Otology & Neurotology 35:e337Ye342 Ó 2014, Otology & Neurotology, Inc.
Identification of Endolymphatic Hydrops in Me´nie`re’s Disease Utilizing Delayed Postcontrast 3D FLAIR and Fused 3D FLAIR and CISS Color Maps *Mari Hagiwara, †J. Thomas Roland Jr, *Xin Wu, *Annette Nusbaum, *James S. Babb, ‡Pamela C. Roehm, †Paul Hammerschlag, §Anil K. Lalwani, and *Girish Fatterpekar *Departments of Radiology and ÞOtolaryngology, New York University School of Medicine, New York, New York, U.S.A.; þDepartment of Otolaryngology, Temple University, Philadelphia, Pennsylvania, and §Department of Otolaryngology, Columbia University, New York, New York, U.S.A.
Objective: The preferential delayed enhancement of the perilymphatic space enables detection of the non-enhancing endolymphatic hydrops present in patients with Me´nie`re’s disease. The aim of this study was to evaluate the diagnostic utility of delayed postcontrast 3D FLAIR images and a color map of fused postcontrast FLAIR and constructive interference steady state (CISS) images in the identification of endolymphatic hydrops in patients with clinically diagnosed Me´nie`re’s disease. Study Design: Case control, blinded study. Setting: Tertiary referral center. Patients: Ten patients with Me´nie`re’s disease and five volunteer controls. Intervention: Diagnostic. Main Outcome Measure: Two neuroradiologists blinded to the clinical history independently evaluated for the presence of endolymphatic hydrops on the images of both inner ears for test and control subjects. Both the standard gray-scale FLAIR images and the fused color map images were independently reviewed.
Results: The gray-scale 3D FLAIR images demonstrated 68.2% sensitivity and 97.4% specificity, and the fused color map images demonstrated 85.0% sensitivity and 88.9% specificity in the identification of endolymphatic hydrops in Me´nie`re’s disease. There was significant correlation between the gray-scale 3D FLAIR images and fused color map images with the categorization of involvement (p = 0.002). Inter-evaluator reliability was excellent (kappa = 0.83 for gray-scale images, kappa = 0.81 for fused color map). Conclusion: Delayed 3D FLAIR and fused 3D FLAIR-CISS color map images of the inner ears after intravenous contrast administration are potentially useful diagnostic tools in the evaluation of patients with suspected Me´nie`re’s disease. Key Words: Delayed postcontrast FLAIRVEndolymphatic hydropsV Me´nie`re’s disease.
Me´nie`re’s disease is an idiopathic disorder of the inner ear characterized by recurrent episodic vertigo, fluctuating hearing loss, aural fullness, and tinnitus secondary to enlargement of the endolymphatic space (endolymphatic hydrops). The current accepted method of diagnosing
Me´nie`re’s disease with certainty requires a postmortem histopathologic confirmation of endolymphatic hydrops (1). Me´nie`re’s disease has therefore remained a clinical diagnosis, made with varying degrees of certainty based on the criteria reported by the Committee on Hearing and Equilibrium (1). Absence of a reliable diagnostic test has hampered not only the ability to definitively diagnose Me´nie`re’s disease but also to design and assess the effectiveness of treatment. However, imaging studies within the past decade have demonstrated visualization of endolymphatic hydrops in patients utilizing delayed 3D FLAIR imaging after the intratympanic and intravenous injections of gadolinium contrast (2,3). The delayed preferential enhancement of the perilymphatic space has enabled distinction between the
Otol Neurotol 35:e337Ye342, 2014.
Address correspondence and reprint requests to Mari Hagiwara, M.D., 660 1st Avenue, 2nd floor Radiology, NYU Langone Medical Center, New York, NY 10016, U.S.A.; E-mail: [email protected] The study was partly funded by an $8,000 seed grant (account number 51-C63370-43757-NYUPRJ) from the Department of Radiology at the NYU Langone Medical Center. There was no salary support from the seed grant. The authors disclose no conflicts of interest.
e337
Copyright © 2014 Otology & Neurotology, Inc. Unauthorized reproduction of this article is prohibited.
e338
M. HAGIWARA ET AL.
fluid-filled perilymphatic and endolymphatic spaces, and thereby has allowed for the detection of an enlarged nonenhanced endolymphatic space (4). The unenhanced endolymphatic space is difficult to distinguish from bone on the 3D FLAIR images. We have found that fusion of the 3D FLAIR sequence with the constructive interference steady state (CISS) sequence can help differentiate the unenhanced endolymphatic space from bone, and using a color map overlay in the areas of increased signal on the FLAIR sequence can increase the conspicuity of the abnormal endolymphatic space in patients with Me´nie`re’s disease. The purpose of this study was to evaluate the diagnostic utility of delayed postcontrast high-resolution FLAIR images and a color map of fused postcontrast FLAIR and CISS images in the identification of endolymphatic hydrops in patients with Me´nie`re’s disease. MATERIALS AND METHODS Study Subjects Ten patients with high clinical suspicion of Me´nie`re’s disease and five controls without Me´nie`re’s disease and with normal hearing were included in the study. Criteria for clinically suspected Me´nie`re’s disease were based on the guidelines established by the Committee on Hearing and Equilibrium in 1995; the criteria includes two or more definitive spontaneous episodes of vertigo lasting 20 minutes or longer, audiometrically documented lowfrequency hearing loss on at least one occasion, tinnitus and aural fullness in the affected ear, and exclusion of other causes (1). All patients had normal conventional MR imaging of the vestibulocochlear complexes and internal auditory canals that excluded an underlying lesion which could mimic Me´nie`re’s disease. None of the patients had a history or current symptomatology of migraine or vestibular migraine. All patients were undergoing medical treatment for Me´nie`re’s disease. Exclusion criteria were based on contraindications to MRI and intravenous contrast administration, including patient history of cardiac pacemaker, intracranial clips, pregnant or nursing, and renal disease. Both the patient population and the volunteers had normal renal function tests and did not suffer from any renal disease in the past. None of the study subjects had contraindications to MRI. We evaluated 20 inner ears (right and left) of the 10 patients and 10 inner ears of the five control asymptomatic volunteers. Bilateral symptoms were reported in one patient. Unilateral right-sided symptoms were reported in six patients, and unilateral left-sided symptoms were reported in three patients. The age range of the 10 patients with Me´nie`re’s disease was 31 to 72 years old, with five men and five women. The age range of the five volunteer controls was 27 to 30 years old; all five controls were men. All patients and control subjects provided institutional review boardYapproved written consent, and the study was Health Insurance Portability and Accountability Act compliant. On the day of imaging, all patients filled out a questionnaire specifying the side of involvement, symptomatology, date of the initial presentation, and date of the most recent episode.
