| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Head and Neck Imaging |
1 From the Department of Radiology, Brooke Army Medical Center, San Antonio, Tex (C.J.B.); Dr Noah Weg & Associates, Suffern, NY (N.W.); and Departments of Otolaryngology-Head and Neck Surgery (L.B.M., S.J.Z.) and Radiology (S.J.Z.), Johns Hopkins University School of Medicine, 601 N Caroline St, Rm 6253, Baltimore, MD 21287-0910. From the 1998 RSNA scientific assembly. Received May 4, 2001; revision requested July 9; final revision received July 16, 2002; accepted August 9. Address correspondence to L.B.M. (e-mail: lminor@jhmi.edu).
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTS DISCUSSION REFERENCES |
|---|
MATERIALS AND METHODS: Temporal bone CT scans with 1.0-mm and/or 0.5-mm collimation were obtained in 50 patients with sound- and/or pressure-induced vestibular symptoms. The control population consisted of 50 patients undergoing CT at 1.0-mm collimation and 57 patients undergoing CT at 0.5-mm collimation for other reasons.
RESULTS: SSC dehiscence was documented on CT scans in all 36 patients with the clinical syndrome, with bilateral findings in six patients. Six other patients without specific clinical signs appeared to have dehiscence on 1.0-mm-collimated scans. Intact bone overlaying the SSC was subsequently identified with 0.5-mm-collimated CT in each case. On the 1.0-mm-collimated scans in 50 control patients, an area judged as possible or definite dehiscence was identified in 18 of 100 ears. The bone overlaying the SSC was intact in each of the 114 control ears evaluated with 0.5-mm-collimated CT. CT findings from the patients with vestibular symptoms combined with those in the control population indicated that the positive predictive value of an apparent dehiscence in the diagnosis of SSC dehiscence syndrome improved from 50% with 1.0-mm-collimated CT with transverse and coronal images to 93% with 0.5-mm-collimated CT with reformation in the plane of the SSC.
CONCLUSION: The positive predictive value of CT in identification of SSC dehiscence syndrome improves with 0.5-mm-collimated helical CT and reformation in the SSC plane.
© RSNA, 2003
Index terms: Computed tomography (CT), helical, 2131.12115 • Computed tomography (CT), thin-section, 2131.12118 • Ear, abnormalities, 2131.218 • Temporal bone, abnormalities, 2131.218 • Temporal bone, CT, 2131.12115, 2131.12118
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS DISCUSSION REFERENCES |
|---|
The vertical-torsional eye movements that align with the plane of the SSC and are evoked by sound and/or pressure stimuli establish a definitive standard for making the diagnosis of SSC dehiscence syndrome (3–7). The sensitivity and specificity of findings in other diagnostic tests, including imaging findings, can be evaluated relative to these vestibulo-ocular findings.
The dehiscence in the SSC creates a "third mobile window" into the inner ear (in addition to the oval and round windows). As a consequence of the dehiscence, motion of endolymph (the fluid within the membranous semicircular canal) is induced by the sound and pressure stimuli. Eye movements in the plane of the SSC result from this motion of endolymph within the dehiscent canal (1,2,5).
A dehiscence in the bone overlaying the SSC of the affected ear has been identified on temporal bone computed tomographic (CT) scans in these patients (1,3,7,8). Imaging has an important role in the diagnosis of this syndrome, particularly when surgical correction of the abnormality is contemplated. The purpose of this study was to describe the CT findings at different collimation widths associated with the SSC dehiscence syndrome and to determine the frequency of these findings in a control population.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS DISCUSSION REFERENCES |
|---|
In the present study, CT was performed for accepted clinical indications for all patients. This study was a review of existing clinical data with patient identifiers removed. It qualified for exemption from an institutional review board protocol based on Department of Health and Human Services criteria 45 CFR 46.101(b)(4). The determination that the study was exempt from a protocol requirement was made by the Joint Committee on Clinical Investigation of the Johns Hopkins University School of Medicine.
In the patients with symptoms of sound- and/or pressure-induced vestibular symptoms, temporal bone CT was a part of the clinical evaluation of their symptoms. Eight of the 14 patients who had sound- and/or pressure-induced vestibular symptoms but who did not have SSC dehiscence syndrome underwent standard 1.0-mm-collimated CT of the temporal bone and then, at a later date, 0.5-mm-collimated CT. The thin-section scans were obtained in these patients because of inconclusive findings on the 1.0-mm-collimated scans. Temporal bone CT scans with 0.5-mm but not 1.0-mm collimation were obtained in six of these 14 patients. For the 36 patients with SSC dehiscence syndrome, 13 underwent standard 1.0-mm-collimated CT alone, 11 underwent 0.5-mm-collimated CT alone, and 12 underwent 1.0-mm-collimated CT followed by 0.5-mm-collimated CT. In the patients who underwent CT with both collimation parameters, the thin-section scans were needed for more detailed visualization of the dehiscence and the surrounding structures, typically to prepare for a surgical procedure.
For the scans that served as controls in this study, 50 consecutive patients (100 ears) underwent standard 1.0-mm-collimated CT of the temporal bone at Johns Hopkins Hospital, and 57 consecutive patients (114 ears) underwent 0.5-mm-collimated CT of the temporal bone at the facility of Dr Noah Weg & Associates. CT in these patients had been performed for evaluation of otologic disorders that did not involve sound- or pressure-induced symptoms and also did not involve bone erosions.
The data gathered in this study included age and sex of all patients whose CT scans were analyzed. True-positive scans were those that demonstrated dehiscence in an ear from a patient with the diagnosis of SSC dehiscence syndrome that was established by the characteristic evoked eye movements. False-positive scans were those that appeared to demonstrate dehiscence radiographically, but typical eye movements could not be evoked by sound or pressure stimuli.
CT Scanning
Standard temporal bone CT was performed with either a Somatom Plus (Siemens Medical Systems, Erlangen, Germany) or a model 9800 (GE Medical Systems, Milwaukee, Wis) CT scanner with use of the following parameters: 1.0-mm collimation for scans in the transverse and coronal planes, 1.0-mm table increment, 330 mAs, 120 kVp.
Thin-section helical CT of the temporal bones was performed with a MxTwin CT scanner (Philips Medical Systems, Best, the Netherlands) in 37 of the 50 patients. This scanner became available after the initial description of SSC dehiscence syndrome (1). It is capable of scanning both single and helical sections with 0.5-mm thin collimation of the x-ray beam. In its Ultra High Resolution scanning mode, the system produces images with 20 line pairs per centimeter at cutoff, in-plane resolution.
All helical 0.5-mm-collimated scans were obtained in the transverse plane by using the Ultra High Resolution mode with a pitch of 0.7, resulting in a section thickness of 0.55 mm at full width at half maximum. A volume between 3.5 and 4.2 cm in its vertical height was acquired through the temporal bones with a 250-mm scan circle. The acquisition requires 50–60-second scanning time. Reformations from these helical data have a resolution of 18 line pairs per centimeter perpendicular to the scan plane (the z axis). The voxels produced with these techniques are practically isotropic (0.5 x 0.5 x 0.55 mm), allowing reformations in any plane from this volumetric data set with virtually no loss in resolution.
A full series of data was reconstructed separately over each temporal bone at 0.2-mm increments, yielding approximately three sections per collimation width. This technique provides between 175 and 210 transverse images through each temporal bone. The images were transferred to a MxView workstation (Philips Medical Systems) that is based on a Silicon Graphics O2 platform. Individual images were zoomed x6 (x1.96 in reconstruction from the raw data with the balance of the enlargement in the display), providing a 4-cm field of view. Twelve coronal reformations were acquired, spaced 0.1 mm apart, and rotated 40°–50° around a vertical axis to be parallel to the SSC. Slight angulation in the sagittal plane was required to display the entire ring of the SSC in the plane of the images. The reformation plane was then rotated 90° so that the dome of the SSC was seen in cross section. Fifteen sections were spaced 0.6 mm apart, which encompassed the entirety of the SSC, including both the anterior crus and the posterior or common crus.
Image Evaluation
For standard 1.0-mm-collimated CT scans, the bone covering the SSC was evaluated independently by two neuroradiologists (C.J.B., S.J.Z.) on the basis of the image on which the bone appeared thinnest, typically the coronal view. A three-point rating system was used: clearly intact, possible dehiscence, or definite dehiscence (Fig 1). Ratings were also reported as an average of the two readers.
|
|
|
Statistical Analyses
Analysis of interrater agreement for the control scans obtained with 1.0-mm collimation was performed with the 
statistic. The value of the statistic was interpreted according to established guidelines (11): 0.76 or higher, excellent agreement relative to chance; 0.41–0.75, good agreement; 0.40 or less, poor agreement. For the diagnosis of SSC dehiscence syndrome, the reference standard was the observation of vertical-torsional eye movements in the plane of the affected SSC that were evoked by sound and/or pressure stimuli. Sensitivity, specificity, and predictive values were calculated from the findings on images in reference to this physiologic standard.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS DISCUSSION REFERENCES |
|---|
|
|
|
All three layers of bone that can overlay the SSC were absent in the case of dehiscence. Figure 3 shows representative sections of temporal bone CT performed with 0.5-mm collimation in the coronal and oblique coronal planes in a patient with SSC dehiscence syndrome affecting the left ear. Figure 3c and 3d shows the dehiscence of the left SSC. An intact layer of bone covering the right SSC is noted in Figure 3a and 3b.
|
|
|
|
Among the 14 patients with symptoms of sound- and/or pressure-induced vertigo but without the characteristic eye movements indicative of SSC dehiscence syndrome, an area suspicious for dehiscence overlaying the SSC on 1.0-mm-collimated images was identified in six patients. This apparent area of dehiscence was unilateral in five patients and bilateral in one patient. CT with 0.5-mm collimation was performed in each of these 14 patients, and an intact layer of bone overlaying the SSC was identified in every case.
Of the 36 patients in whom SSC dehiscence syndrome was diagnosed on the basis of the characteristic evoked eye movements, there were three patients with eye movements evoked from stimuli in only one ear who also had apparent dehiscence on 1.0-mm-collimated scans of the contralateral temporal bone. Scans obtained with 0.5-mm collimation demonstrated an intact layer of bone overlaying this contralateral SSC and dehiscence overlaying the SSC on the symptomatic side in each case.
SSC dehiscence syndrome was diagnosed on the basis of vertical-torsional eye movements in the plane of the affected SSC that were evoked by sound and/or pressure stimuli in 23 of the 24 patients for whom the 0.5-mm-collimated CT scan indicated dehiscence. For these 23 patients, the CT findings indicated unilateral dehiscence in 19 and bilateral dehiscence in four. These findings were consistent with eye movements evoked by sound and pressure stimuli in all but one case. One patient had symptoms and signs of left SSC dehiscence syndrome but did not have these symptoms and signs associated with the right ear. The 0.5-mm-collimated CT scan indicated bilateral dehiscence. Also, one patient had dehiscence of bone overlaying the left SSC on the CT scan but lacked the characteristic eye movements evoked by sound and pressure stimuli. The patient had previously experienced vertigo and oscillopsia induced by loud noises in the left ear, but these symptoms had resolved 1 month before consultation.
Table 1 gives the sensitivity, specificity, and positive and negative predictive values of 1.0- and 0.5-mm-collimated CT performed in these patients with symptoms of sound- or pressure-induced vertigo and oscillopsia. The numbers in Table 1 reflect data from each ear.
|
|
|
= 0.525).
|
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS DISCUSSION REFERENCES |
|---|
The Tullio phenomenon has been described in many different conditions affecting the middle and inner portions of the ear, including syphilis (16,17), congenital deafness (18), Ménière disease (19,20), perilymph fistulas (21,22), trauma (19,23), and Lyme disease (24). Patients with some of these same abnormalities have also been reported to have vestibular symptoms and possibly eye movements evoked by stimuli that result in changes in middle ear or intracranial pressure (Hennebert sign) (25).
In addition to the disorders that can lead to signs such as the Tullio phenomenon and Hennebert sign, symptoms of sound- and/or pressure-induced vertigo and oscillopsia can be seen in a variety of disorders affecting the vestibular system, including autoimmune inner ear disease (26), perilymphatic fistula (27), and vestibular migraine (28), in addition to SSC dehiscence syndrome. Patients with a dehiscence of the bone overlaying the SSC have been shown to develop vertical-torsional eye movements that align with the plane of the affected SSC in response to loud noises or stimuli that result in changes in middle ear or intracranial pressure (1,2,5). These evoked eye movements are highly specific for this disorder (3,5–7). Surgical exploration of the roof of the SSC through a middle cranial fossa approach has confirmed the presence of dehiscence overlaying this canal in patients with this syndrome. Plugging or resurfacing of the SSC has led to resolution or reduction in signs and symptoms (1–3). Since the initial description of this syndrome, patients with this disorder have been described in reports from six other centers (3,6–8,29,30).
The results of our study show that standard temporal bone CT scans obtained with 1.0-mm collimation and viewed as transverse and coronal images have high sensitivity but are lacking in specificity in that thin bone may appear to be dehiscent. These findings provide an explanation for an observation by Watson et al (30). These investigators noted that two of the three unaffected ears in their four patients with SSC dehiscence syndrome appeared to have dehiscence of bone overlaying the SSC on temporal bone CT scans with 1.5-mm collimation. It is likely that these "apparent" dehiscences in ears without the symptoms and signs that are characteristic for SSC dehiscence syndrome are due to partial volume averaging associated with 1.5-mm-collimated scans. The findings of Watson et al (30) and the results of the present study clearly indicate that transverse and coronal images obtained from temporal bone CT examinations performed with 1.0- or 1.5-mm collimation should not be used as the sole criterion for establishing a diagnosis of SSC dehiscence syndrome.
Temporal bone CT scans with 0.5-mm collimation and reformatting of data into the plane of the SSC improve the specificity and positive predictive value of a finding of dehiscence. These scans are particularly valuable when surgical correction of the abnormality is being contemplated. There will undoubtedly be cases in which the bone will be so thin that a determination of whether or not the roof of the SSC is intact will not be possible, even on scans obtained with 0.5-mm collimation. We encountered two patients in whom this may have been the case. In one patient, there were symptoms and signs that established a diagnosis of left SSC dehiscence syndrome but none associated with the right ear. The 0.5-mm-collimated CT scan revealed bilateral dehiscence. A second patient had previously experienced symptoms that could have been associated with left SSC dehiscence syndrome, but these symptoms had resolved by the time the patient was referred for consultation. On neuro-otologic examination, the patient lacked the evoked eye movements that are characteristic for SSC dehiscence syndrome. In each patient, an intact layer of bone measuring less than 0.1 mm may have appeared as dehiscent on the scan. Alternatively, dehiscence may have existed but was not causing symptoms and signs because the canal became functionally inactivated either through fibrosis or the effects of the overlaying dura compressing the membranous semicircular canal.
The diagnosis of SSC dehiscence syndrome is based on characteristic symptoms and the finding of vertical-torsional eye movements evoked by sound or pressure stimuli. A finding of a low threshold for vestibular myogenic potentials in patients with SSC dehiscence syndrome may also be helpful in making the diagnosis (3,30,31). Thus, in the uncommon situation of 0.5-mm-collimated scans showing an apparent dehiscence that does not appear to be associated with the patient’s symptoms, the diagnosis will be apparent from the clinical findings in the patient.
The treatment for patients with this syndrome varies according to the character and severity of symptoms. Uncertainty about the cause of these symptoms, some of which may seem rather bizarre, has been one of the more troublesome features for many patients. Once the diagnosis and cause of the symptoms have been explained, many patients have found that avoidance of the provocative stimuli provides adequate treatment. For others, the persistent dysequilibrium is disabling, even when sound- and/or pressure-evoked symptoms are not occurring. Surgical plugging or resurfacing of the dehiscent canal has been shown to be beneficial in these cases (2).
Identification of SSC Dehiscence on Temporal Bone CT Scans
Isotropic CT facilitates viewing complex anatomic structures in the optimal plane. Kalender (32) described the value of this technique for thin-section CT. Venema et al (33) reported CT of the temporal bone with use of 0.5-mm collimation, but their study was limited to images in the coronal plane.
The following considerations explain the improved visualization of these thin bones on 0.5-mm-collimated CT scans. Direct coronal 1.0-mm-collimated CT performed with a system with high in-plane resolution (20 line pairs per centimeter) only depicts a small portion of the curved canal roof, its dome, perpendicular to the plane of the scan. The remainder of the canal roof is oblique to that plane. Since a voxel is the smallest volume that can be displayed by an imaging system, structures that fill part of a voxel are still displayed as the full size of a voxel. The attenuation of these small structures, however, is averaged with the attenuation of the other structures that contribute to that voxel, based exactly on the proportion of the voxel each tissue occupies. This is the well-known phenomenon of partial volume averaging. Since most of the roof of an SSC (except for its dome) is oblique to the coronal plane of section and therefore only fills parts of voxels, thin sections of a roof are subject to partial volume averaging over a voxel that is 1.0-mm in thickness. For a layer of bone with a thickness of 0.1 mm, such averaging results in a substantial decrease in its attenuation on the image (depending on the angle of the bone to the plane of the scan) and may lead to an impression that the bone is absent. This mechanism can explain the false-positive scans in our series of scans obtained with 1.0-mm collimation in the direct coronal plane. The increased attenuation of such thin bone by means of partial volume averaging still applies to scans obtained with 0.5-mm collimation, but its effect is halved.
Scanning with 0.5-mm collimation as we did for the present study yields the following attenuations. Otic capsule bone is extremely dense, measuring 1,800–2,000 HU (34,35). The surrounding fluid and brain tissues have attenuation measurements of 0–40 HU. The Ultra High Resolution scans are relatively noisy, with SDs for attenuation measurements varying between 40 and 80 HU. With these values, a thin layer of bone measuring 0.1 mm in thickness would occupy approximately one-fifth of a voxel, with the remainder of the voxel filled with fluid and adjacent tissues. This voxel would therefore measure 360–400 HU (1,800/5 to 2,000/5) ± 40–80 (noise). Such a voxel is readily visible against a background of 0–40 HU ± 40–80. Thus, bone layers with values as thin as these can be visualized.
The preceding partial volume calculations are predicated on the fact that even when the bone covering the SSC is thin, it is still extremely dense (approximately 2,000 HU), as is typical of otic capsule bone. This observation has been confirmed in a temporal bone histologic study (36).
Scanning in the transverse plane with 1.0-mm rather than 0.5-mm collimation and performing multiplanar reformations in the plane of the SSC decreases the z-axis resolution by 50%. With such a system, the thinnest measurements that could be made with confidence would be approximately double those noted above (ie, 0.2 mm). Since many of the roofs of the SSCs are thinner than this measurement, they could not be detected on reformations in the plane of the SSC derived from data sets scanned with 1.0-mm collimation and would appear as dehiscent. Images obtained with 1.0-mm collimation could not be analyzed when the data were reformatted into the plane of the SSC because the reformations were too blurry to be of value. We performed these reformations from the data of 1.0-mm-collimated scans with varying degrees of overlap, and the image quality was not sufficiently improved to permit analysis of the integrity of the roof of the SSC.
Temporal bone CT can demonstrate potentially treatable dehiscence in the bone overlaying an SSC in patients with vestibular symptoms and signs induced by sound and/or pressure. Standard 1.0- or 1.5-mm-collimated temporal bone CT has a high sensitivity for detection of bone dehiscence overlaying the SSC but relatively low specificity. The specificity of CT is improved when scanning with 0.5-mm collimation and reformation into the plane of the SSC are used.
|
FOOTNOTES |
|---|
Abbreviation: SSC = superior semicircular canal
Author contributions: Guarantors of integrity of entire study, C.J.B., N.W., L.B.M.; study concepts and design, all authors; literature research, L.B.M., N.W., C.J.B.; clinical studies, L.B.M., N.W.; data acquisition and analysis/interpretation, N.W., L.B.M., C.J.B.; statistical analysis, L.B.M., N.W., C.J.B.; manuscript preparation, C.J.B., L.B.M., N.W.; manuscript definition of intellectual content, editing, revision/review, and final version approval, all authors.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS DISCUSSION REFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  C E OFFIAH, R T RAMSDEN, and J E GILLESPIE Imaging appearances of unusual conditions of the middle and inner ear Br. J. Radiol., June 1, 2008; 81(966): 504 - 514. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S M Rosengren, S T Aw, G M Halmagyi, N P M. Todd, and J G Colebatch Ocular vestibular evoked myogenic potentials in superior canal dehiscence J. Neurol. Neurosurg. Psychiatry, May 1, 2008; 79(5): 559 - 568. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. S. Welgampola, O. A. Myrie, L. B. Minor, and J. P. Carey Vestibular-evoked myogenic potential thresholds normalize on plugging superior canal dehiscence Neurology, February 5, 2008; 70(6): 464 - 472. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. T. Aw, M. J. Todd, G. E. Aw, J. S. Magnussen, I. S. Curthoys, and G. M. Halmagyi Click-evoked vestibulo-ocular reflex: Stimulus-response properties in superior canal dehiscence Neurology, April 11, 2006; 66(7): 1079 - 1087. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 B. F. Branstetter IV, C. Harrigal, E. J. Escott, and B. E. Hirsch Superior Semicircular Canal Dehiscence: Oblique Reformatted CT Images for Diagnosis Radiology, March 1, 2006; 238(3): 938 - 942. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. I. Lane, E. P. Lindell, R. J. Witte, D. R. DeLone, and C. L. W. Driscoll Middle and Inner Ear: Improved Depiction with Multiplanar Reconstruction of Volumetric CT Data RadioGraphics, January 1, 2006; 26(1): 115 - 124. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
H. D. Curtin Superior Semicircular Canal Dehiscence Syndrome and Multi-Detector Row CT Radiology, February 1, 2003; 226(2): 312 - 314. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME |
