngology—Head and Neck Surgery for VS.¹⁸ Auditory brainstem response (ABR) testing was performed using Interacoustics EP15 Eclipse (Interacoustics A/S). Rarefaction polarity clicks between 1000 and 2000 are used as stimuli, and responses are recorded with surface electrodes according to the international 10–20 system.^19,20 For interpretation, we divided results into normal impulse transmission (no alteration of waves), altered transmission (alteration of amplitude or latency but waveform maintained) and lack of response (no discernible waves formed).

Statistical analyses

Statistical analyses were performed with Stata/SE v.14.1 software (StataCorp LP) using parametrical and non-parametrical tests. Descriptive statistics are presented as mean, median, minimum, maximum and standard deviation. The unpaired t-test and Mann–Whitney U-test were used to compare quantitative variables. Fisher’s exact test and the chi-square test were used for categorical variables. Statistical significance was defined as p ≤ 0.05.

Results

Of the 54 VS patients included in the study, 26 were men and 28 were women. The mean age was 50.5 ± 12.3 years (range 18–73 years). Forty-one patients were examined with a 1.5 Tesla scanner and 13 with a 3.0 Tesla scanner.

Focal hyperintensity in the dorsal brainstem

Focal HDB was found in 12/54 (22%) patients (10 patients, 1.5 Tesla; two patients, 3.0 Tesla). Tumours associated with HDB had a significantly larger mean diameter than those not associated with HDB (28 ± 7.4 mm vs. 17.4 ± 8.5 mm; p < 0.001). Cystic morphology (p = 0.004) and a greater distance from the tumour to the cochlear aperture (p = 0.039) also showed significant association with the occurrence of HDB (Table 1). Other variables, including age, sex and laterality of tumour, were not associated with HDB.

Table 1.

Tumour characteristics and HDB.

	Classification	Mean	HDB		p
			No (N = 42)	Yes (N = 12)
Size (mm)		19.9 ± 9.5 (3–45)	17.4 ± 8.5 (3–45)	28.8 ± 7.4 (18–45)	<0.001
	Homogeneous	22 (40.7)	22 (52.4)	0 (0)
Tumour morphology	Heterogeneous	15 (27.8)	10 (23.8)	5 (41.7)
	Cystic	17 (31.5)	10 (23.8)	7 (58.3)	0.004
Cochlear aperture distance (mm)		2.5 ± 2.2 (0–7.8)	2.1 ± 2.1 (0–7.8)	3.6 ± 2.5 (0–7.5)	0.039

HDB: hyperintensity in the dorsal brainstem.

Inner-ear MRI signal intensity

FLAIR-SIRc was significantly higher in patients with HDB (p = 0.014), and this difference persisted in subgroup analyses: patients evaluated with a 1.5 Tesla scanner with 3D-FLAIR sequences (p = 0.026); patients with Hannover T3 and T4 tumours (p = 0.016); and patients with Hannover T3/T4 tumours with cystic or heterogeneous components (p = 0.017; Table 2 and Figure 2). There was no association between HDB and FLAIR-SIRv, FIESTA/CUBE-SIRc or SIRv.

Table 2.

Inner-ear signal intensity ratios and HDB.

	HDB	n	Mean	Median	Minimum	Maximum	SD	p
FIESTA/CUBE-SIRc	No	42	4.18	3.97	2.53	7.77	0.92
	Yes	12	4.19	4.21	3.13	6.63	0.93	0.981
FIESTA/CUBE-SIRv	No	42	5.21	4.86	2.97	11.92	1.43
	Yes	12	5.30	5.21	4.10	7.66	0.94	0.837
FLAIR-SIRc	No	41	0.81	0.81	0.27	1.63	0.33
	Yes	11	1.11	0.99	0.65	2.08	0.41	0.014
FLAIR-SIRv	No	41	0.56	0.51	0.12	1.42	0.32
	Yes	11	0.77	0.66	0.23	1.39	0.38	0.070

Two patients were excluded from FLAIR analyses because of inadequate examinations.

FIESTA/CUBE: fast imaging employing steady-state acquisition/fast spin-echo imaging with variable flip angles; FLAIR: fluid-attenuated inversion recovery; SIRc: signal intensity ratio of the cochlea; SIRv: signal intensity ratio of the vestibule.

Figure 2.