MR Imaging All imaging was performed on a 3-Tesla MR imaging scanner (Tim-Trio; Siemens, Erlangen, Germany) using a 12-channel array head coil. All patients underwent MR scanning before and
after the intravenous administration of double-dose (0.2 mmol/kg body weight) gadolinium (gadopentetate dimeglumine, Magnevist; Bayer Health Care, Wayne, NJ). Four hours after the intravenous administration of contrast, a 3D FLAIR sequence and a CISS sequence of the inner ears were obtained. The parameters for the 3D FLAIR sequence were as follows: TR 9,000 ms, TE 315 ms, TI 2,500 ms, matrix 256 242, 16 1-mm-thick slices, FOV 17.5 16.4 cm, variable flip angle, echo train length 121, NEX 2.0, iPAT off, voxel size 0.7 0.7 1.0 mm, and acquisition time 12 minutes. The parameters for the CISS sequence were as follows: TR 9.12, TE 4.56, matrix 448 358, 32 0.6-mm-thick slices, FOV 14 14 cm, flip angle 50 degrees, echo train length 1, NEX 1, iPAT off, voxel size 0.4 0.3 0.6 mm, and acquisition time 4 minutes 30 seconds.
Image Analysis Fusion of the 3D FLAIR and CISS images and color overlay maps were generated using Sphere v2.2 post-processing software (Olea Medical, La Ciotat, France). The fused color maps were obtained in all patients and controls except for one patient due to technical error. Two neuroradiologists (M.H., 5 years’ experience and G.F., 12 years’ experience) blinded to clinical history independently reviewed the images of both (right and left) inner ears for test and control subjects. The readers subjectively determined the presence or absence of inner ear hydrops and specified whether the vestibule and/or cochlea were involved. Both the standard gray-scale FLAIR images and the fused color map images were separately reviewed in this manner, with randomization of the presentation order.
Gray-scale FLAIR Identification of an enlarged non-enhancing endolymphatic compartment seen as ‘‘black circles’’ in the cochlea was used to determine the presence of endolymphatic hydrops (Fig. 1). Nonvisualization of the vestibule was used to determine hydrops within the vestibule.
Fused Color Map Hyperintense (enhancing) signal on the FLAIR sequence was colorized, while the CISS sequence was left in gray scale. Identification of inward concavities seen as ‘‘bites’’ within the cochlea was used to designate the presence of hydrops within the cochlea
FIG. 1. Gray-scale 3D FLAIR. A, Identification of an enlarged non-enhancing endolymphatic compartment seen as ‘‘black circles’’ (solid arrows) in the cochlea was used to determine the presence of endolymphatic hydrops. Lack of visualization of the vestibule in its expected location (dashed arrow) was the determining factor used to determine hydrops within the vestibule. B, Lack of an enlarged discretely identifiable endolymphatic space in the cochlea (solid arrow) and visualization of the normal enhancing vestibule (dashed arrow) were used to determine the absence of endolymphatic hydrops.
Otology & Neurotology, Vol. 35, No. 10, 2014
Copyright © 2014 Otology & Neurotology, Inc. Unauthorized reproduction of this article is prohibited.
ENDOLYMPHATIC HYDROPS IN ME´NIE`RE_S DISEASE (Fig. 2). Lack of color (reflecting absence of enhancement on the FLAIR sequence) within the visualized vestibule was used to determine the presence of hydrops within the vestibule.
TABLE 1. The sensitivity and specificity achieved when ears were classified as affected (test positive for Me´nie`re’s disease) when a specific abnormal finding was detected Finding
Statistical Analysis Logistic regression for correlated data was used to assess the diagnostic utility of imaging findings for the detection of affected ears. Specifically, generalized estimating equations based on binary logistic regression was used to model the status of each ear (affected versus unaffected) as a function of reader and each imaging finding. Statistical dependencies among the multiple results recorded for each individual subject (person) were accounted for by assuming results to be correlated when derived for the same subject and independent when derived for different subjects. Interreader agreement was assessed in terms of simple kappa coefficients with kappa greater than 0.6 interpreted as substantial agreement (5). All statistical tests were conducted at the two-sided 5% significance level using SAS 9.3 (SAS Institute, Cary, NC).
RESULTS There was no significant difference ( p 9 0.4) between the unaffected ears of the Me´nie`re’s patients and the unaffected ears of controls in terms of the prevalence of any imaging finding. Because of lack of significant difference, these two groups were then combined in one unaffected ear category. Significant correlation ( p G 0.001) was demonstrated between categorization of involvement with both the grayscale FLAIR images and the fused color map images. For overall assessment of the presence of Me´nie`re’s disease, the gray-scale 3D FLAIR images demonstrated lower sensitivity (68.2% versus 85%) and greater specificity (97.4% versus 88.9%) compared with the fused FLAIR-CISS color maps (Table 1). For evaluation of the vestibule, the grayscale FLAIR images showed lower sensitivity (63.6% versus 75.0%) and slightly greater specificity (97.4% versus 94.4%) compared with the fused color maps. For evaluation of the cochlea, the gray-scale FLAIR images demonstrated lower sensitivity (68.2% versus 80.0%) and greater specificity (94.7% versus 77.8%) compared with the fused color maps.
e339
Overall gray scale Vestibule gray scale Cochlea gray scale Overall color map Vestibule color map Cochlea color map
Specificity 97.4% 97.4% 94.7% 88.9% 94.4% 77.8%
(37/38) (37/38) (36/38) (32/36) (34/36) (28/36)
Sensitivity
p
68.2% (15/22) 63.6% (14/22) 68.2% (15/22) 85.0% (17/20) 75.0% (15/20) 80.0% (16/20)
G0.001 0.003 G0.001 G0.001 G0.001 0.001
Numerators and denominators are counts of evaluations (summed over readers) rather than counts of ears or subjects; variation in the denominators within a single column is a result of missing data. p values are from the generalized estimating equations analysis to test whether the relevant finding was a predictor of whether or not an ear was affected. There was no set of two or more imaging measures representing independent predictors of whether or not ears were affected.
The kappa for inter-reader agreement was 0.83 for the gray-scale 3D FLAIR images and 0.81 for fused color map images, indicating substantial agreement between readers. More specifically, of the 20 inner ears of patients with Me´nie`re’s disease, there was in consensus between the two readers on the 3D FLAIR images: seven true-positives, eight true-negatives, and three false-negatives. There was one patient where there was discrepancy between the two readers for both the affected and unaffected inner ears. Of the 18 inner ears of patients with Me´nie`re’s disease, there was in consensus on the fused color maps: eight truepositives, seven true-negatives, one false-negative, and one false-positive (Fig. 3). There was discrepancy between the readers for one of the affected ears. Of the 10 inner ears of the five control patients, there were eight true-negatives in consensus on the 3D FLAIR and fused color images (Fig. 4). On the fused color maps, there was discrepancy between readers for one control subject, with both inner ears incorrectly determined as abnormal by one reader. All of the false-negative patients with Me´nie`re’s disease based on the gray-scale FLAIR images had their initial presentation within the past 2 years. All of the true-positive inner ears were in patients with the initial presentation greater than 3 years prior (range 3Y28 years). One false-positive on the color map was present in the contralateral ear of a patient with Me´nie`re’s disease with unilateral symptoms. DISCUSSION
FIG. 2. Fused color map. A, Identification of ‘‘bites’’ within the cochlea (solid arrows) was used to designate the presence of hydrops within the cochlea. Lack of color (reflecting lack of enhancement on the FLAIR sequence) within the visualized vestibule (dashed arrow) was used to determine the presence of hydrops within the vestibule. B, Lack of ‘‘bites’’ within the cochlea with normal convex margins (solid arrows) and visualization of color signal within the vestibule (dashed arrow) were used to determine absence of endolymphatic hydrops.
Me´nie`re’s disease is an idiopathic debilitating disorder of the inner ear characterized by recurrent episodic vertigo, hearing loss, aural fullness, and tinnitus. The currently accepted criteria to diagnose Me´nie`re’s disease with certainty require postmortem histopathologic evaluation of the temporal bones, and therefore Me´nie`re’s disease has largely remained a clinical diagnosis (1). Identification of endolymphatic hydrops on imaging can not only enable a definitive diagnosis of Me´nie`re’s disease in living patients but can also potentially help determine therapeutic strategies and evaluate treatment response, as well as provide Otology & Neurotology, Vol. 35, No. 10, 2014
Copyright © 2014 Otology & Neurotology, Inc. Unauthorized reproduction of this article is prohibited.
e340
M. HAGIWARA ET AL.
further understanding of the pathophysiology of the poorly understood disorder (6,7). Distinction of the fluid-filled endolymphatic and perilymphatic spaces has been limited on conventional imaging because of its small size and identical signal intensities. Within the past decade, research studies have investigated the use of delayed high-resolution imaging to visualize endolymphatic hydrops after the intratympanic and intravenous injections of gadolinium contrast (2). Preferential transfer of gadolinium into the perilymphatic space has enabled visual distinction of the endolymphatic and perilymphatic spaces, and thereby has allowed for the detection of an enlarged non-enhanced endolymphatic space (4,7). Intratympanic injection of contrast has been shown to provide better image contrast compared with intravenous injection because of a greater degree of perilymphatic enhancement (8,9). However, this method of contrast administration is more invasive, requires off-label use of the contrast material, requires 24-hour interval between the time of injection and of imaging, and enables evaluation of only one inner ear at a time (9,10). Intravenous injection of contrast is the standard use of contrast material, requires only a 4-hour waiting period, and enables evaluation of both inner ears at a time (9); intravenous contrast administration thus is the ideal method of choice in the evaluation of Me´nie`re’s disease. The delayed postcontrast MR imaging findings seen in patients with Me´nie`re’s disease are thought to correspond to an enlarged endolymphatic space. The ‘‘black circles’’ in the cochlea are thought to reflect the dilated unenhanced endolymphatic scala media compartment, which is barely perceptible in the normal patient. On the fused color map images, this enlarged unenhanced endolymphatic space resulted in the appearance of ‘‘bites’’ in the colorized cochlea. Non-visualization of the vestibule on the 3D FLAIR
FIG. 3. Meniere’s disease patient with right-sided symptoms. Gray-scale 3D FLAIR (A) and fused color map (B) demonstrates endolymphatic hydrops within the cochlea and vestibule on the right side and normal-appearing endolymphatic space on the left side. Note the enlarged rounded endolymphatic space on the gray-scale image and apparent ‘‘bites’’ on the color map within the right cochlea (solid arrows) compared with the normal left side. Also note the lack of enhancement on the gray-scale image and lack of color on the color map within the right vestibule (dashed arrows), compared with the normal left side.
FIG. 4. Control patient. Gray-scale 3D FLAIR (A) and fused color map (B) demonstrate normal-appearing cochleae (solid arrows) and vestibules (dashed arrows) bilaterally.
sequence has also been speculated to reflect marked enlargement of the endolymphatic space, which either completely replaces the fluid in the perilymphatic space and/or prevents the movement of gadolinium from entering the perilymphatic space (11). On the fused color map images, this was seen as lack of color within the vestibule, corresponding to the absence of gadolinium contrast. Our caseYcontrol, blinded study demonstrated good interobserver reliability as well as good specificity in the identification of Me´nie`re’s disease for the 3D FLAIR images after the intravenous administration of contrast. The gray-scale 3D FLAIR images demonstrated 68.2% sensitivity and 97.4% specificity in the identification of endolymphatic hydrops in Me´nie`re’s disease. In their study of 207 ears (31 with Me´nie`re’s disease symptoms), Fang et al. demonstrated greater sensitivity (97.5%) and specificity (98%) using gray-scale 3D FLAIR images after the intratympanic injection of contrast (12). The decreased sensitivity and specificity in our study may be related to smaller sample size and improved contrast resolution with intratympanic injections. Based on our literature search, this is the only study to evaluate the sensitivity and specificity of MRI after the intravenous administration of contrast. This is also the first study to use fused color maps of contrast-enhanced images and CISS images for the identification of endolymphatic hydrops. Using the fused color images, we noticed that the sensitivity to identify endolymphatic hydrops increased from 68.2% to 85.0%. We speculate this increased sensitivity to result from the ability of the fused color images to better distinguish the unenhanced endolymphatic space from the underlying bone. Previous studies have suggested that the unenhanced hypointense endolymphatic space is difficult to distinguish from the hypointense bone on the 3D FLAIR images (13,14). Naganawa et al. optimized the inversion time of a 3D inversion-recovery turbo spin-echo sequence so that the
Otology & Neurotology, Vol. 35, No. 10, 2014
Copyright © 2014 Otology & Neurotology, Inc. Unauthorized reproduction of this article is prohibited.
ENDOLYMPHATIC HYDROPS IN ME´NIE`RE_S DISEASE hypointense endolymphatic space was of a different signal (black) relative to the signal (gray) of the surrounding bone (14,15). We found that fusion of the 3D FLAIR sequence with the CISS sequence more clearly differentiates the unenhanced endolymphatic space from the bone, as the CISS sequence clearly distinguishes fluid versus non-fluid. Using a color map overlay in the areas of enhancing signal on the FLAIR sequence further increased the conspicuity of the abnormal endolymphatic space in patients with Me´nie`re’s disease. The fused color map demonstrated decreased specificity to diagnose Me´nie`re’s disease compared with the gray-scale FLAIR images (88.9% fused color map versus 97.4% gray-scale image). The asymptomatic inner ear of one patient with Me´nie`re’s disease was thought to be abnormal by both readers on the fused color map, because of perceived ‘‘bites’’ within the cochlea. The identification of ‘‘bites’’ within the cochlea may therefore not be as reliable in the identification of hydrops compared with the identification of an enlarged endolymphatic space on the gray-scale 3D FLAIR images. It is also possible that the asymptomatic inner ear of the patient with Me´nie`re’s disease in fact has associated abnormal findings, which perhaps was better evaluated by the color map. In addition, our study demonstrated increased diagnostic accuracy in patients with a longer time interval since the time of initial presentation (Q3 years), suggesting that the duration of symptomatology correlates with the presence of appreciable endolymphatic hydrops on imaging. In other words, the degree of endolymphatic hydrops may be minimal in patients with early Me´nie`re’s disease such that it is not appreciated on MR. However, this assessment needs to be further evaluated with a larger sample size. Limitations in our study include a small number of subjects, with our controls not age- or sex-matched to our patient population. As with other imaging studies of Me´nie`re’s disease, a limitation with our study is the lack of histological confirmation of Me´nie`re’s disease in our patients. There is currently no gold standard for the diagnosis aside from postmortem histological evaluation. However, in our study, we only scanned patients with a high clinical suspicion of Me´nie`re’s disease selected by experienced neuro-otologists. Most studies using intravenous administration of contrast to evaluate endolymphatic hydrops utilized a double dose of contrast, as was done in our study, given the limited perilymphatic contrast uptake with a single dose (2,7,16). Our study did not include any patients with risk factors for renal disease, but nephrogenic systemic sclerosis remains a potential risk in patients receiving a double dose of gadolinium. More recent studies have investigated the optimization of imaging techniques using a single dose of contrast to visualize endolymphatic hydrops, including the HYDROPS and HYDROPS2 technique developed by Naganawa et al. (2,17Y19). We plan to further investigate single-dose methods to evaluate the diagnosis of Me´nie`re’s disease in a prospective manner. The scanning time of the 3D FLAIR sequence was 12 minutes and the CISS sequence was 4 minutes 30 seconds,
e341
for a total scanning time of 16 minutes 30 seconds. Motion artifact was not an issue for any of our patients in the study during each sequence or between sequences. The utilized post-processing software also corrects for motion between sequences, though this was not necessary for any of our patients. However, a patient with Me´nie`re’s disease during an acute episode may be unable to remain immobile during image acquisition, which would limit the utility of an MR examination. Another limitation to our study is the lack of objective analysis of the inner ears and lack of grading of the degree of hydrops in our patients. We felt that the small size and limited resolution mitigated against reliable and accurate measurements. We also wanted to assess the predictive value using subjective analysis of the inner ears. Nakashima et al. have proposed a three-stage grading system for hydrops in both the cochlea and vestibule: none, mild, and significant. In the vestibule, the grading system was determined based on ratio of the area of the endolymphatic space to that of total fluid space within the vestibule, with less than 33.3% ratio designated as no hydrops (20). Using this grading system, a recent multicenter study of Me´nie`re’s disease demonstrated 65% of asymptomatic contralateral inner ears of patients with Me´nie`re’s disease showed endolymphatic hydrops on MRI (21). However, a separate cadaveric study of 67 autopsies with histologic endolymphatic hydrops demonstrated bilaterality in only 20 (29.9%) specimens (22), raising the possibility that the inclusion of all patients with greater than 33.3% area ratio results in an increased falsepositive rate. In our experience, we have found that the area ratio can vary significantly depending on slice selection of the vestibule and potentially could be distorted by different slice angles. In addition, some of our control patients had area ratios exceeding 33.3%, though this may be related to differences in imaging technique between the studies. We used the absence of visualization of the contrast-enhanced vestibular perilymphatic space as the most reliable criterion to identify endolymphatic hydrops. Volume calculations may be a more useful and reliable method of grading hydrops in future studies. In conclusion, we have confirmed that delayed 3D FLAIR imaging of the inner ears obtained 4 hours after the intravenous administration of contrast is a potentially useful diagnostic tool in the evaluation of patients with Me´nie`re’s disease. Fused 3D FLAIR-CISS color map images can help increase the sensitivity for diagnosis, though is less specific compared with the gray-scale 3D FLAIR images.
REFERENCES 1. Committee on Hearing and Equilibrium. Committee on Hearing and Equilibrium guidelines for the diagnosis and evaluation of therapy in Meniere’s disease. Otolaryngol Head Neck Surg 1995;113:181Y5. 2. Sano R, Teranishi M, Yamazaki M, et al. Contrast enhancement of the inner ear in magnetic resonance images taken at 10 minutes or 4 hours after intravenous gadolinium injection. Acta Otolaryngol 2012;132:241Y6. Otology & Neurotology, Vol. 35, No. 10, 2014
Copyright © 2014 Otology & Neurotology, Inc. Unauthorized reproduction of this article is prohibited.
e342
M. HAGIWARA ET AL.
3. Naganawa S, Kawai H, Sone M, Nakashima T. Increased sensitivity to low concentration gadolinium contrast by optimized heavily T2weighted 3D-FLAIR to visualize endolymphatic space. Magn Reson Med Sci 2010;9:73Y80. 4. Niyazov D, Andrews JC, Strelioff D, Sinha S, Lufkin R. Diagnosis and endolymphatic hydrops in vivo with magnetic resonance imaging. Otol Neurotol 2001;22:813Y7. 5. Landis JR, Koch GG. The measurement of observer agreement for categorical data. Biometrics 1977;33:159Y74. 6. Carfrae M, Holtzman A, Eames F, Parnes SM, Lupinetti A. 3 Tesla delayed contrast magnetic resonance imaging evaluation of Meniere’s disease. Laryngoscope 2008;118:501Y5. 7. Naganawa S, Komada T, Fukatsu H, Ishigaki T, Takizawa O. Observation of contrast enhancement in the cochlear fluid space of healthy subjects using a 3D-FLAIR sequence at 3 Tesla. Eur Radiol 2006;16:733Y7. 8. Nakashima T, Naganawa S, Teranishi M, et al. Endolymphatic hydrops revealed by intravenous gadolinium injection in patients with Meniere’s disease. Acta Otolaryngol 2010;130:338Y43. 9. Yamazaki M, Naganawa S, Tagaya M, et al. Comparison of contrast effect on the cochlear perilymph after intratympanic and intravenous gadolinium injection. AJNR Am J Neuroradiol 2012;33:773Y8. 10. Naganawa S, Yamazaki M, Kawai H, Bokura K, Sone M, Nakashima T. Imaging of endolymphatic and perilymphatic fluid after intravenous administration of single-dose gadodiamide. Magn Reson Med Sci 2012;11:145Y50. 11. Nakashima T, Naganawa S, Sugiura M, et al. Visualization of endolymphatic hydrops in patients with Meniere’s disease. Laryngoscope 2007;117:415Y20. 12. Fang ZM, Chen X, Gu X, et al. A new magnetic resonance imaging scoring system for perilymphatic space appearance after intratympanic gadolinium injection, and its clinical application. J Laryngol Otol 2012;126:454Y9. 13. Naganawa S, Sugiura M, Kawamura M, Fukatsu H, Sone M, Nakashima T. Imaging of endolymphatic and perilymphatic fluid at 3T after intratympanic administration of gadolinium-diethylene-triamine pentaacetic acid. AJNR Am J Neuroradiol 2008;29:724Y6.
14. Naganawa S, Satake H, Kawamura M, Fukatsu H, Sone M, Nakashima T. Separate visualization of endolymphatic space, perilymphatic space and bone by a single pulse sequence; 3D-inversion recovery imaging utilizing real reconstruction after intratympanic Gd-DTPA administration at 3 Tesla. Eur Radiol 2008;18:920Y4. 15. Nakashima T, Naganawa S, Katayama N, et al. Clinical significance of endolymphatic imaging after intratympanic gadolinium injection. Acta Otolaryngol Suppl 2009;129:9Y14. 16. Tagaya M, Teranishi M, Naganawa S, et al. 3 Tesla magnetic resonance imaging obtained 4 hours after intravenous gadolinium injection in patients with sudden deafness. Acta Otolaryngol 2010; 130:665Y9. 17. Naganawa S, Yamazaki M, Kawai H, Bokura K, Sone M, Nakashima T. Imaging of Meniere’s disease after intravenous administration of single-dose gadodiamide: utility of subtraction images with different inversion time. Magn Reson Med Sci 2012; 11:213Y9. 18. Naganawa S, Yamazaki M, Kawai H, Bokura K, Sone M, Nakashima T. Imaging of Meniere’s disease by subtraction of MR cisternography from positive perilymph image. Magn Reson Med Sci 2012;11:303Y9. 19. Tanigawa T, Tamaki T, Yamamuro O, et al. Visualization of endolymphatic hydrops after administration of a standard dose of an intravenous gadolinium-based contrast agent. Acta Otolaryngol 2011; 131:596Y601. 20. Nakashima T, Naganawa S, Pyykko I, et al. Grading of endolymphatic hydrops using magnetic resonance imaging. Acta Otolaryngol Suppl 2009;129:5Y8. 21. Pyykko I, Nakashima T, Yoshida T, Zou J, Naganawa S. Meniere’s disease: a reappraisal supported by a variable latency of symptoms and the MRI visualisation of endolymphatic hydrops. BMJ Open 2013;3:1Y10. 22. Yazawa Y, Kitahara M. Bilateral endolymphatic hydrops in Meniere’s disease: review of temporal bone autopsies. Ann Otol Rhinol Laryngol 1990;99(7 Pt 1):524Y8.
Otology & Neurotology, Vol. 35, No. 10, 2014
Copyright © 2014 Otology & Neurotology, Inc. Unauthorized reproduction of this article is prohibited.
Identification of Endolymphatic Hydrops in Me´nie`re’s Disease Utilizing Delayed Postcontrast 3D FLAIR and Fused 3D FLAIR and CISS Color Maps *Mari Hagiwara, †J. Thomas Roland Jr, *Xin Wu, *Annette Nusbaum, *James S. Babb, ‡Pamela C. Roehm, †Paul Hammerschlag, §Anil K. Lalwani, and *Girish Fatterpekar *Departments of Radiology and ÞOtolaryngology, New York University School of Medicine, New York, New York, U.S.A.; þDepartment of Otolaryngology, Temple University, Philadelphia, Pennsylvania, and §Department of Otolaryngology, Columbia University, New York, New York, U.S.A.
Objective: The preferential delayed enhancement of the perilymphatic space enables detection of the non-enhancing endolymphatic hydrops present in patients with Me´nie`re’s disease. The aim of this study was to evaluate the diagnostic utility of delayed postcontrast 3D FLAIR images and a color map of fused postcontrast FLAIR and constructive interference steady state (CISS) images in the identification of endolymphatic hydrops in patients with clinically diagnosed Me´nie`re’s disease. Study Design: Case control, blinded study. Setting: Tertiary referral center. Patients: Ten patients with Me´nie`re’s disease and five volunteer controls. Intervention: Diagnostic. Main Outcome Measure: Two neuroradiologists blinded to the clinical history independently evaluated for the presence of endolymphatic hydrops on the images of both inner ears for test and control subjects. Both the standard gray-scale FLAIR images and the fused color map images were independently reviewed.
Results: The gray-scale 3D FLAIR images demonstrated 68.2% sensitivity and 97.4% specificity, and the fused color map images demonstrated 85.0% sensitivity and 88.9% specificity in the identification of endolymphatic hydrops in Me´nie`re’s disease. There was significant correlation between the gray-scale 3D FLAIR images and fused color map images with the categorization of involvement (p = 0.002). Inter-evaluator reliability was excellent (kappa = 0.83 for gray-scale images, kappa = 0.81 for fused color map). Conclusion: Delayed 3D FLAIR and fused 3D FLAIR-CISS color map images of the inner ears after intravenous contrast administration are potentially useful diagnostic tools in the evaluation of patients with suspected Me´nie`re’s disease. Key Words: Delayed postcontrast FLAIRVEndolymphatic hydropsV Me´nie`re’s disease.
Me´nie`re’s disease is an idiopathic disorder of the inner ear characterized by recurrent episodic vertigo, fluctuating hearing loss, aural fullness, and tinnitus secondary to enlargement of the endolymphatic space (endolymphatic hydrops). The current accepted method of diagnosing
Me´nie`re’s disease with certainty requires a postmortem histopathologic confirmation of endolymphatic hydrops (1). Me´nie`re’s disease has therefore remained a clinical diagnosis, made with varying degrees of certainty based on the criteria reported by the Committee on Hearing and Equilibrium (1). Absence of a reliable diagnostic test has hampered not only the ability to definitively diagnose Me´nie`re’s disease but also to design and assess the effectiveness of treatment. However, imaging studies within the past decade have demonstrated visualization of endolymphatic hydrops in patients utilizing delayed 3D FLAIR imaging after the intratympanic and intravenous injections of gadolinium contrast (2,3). The delayed preferential enhancement of the perilymphatic space has enabled distinction between the
Otol Neurotol 35:e337Ye342, 2014.
Address correspondence and reprint requests to Mari Hagiwara, M.D., 660 1st Avenue, 2nd floor Radiology, NYU Langone Medical Center, New York, NY 10016, U.S.A.; E-mail: [email protected] The study was partly funded by an $8,000 seed grant (account number 51-C63370-43757-NYUPRJ) from the Department of Radiology at the NYU Langone Medical Center. There was no salary support from the seed grant. The authors disclose no conflicts of interest.
e337
Copyright © 2014 Otology & Neurotology, Inc. Unauthorized reproduction of this article is prohibited.
e338
M. HAGIWARA ET AL.
fluid-filled perilymphatic and endolymphatic spaces, and thereby has allowed for the detection of an enlarged nonenhanced endolymphatic space (4). The unenhanced endolymphatic space is difficult to distinguish from bone on the 3D FLAIR images. We have found that fusion of the 3D FLAIR sequence with the constructive interference steady state (CISS) sequence can help differentiate the unenhanced endolymphatic space from bone, and using a color map overlay in the areas of increased signal on the FLAIR sequence can increase the conspicuity of the abnormal endolymphatic space in patients with Me´nie`re’s disease. The purpose of this study was to evaluate the diagnostic utility of delayed postcontrast high-resolution FLAIR images and a color map of fused postcontrast FLAIR and CISS images in the identification of endolymphatic hydrops in patients with Me´nie`re’s disease. MATERIALS AND METHODS Study Subjects Ten patients with high clinical suspicion of Me´nie`re’s disease and five controls without Me´nie`re’s disease and with normal hearing were included in the study. Criteria for clinically suspected Me´nie`re’s disease were based on the guidelines established by the Committee on Hearing and Equilibrium in 1995; the criteria includes two or more definitive spontaneous episodes of vertigo lasting 20 minutes or longer, audiometrically documented lowfrequency hearing loss on at least one occasion, tinnitus and aural fullness in the affected ear, and exclusion of other causes (1). All patients had normal conventional MR imaging of the vestibulocochlear complexes and internal auditory canals that excluded an underlying lesion which could mimic Me´nie`re’s disease. None of the patients had a history or current symptomatology of migraine or vestibular migraine. All patients were undergoing medical treatment for Me´nie`re’s disease. Exclusion criteria were based on contraindications to MRI and intravenous contrast administration, including patient history of cardiac pacemaker, intracranial clips, pregnant or nursing, and renal disease. Both the patient population and the volunteers had normal renal function tests and did not suffer from any renal disease in the past. None of the study subjects had contraindications to MRI. We evaluated 20 inner ears (right and left) of the 10 patients and 10 inner ears of the five control asymptomatic volunteers. Bilateral symptoms were reported in one patient. Unilateral right-sided symptoms were reported in six patients, and unilateral left-sided symptoms were reported in three patients. The age range of the 10 patients with Me´nie`re’s disease was 31 to 72 years old, with five men and five women. The age range of the five volunteer controls was 27 to 30 years old; all five controls were men. All patients and control subjects provided institutional review boardYapproved written consent, and the study was Health Insurance Portability and Accountability Act compliant. On the day of imaging, all patients filled out a questionnaire specifying the side of involvement, symptomatology, date of the initial presentation, and date of the most recent episode.
MR Imaging All imaging was performed on a 3-Tesla MR imaging scanner (Tim-Trio; Siemens, Erlangen, Germany) using a 12-channel array head coil. All patients underwent MR scanning before and
after the intravenous administration of double-dose (0.2 mmol/kg body weight) gadolinium (gadopentetate dimeglumine, Magnevist; Bayer Health Care, Wayne, NJ). Four hours after the intravenous administration of contrast, a 3D FLAIR sequence and a CISS sequence of the inner ears were obtained. The parameters for the 3D FLAIR sequence were as follows: TR 9,000 ms, TE 315 ms, TI 2,500 ms, matrix 256 242, 16 1-mm-thick slices, FOV 17.5 16.4 cm, variable flip angle, echo train length 121, NEX 2.0, iPAT off, voxel size 0.7 0.7 1.0 mm, and acquisition time 12 minutes. The parameters for the CISS sequence were as follows: TR 9.12, TE 4.56, matrix 448 358, 32 0.6-mm-thick slices, FOV 14 14 cm, flip angle 50 degrees, echo train length 1, NEX 1, iPAT off, voxel size 0.4 0.3 0.6 mm, and acquisition time 4 minutes 30 seconds.
Image Analysis Fusion of the 3D FLAIR and CISS images and color overlay maps were generated using Sphere v2.2 post-processing software (Olea Medical, La Ciotat, France). The fused color maps were obtained in all patients and controls except for one patient due to technical error. Two neuroradiologists (M.H., 5 years’ experience and G.F., 12 years’ experience) blinded to clinical history independently reviewed the images of both (right and left) inner ears for test and control subjects. The readers subjectively determined the presence or absence of inner ear hydrops and specified whether the vestibule and/or cochlea were involved. Both the standard gray-scale FLAIR images and the fused color map images were separately reviewed in this manner, with randomization of the presentation order.
Gray-scale FLAIR Identification of an enlarged non-enhancing endolymphatic compartment seen as ‘‘black circles’’ in the cochlea was used to determine the presence of endolymphatic hydrops (Fig. 1). Nonvisualization of the vestibule was used to determine hydrops within the vestibule.
Fused Color Map Hyperintense (enhancing) signal on the FLAIR sequence was colorized, while the CISS sequence was left in gray scale. Identification of inward concavities seen as ‘‘bites’’ within the cochlea was used to designate the presence of hydrops within the cochlea
FIG. 1. Gray-scale 3D FLAIR. A, Identification of an enlarged non-enhancing endolymphatic compartment seen as ‘‘black circles’’ (solid arrows) in the cochlea was used to determine the presence of endolymphatic hydrops. Lack of visualization of the vestibule in its expected location (dashed arrow) was the determining factor used to determine hydrops within the vestibule. B, Lack of an enlarged discretely identifiable endolymphatic space in the cochlea (solid arrow) and visualization of the normal enhancing vestibule (dashed arrow) were used to determine the absence of endolymphatic hydrops.
Otology & Neurotology, Vol. 35, No. 10, 2014
Copyright © 2014 Otology & Neurotology, Inc. Unauthorized reproduction of this article is prohibited.
ENDOLYMPHATIC HYDROPS IN ME´NIE`RE_S DISEASE (Fig. 2). Lack of color (reflecting absence of enhancement on the FLAIR sequence) within the visualized vestibule was used to determine the presence of hydrops within the vestibule.
TABLE 1. The sensitivity and specificity achieved when ears were classified as affected (test positive for Me´nie`re’s disease) when a specific abnormal finding was detected Finding
Statistical Analysis Logistic regression for correlated data was used to assess the diagnostic utility of imaging findings for the detection of affected ears. Specifically, generalized estimating equations based on binary logistic regression was used to model the status of each ear (affected versus unaffected) as a function of reader and each imaging finding. Statistical dependencies among the multiple results recorded for each individual subject (person) were accounted for by assuming results to be correlated when derived for the same subject and independent when derived for different subjects. Interreader agreement was assessed in terms of simple kappa coefficients with kappa greater than 0.6 interpreted as substantial agreement (5). All statistical tests were conducted at the two-sided 5% significance level using SAS 9.3 (SAS Institute, Cary, NC).
RESULTS There was no significant difference ( p 9 0.4) between the unaffected ears of the Me´nie`re’s patients and the unaffected ears of controls in terms of the prevalence of any imaging finding. Because of lack of significant difference, these two groups were then combined in one unaffected ear category. Significant correlation ( p G 0.001) was demonstrated between categorization of involvement with both the grayscale FLAIR images and the fused color map images. For overall assessment of the presence of Me´nie`re’s disease, the gray-scale 3D FLAIR images demonstrated lower sensitivity (68.2% versus 85%) and greater specificity (97.4% versus 88.9%) compared with the fused FLAIR-CISS color maps (Table 1). For evaluation of the vestibule, the grayscale FLAIR images showed lower sensitivity (63.6% versus 75.0%) and slightly greater specificity (97.4% versus 94.4%) compared with the fused color maps. For evaluation of the cochlea, the gray-scale FLAIR images demonstrated lower sensitivity (68.2% versus 80.0%) and greater specificity (94.7% versus 77.8%) compared with the fused color maps.
e339
Overall gray scale Vestibule gray scale Cochlea gray scale Overall color map Vestibule color map Cochlea color map
Specificity 97.4% 97.4% 94.7% 88.9% 94.4% 77.8%
(37/38) (37/38) (36/38) (32/36) (34/36) (28/36)
Sensitivity
p
68.2% (15/22) 63.6% (14/22) 68.2% (15/22) 85.0% (17/20) 75.0% (15/20) 80.0% (16/20)
G0.001 0.003 G0.001 G0.001 G0.001 0.001
Numerators and denominators are counts of evaluations (summed over readers) rather than counts of ears or subjects; variation in the denominators within a single column is a result of missing data. p values are from the generalized estimating equations analysis to test whether the relevant finding was a predictor of whether or not an ear was affected. There was no set of two or more imaging measures representing independent predictors of whether or not ears were affected.
The kappa for inter-reader agreement was 0.83 for the gray-scale 3D FLAIR images and 0.81 for fused color map images, indicating substantial agreement between readers. More specifically, of the 20 inner ears of patients with Me´nie`re’s disease, there was in consensus between the two readers on the 3D FLAIR images: seven true-positives, eight true-negatives, and three false-negatives. There was one patient where there was discrepancy between the two readers for both the affected and unaffected inner ears. Of the 18 inner ears of patients with Me´nie`re’s disease, there was in consensus on the fused color maps: eight truepositives, seven true-negatives, one false-negative, and one false-positive (Fig. 3). There was discrepancy between the readers for one of the affected ears. Of the 10 inner ears of the five control patients, there were eight true-negatives in consensus on the 3D FLAIR and fused color images (Fig. 4). On the fused color maps, there was discrepancy between readers for one control subject, with both inner ears incorrectly determined as abnormal by one reader. All of the false-negative patients with Me´nie`re’s disease based on the gray-scale FLAIR images had their initial presentation within the past 2 years. All of the true-positive inner ears were in patients with the initial presentation greater than 3 years prior (range 3Y28 years). One false-positive on the color map was present in the contralateral ear of a patient with Me´nie`re’s disease with unilateral symptoms. DISCUSSION
FIG. 2. Fused color map. A, Identification of ‘‘bites’’ within the cochlea (solid arrows) was used to designate the presence of hydrops within the cochlea. Lack of color (reflecting lack of enhancement on the FLAIR sequence) within the visualized vestibule (dashed arrow) was used to determine the presence of hydrops within the vestibule. B, Lack of ‘‘bites’’ within the cochlea with normal convex margins (solid arrows) and visualization of color signal within the vestibule (dashed arrow) were used to determine absence of endolymphatic hydrops.
Me´nie`re’s disease is an idiopathic debilitating disorder of the inner ear characterized by recurrent episodic vertigo, hearing loss, aural fullness, and tinnitus. The currently accepted criteria to diagnose Me´nie`re’s disease with certainty require postmortem histopathologic evaluation of the temporal bones, and therefore Me´nie`re’s disease has largely remained a clinical diagnosis (1). Identification of endolymphatic hydrops on imaging can not only enable a definitive diagnosis of Me´nie`re’s disease in living patients but can also potentially help determine therapeutic strategies and evaluate treatment response, as well as provide Otology & Neurotology, Vol. 35, No. 10, 2014
Copyright © 2014 Otology & Neurotology, Inc. Unauthorized reproduction of this article is prohibited.
e340
M. HAGIWARA ET AL.
further understanding of the pathophysiology of the poorly understood disorder (6,7). Distinction of the fluid-filled endolymphatic and perilymphatic spaces has been limited on conventional imaging because of its small size and identical signal intensities. Within the past decade, research studies have investigated the use of delayed high-resolution imaging to visualize endolymphatic hydrops after the intratympanic and intravenous injections of gadolinium contrast (2). Preferential transfer of gadolinium into the perilymphatic space has enabled visual distinction of the endolymphatic and perilymphatic spaces, and thereby has allowed for the detection of an enlarged non-enhanced endolymphatic space (4,7). Intratympanic injection of contrast has been shown to provide better image contrast compared with intravenous injection because of a greater degree of perilymphatic enhancement (8,9). However, this method of contrast administration is more invasive, requires off-label use of the contrast material, requires 24-hour interval between the time of injection and of imaging, and enables evaluation of only one inner ear at a time (9,10). Intravenous injection of contrast is the standard use of contrast material, requires only a 4-hour waiting period, and enables evaluation of both inner ears at a time (9); intravenous contrast administration thus is the ideal method of choice in the evaluation of Me´nie`re’s disease. The delayed postcontrast MR imaging findings seen in patients with Me´nie`re’s disease are thought to correspond to an enlarged endolymphatic space. The ‘‘black circles’’ in the cochlea are thought to reflect the dilated unenhanced endolymphatic scala media compartment, which is barely perceptible in the normal patient. On the fused color map images, this enlarged unenhanced endolymphatic space resulted in the appearance of ‘‘bites’’ in the colorized cochlea. Non-visualization of the vestibule on the 3D FLAIR
FIG. 3. Meniere’s disease patient with right-sided symptoms. Gray-scale 3D FLAIR (A) and fused color map (B) demonstrates endolymphatic hydrops within the cochlea and vestibule on the right side and normal-appearing endolymphatic space on the left side. Note the enlarged rounded endolymphatic space on the gray-scale image and apparent ‘‘bites’’ on the color map within the right cochlea (solid arrows) compared with the normal left side. Also note the lack of enhancement on the gray-scale image and lack of color on the color map within the right vestibule (dashed arrows), compared with the normal left side.
FIG. 4. Control patient. Gray-scale 3D FLAIR (A) and fused color map (B) demonstrate normal-appearing cochleae (solid arrows) and vestibules (dashed arrows) bilaterally.
sequence has also been speculated to reflect marked enlargement of the endolymphatic space, which either completely replaces the fluid in the perilymphatic space and/or prevents the movement of gadolinium from entering the perilymphatic space (11). On the fused color map images, this was seen as lack of color within the vestibule, corresponding to the absence of gadolinium contrast. Our caseYcontrol, blinded study demonstrated good interobserver reliability as well as good specificity in the identification of Me´nie`re’s disease for the 3D FLAIR images after the intravenous administration of contrast. The gray-scale 3D FLAIR images demonstrated 68.2% sensitivity and 97.4% specificity in the identification of endolymphatic hydrops in Me´nie`re’s disease. In their study of 207 ears (31 with Me´nie`re’s disease symptoms), Fang et al. demonstrated greater sensitivity (97.5%) and specificity (98%) using gray-scale 3D FLAIR images after the intratympanic injection of contrast (12). The decreased sensitivity and specificity in our study may be related to smaller sample size and improved contrast resolution with intratympanic injections. Based on our literature search, this is the only study to evaluate the sensitivity and specificity of MRI after the intravenous administration of contrast. This is also the first study to use fused color maps of contrast-enhanced images and CISS images for the identification of endolymphatic hydrops. Using the fused color images, we noticed that the sensitivity to identify endolymphatic hydrops increased from 68.2% to 85.0%. We speculate this increased sensitivity to result from the ability of the fused color images to better distinguish the unenhanced endolymphatic space from the underlying bone. Previous studies have suggested that the unenhanced hypointense endolymphatic space is difficult to distinguish from the hypointense bone on the 3D FLAIR images (13,14). Naganawa et al. optimized the inversion time of a 3D inversion-recovery turbo spin-echo sequence so that the
Otology & Neurotology, Vol. 35, No. 10, 2014
Copyright © 2014 Otology & Neurotology, Inc. Unauthorized reproduction of this article is prohibited.
ENDOLYMPHATIC HYDROPS IN ME´NIE`RE_S DISEASE hypointense endolymphatic space was of a different signal (black) relative to the signal (gray) of the surrounding bone (14,15). We found that fusion of the 3D FLAIR sequence with the CISS sequence more clearly differentiates the unenhanced endolymphatic space from the bone, as the CISS sequence clearly distinguishes fluid versus non-fluid. Using a color map overlay in the areas of enhancing signal on the FLAIR sequence further increased the conspicuity of the abnormal endolymphatic space in patients with Me´nie`re’s disease. The fused color map demonstrated decreased specificity to diagnose Me´nie`re’s disease compared with the gray-scale FLAIR images (88.9% fused color map versus 97.4% gray-scale image). The asymptomatic inner ear of one patient with Me´nie`re’s disease was thought to be abnormal by both readers on the fused color map, because of perceived ‘‘bites’’ within the cochlea. The identification of ‘‘bites’’ within the cochlea may therefore not be as reliable in the identification of hydrops compared with the identification of an enlarged endolymphatic space on the gray-scale 3D FLAIR images. It is also possible that the asymptomatic inner ear of the patient with Me´nie`re’s disease in fact has associated abnormal findings, which perhaps was better evaluated by the color map. In addition, our study demonstrated increased diagnostic accuracy in patients with a longer time interval since the time of initial presentation (Q3 years), suggesting that the duration of symptomatology correlates with the presence of appreciable endolymphatic hydrops on imaging. In other words, the degree of endolymphatic hydrops may be minimal in patients with early Me´nie`re’s disease such that it is not appreciated on MR. However, this assessment needs to be further evaluated with a larger sample size. Limitations in our study include a small number of subjects, with our controls not age- or sex-matched to our patient population. As with other imaging studies of Me´nie`re’s disease, a limitation with our study is the lack of histological confirmation of Me´nie`re’s disease in our patients. There is currently no gold standard for the diagnosis aside from postmortem histological evaluation. However, in our study, we only scanned patients with a high clinical suspicion of Me´nie`re’s disease selected by experienced neuro-otologists. Most studies using intravenous administration of contrast to evaluate endolymphatic hydrops utilized a double dose of contrast, as was done in our study, given the limited perilymphatic contrast uptake with a single dose (2,7,16). Our study did not include any patients with risk factors for renal disease, but nephrogenic systemic sclerosis remains a potential risk in patients receiving a double dose of gadolinium. More recent studies have investigated the optimization of imaging techniques using a single dose of contrast to visualize endolymphatic hydrops, including the HYDROPS and HYDROPS2 technique developed by Naganawa et al. (2,17Y19). We plan to further investigate single-dose methods to evaluate the diagnosis of Me´nie`re’s disease in a prospective manner. The scanning time of the 3D FLAIR sequence was 12 minutes and the CISS sequence was 4 minutes 30 seconds,
e341
for a total scanning time of 16 minutes 30 seconds. Motion artifact was not an issue for any of our patients in the study during each sequence or between sequences. The utilized post-processing software also corrects for motion between sequences, though this was not necessary for any of our patients. However, a patient with Me´nie`re’s disease during an acute episode may be unable to remain immobile during image acquisition, which would limit the utility of an MR examination. Another limitation to our study is the lack of objective analysis of the inner ears and lack of grading of the degree of hydrops in our patients. We felt that the small size and limited resolution mitigated against reliable and accurate measurements. We also wanted to assess the predictive value using subjective analysis of the inner ears. Nakashima et al. have proposed a three-stage grading system for hydrops in both the cochlea and vestibule: none, mild, and significant. In the vestibule, the grading system was determined based on ratio of the area of the endolymphatic space to that of total fluid space within the vestibule, with less than 33.3% ratio designated as no hydrops (20). Using this grading system, a recent multicenter study of Me´nie`re’s disease demonstrated 65% of asymptomatic contralateral inner ears of patients with Me´nie`re’s disease showed endolymphatic hydrops on MRI (21). However, a separate cadaveric study of 67 autopsies with histologic endolymphatic hydrops demonstrated bilaterality in only 20 (29.9%) specimens (22), raising the possibility that the inclusion of all patients with greater than 33.3% area ratio results in an increased falsepositive rate. In our experience, we have found that the area ratio can vary significantly depending on slice selection of the vestibule and potentially could be distorted by different slice angles. In addition, some of our control patients had area ratios exceeding 33.3%, though this may be related to differences in imaging technique between the studies. We used the absence of visualization of the contrast-enhanced vestibular perilymphatic space as the most reliable criterion to identify endolymphatic hydrops. Volume calculations may be a more useful and reliable method of grading hydrops in future studies. In conclusion, we have confirmed that delayed 3D FLAIR imaging of the inner ears obtained 4 hours after the intravenous administration of contrast is a potentially useful diagnostic tool in the evaluation of patients with Me´nie`re’s disease. Fused 3D FLAIR-CISS color map images can help increase the sensitivity for diagnosis, though is less specific compared with the gray-scale 3D FLAIR images.
REFERENCES 1. Committee on Hearing and Equilibrium. Committee on Hearing and Equilibrium guidelines for the diagnosis and evaluation of therapy in Meniere’s disease. Otolaryngol Head Neck Surg 1995;113:181Y5. 2. