Brought to you by

Letter to the Editor: “Visualization of the saccule and utricle with non-contrast-enhanced FLAIR sequences”

by Shinji Naganawa, Michihiko Sone (naganawa@med.nagoya-u.ac.jp)

Visualization of the saccule and utricle with non-contrast-enhanced FLAIR sequences

Dear Editor-in-Chief,

We read with great interest the recently published article by Hikaru Fukutomi, et al. entitled “Visualization of the saccule and utricle with non contrast enhanced FLAIR sequences” [1]. We believe that the challenge of diagnosing Meniere’s disease with non-contrast enhanced MR imaging, if realized, would be of great value for patients.

There have been two methods proposed to increase the efficacy of 3D-FLAIR by introducing T2-selectivity to an inversion recovery pulse. One is a method using a T2-selective inversion recovery pulse [2], and the other similar method utilizes T2-preparation pulses before the application of the inversion recovery pulse [3]. The accuracy for the inversion of the spins is susceptible to local field inhomogeneity in both methods. For the 3D-FLAIR using a T2-selective inversion recovery pulse, it has been shown that the cerebrospinal fluid (CSF) near the sphenoidal ridge and the lateral part of the vestibular fluid, has high signal intensity due to an incomplete inversion of the spins [4]. We have also encountered cases that show a difference in the shape and size of the endolymphatic space visualized in the non-contrast enhanced 3D-FLAIR image with a T2-selective inversion pulse, which were visualized on the contrast enhanced conventional 3D-FLAIR. In a phantom experiment using test tubes, we found that although 3D-FLAIR imaging using T2-preparation pulses tends to suffer less artifacts than 3D-FLAIR using T2-selective inversion recovery, the fluid signal near air persists in the images using either method due to the incomplete signal suppression. Based on these experiences, we assume that the presumed perilymph signal shown in this paper would be the incompletely suppressed fluid signal due to the field inhomogeneity. We have a few specific concerns regarding this paper as stated below.

The authors’ method relies on the hypothesis that the perilymph and endolymph fluids have similar T1 values and different T2 values. They seem to rely on the protein concentration difference between the perilymph and endolymph. The r1 and r2 of albumin has been reported to be 0.010 and 0.066 (g/dl)-1s-1. The r2/r1 is 6.6. If the authors assume T2 shortening of the perilymph, the T1 of the perilymph should also shorten to some extent, which can be detected by conventional heavily T2-weighted 3D-FLAIR. The authors might clarify this point.

In Figure 3b using T2 preparation pulses, the lateral wall of the vestibule is convex towards the inside of vestibule at TI = 224 ms. The vestibular high signal area in the image at TI = 224 ms, is far larger than the presumed vestibular endolymph area in the image at TI = 2000 ms. Alternatively, the lateral wall of the vestibule is convex towards the outside of vestibule at TI = 5000 ms. In both images, both the endolymph and most of the perilymph in the cochlea and vestibule show a high signal intensity under these conditions. The perilymph fluid in the lateral part of the vestibule has a low signal intensity in the image at TI = 224 ms. It is unreasonable to assume that only the perilymph in the lateral part of the vestibule has a shorter T2 due to a higher protein concentration.

In Fig. 4a, the band-like high signal intensity along the lateral wall of the vestibule is shown, but not on the medial side of the vestibule and the cochlea. We do not know a theoretical basis for this other than artifacts due to incomplete inversion in the lateral side of the vestibule. In Fig. 4c, the CSF adjacent to the wall of the internal auditory canal and the sphenoidal ridge shows a linear-shaped higher signal intensity than the CSF, apart from the bony wall. It is unlikely that the CSF in these regions close to the bony wall has a higher protein concentration than in other regions.

In Fig 4d, the authors employed a T2-preparation pulses with the preparation time of 125 ms for the conventional injected 3D-FLAIR images according to Table 2. This is not appropriate for the comparison. In our volunteer study, the lateral part of the vestibule had a higher signal intensity even without contrast injection in the images with a T2-prep time of 125ms. Also, in the images in Fig. 4d, the lateral part of the vestibule has a higher signal than the presumed perilymph in the medial part of the vestibule both in the non-contrast and the contrast-enhanced images.

On the hydrops side (Fig. 4d), the shape of the endolymph in the cochlea of the right ear is completely different between the non-contrast and the contrast-enhanced images. In the cochlea, the contrast-enhanced perilymph of the scala vestibuli is not visible in the image, but in the non-contrast image, the scala vestibuli is clearly visualized with a high signal intensity. For the vestibule, in the contrast-enhanced image of the right ear, the low signal intensity of the endolymphatic space is touching the lateral wall of the vestibule, however in the optimized non-contrast image, the presumed endolymphatic space does not touch the lateral wall of the vestibule. On the normal side, the saccule is directly adjacent to the medial side wall of the vestibule in the contrast-enhanced image, however the saccule does not touch the medial side wall of the vestibule in the non-contrast image (Fig. 4d).

For the reasons mentioned above, there is still insufficient evidence to conclude that the endolymph and the perilymph can be separated in non-contrast enhanced 3D-FLAIR images using T2-preparation pulses, since the fluid near the bony wall is prone to incomplete inversion. A detailed correlation with the contrast enhanced 3D-FLAIR without T2-preparation pulses in a large number of cases is desirable in the future.