Positions of ROIs, inner-ear intensity signal and HDB. (a) Cochlea signal intensity (orange circle), vestibule signal intensity (blue circle) and cerebellar signal intensity (yellow circle) on an axial FLAIR MR image. (b) Axial FIESTA image demonstrating linear HDB ipsilateral to the tumour. (c) An oblique sagittal reformatted FLAIR image with maximum intensity projection technique; signal intensity of the labyrinth is higher ipsilateral to the tumour compared with the contralateral side (d). HDB: hyperintensity in the dorsal brainstem; FLAIR: fluid-attenuated inversion recovery.

Figure 3.

HDB morphology. Axial FIESTA MR image with maximum intensity projection technique clearly shows the linear shape and obliquely oriented direction of the HDB.

Vestibulocochlear function tests

We found no statistically significant association between HDB and ABR, audiometry or vHIT except for a marginally lower median VOR value when only patients with Hannover T3/T4 tumours with cystic or heterogeneous components were analysed (0.66 ± 0.17; p = 0.049; Tables 3 and 4). Tumours with higher FLAIR-SIRc had a significantly worse ABR (lack of response, p = 0.026). A similar association was found between FLAIR-SIRv and ABR (p = 0.035; Table 5).

Table 3.

Audiometry and HDB.

	HDB	n	Mean	Median	Minimum	Maximum	SD	p
PTA (dB)	No	42	48.2	41	10	120	27.6
	Yes	11	45.8	44	22	71	16.7	0.790
SDS (%)	No	42	67.7	80	0	100	33.0
	Yes	12	71.2	85	0	100	32.3	0.749

One patient was excluded from PTA analysis for inadequate examination.

PTA: pure-tone audiometry; SDS: speech discrimination scores.

Table 4.

vHIT in HDB versus large (Hannover T3/T4), cystic/heterogeneous cases.

	HDB	n	Mean	Median	Minimum	Maximum	SD	p
Mean VOR	No	13	0.80	0.83	0.45	1.06	0.16
	Yes	12	0.66	0.70	0.40	0.86	0.17	0.049
Number of SC	No	13	0.85	1	0	2	0.80
	Yes	12	1.58	1.5	0	3	1.08	0.098

VOR: vestibular ocular reflex; SC: semi-circular canals.

Table 5.

Labyrinth signal intensity ratio and ABR.

	ABR	n	Mean	Median	Minimum	Maximum	SD	p
FIESTA/CUBE SIRc	LOR	15	4.24	3.88	3.22	6.87	1.07
	Others	27	4.16	3.98	3.17	5.45	0.57	0.790
FIESTA/CUBE SIRv	LOR	15	5.65	4.92	3.99	11.92	1.95
	Others	27	5.03	4.91	4.14	6.14	0.58	0.246
FLAIR SIRc	LOR	13	1.04	0.99	0.43	2.08	0.42
	Others	27	0.73	0.73	0.27	1.28	0.25	0.026
FLAIR SIRv	LOR	13	0.73	0.70	0.23	1.42	0.38
	Others	27	0.47	0.46	0.17	1.06	0.20	0.035

Only patients who performed 3D-FLAIR were used for the analyses.

ABR: auditory brainstem response; LOR: lack of response; Others: increased, normal, only V recognised, increased I latency.

Discussion

Hyperintense signal abnormality in the pons of patients with VS was first described by Okamoto et al. as a finding specific for VS.⁶ However, further studies have also shown this finding in other lesions of the cerebellopontine angle.⁷ The main hypothesis underlying this imaging finding is degeneration of the vestibular nuclei. However, no studies supporting this theory have been published. Vestibulocochlear nerve and nuclei degeneration have been described in VS in anatomical and histopathological studies,^21–23 but its clinical relevance is not well understood, and the presence of symptoms has been difficult to separate from the underlying disease. One previous study reported a correlation between HDB and decreased hearing ability (pure-tone audiometry) and vestibular function (caloric testing). However, these correlations did not reach statistical significance in multivariate analyses.⁷ In our study, we did not find a direct association between HDB and vestibulocochlear function tests, but we did find association between HDB and signal intensity of the labyrinth and other variables.

Tumour size is the most consistent variable associated with HDB in the literature, as demonstrated by Yamamoto et al.⁷ and Okamoto et al.,⁶ possibly reflecting the role of nerve compression in its pathophysiology. We also found a significant correlation between size of VS and presence of HDB, and concur with the hypothesis that chronic compression of the nerve is the most logical cause. The presence of a cystic/heterogeneous component in the tumour, which is only present in large (Hannover T3/T4) VS, has been previously reported by Okamoto et al.⁶ to be associated with HDB. This study found the same, confirming their finding. Cystic degeneration is a known risk factor for aggressive tumour behaviour and has also been associated with increased vascularity,^3,24 which may lead to more adhesion to the brainstem and other neural and vascular structures. The pathophysiology of cystic degeneration involves recurrent haemorrhages and subsequent inflammatory changes, which may account for the rapid nerve degeneration that has been observed in vestibular function tests¹⁵ as well as the difference in VOR values of patients with HDB observed in this study. In addition, the association between tumour distance from the cochlear aperture and HDB found in our study has not been previously described but has been associated with larger tumours and cystic degeneration.^25,26

Inner-ear fluid FLAIR signal hyperintensity has been described in patients with VS using both 2D⁸ and 3D techniques.^12,13 This alteration is thought to be due to increased protein content of the labyrinth, which is a known clinical finding in VS, as labyrinthine taps used to be performed for diagnosis.^27,28 The literature regarding the significance of hyperintense inner-ear fluid signal has been contradictory. Although the majority of studies have found no clear associations, Kim et al. found that intrameatal tumours – namely, small ones – had significantly lower inner-ear fluid signal intensity on FLAIR sequences compared to tumours with extrameatal extension and that inner-ear fluid signal intensity in intrameatal lesions moderately correlated to hearing impairment as measured by pure-tone audiometry. However, the correlation was not significant.¹³ We found that worse ABR results were associated with higher FLAIR-SIRc and FLAIR-SIRv. In addition, we found that a higher FLAIR-SIRc was associated with the presence of HDB, which supports the hypothesis that HDB could be an anatomical and imaging manifestation of degeneration of the auditory pathway. Although vestibular and auditory systems anatomy and functions are closely related and share the same cranial nerve (vestibulocochlear), there are separated brainstem nuclear complexes for each one (four vestibular and two cochlear nuclei). To our knowledge, no other studies have addressed the relationship between inner-ear signal intensity and HDB.

Finally, we found that HDB lesions had similar morphology and size, with several cases demonstrating a linear configuration, obliquely oriented in the sagittal plane (Figures 2(b) and 3). This was also observed by Yamamoto et al.,⁷ who attributed it to degeneration of the vestibular nerve itself. On the basis of our functional and imaging results, we hypothesise that HDB could represent degeneration of the cochlear nuclei. When confronting these findings with histological and neurosurgical anatomical studies of the cochlear nucleus complex, many morphological and topographical similarities arise. The cochlear nucleus complex is located laterally to the inferior cerebellar peduncles in the pontomedullary junction at the level of the entrance of the vestibulocochlear nerves. It is formed by the ventral and dorsal cochlear nuclei, which together have a distorted X-like shape, but separated have linear configurations.^29–31 Because the axis of the ventral nucleus is obliquely oriented to the sagittal plane and is normally few millimetres from the posterior surface of the brainstem, we believe that it could correspond to the same location where HDB were constantly observed in our sample.

Considering that HDB is associated with a higher SIRc, which in turn is associated with degeneration of the auditory pathway, as shown by worse ABR results, we hypothesise that the higher protein concentration in perilymph seen in VS could lead to cochlear damage and contribute to cochlear nuclei degeneration. In fact, a recent laboratory study has shown that secreted factors from VS can cause variable types of direct cochlear damage, including loss of hair cells and neuritis.³² Further studies are currently underway to test this hypothesis further, and the increasing clinical use of 7 Tesla MRI scanners seems promising to delineate brainstem anatomy better, particularly of small structures.^33,34

Our study has several limitations. The use of different scanners with different magnetic fields and variable techniques may have increased the heterogeneity of the sample population data. However, our findings persisted after multiple subgroup analyses. In addition, to measure inner-ear signal intensity, we manually placed small ROIs within its limits, which is prone to manual errors, since inner-ear structures can be difficult to individualise on FLAIR sequences. To compensate for this, we used FIESTA/CUBE sequences as a reference to place the ROIs.

In conclusion, focal HDB was associated with large and cystic VS, as well as tumours farther from the cochlear aperture. Higher FLAIR-SIRc was also associated with HDB, but not directly to vestibulocochlear function tests. Further studies are necessary to delineate its exact location better and to understand its clinical value.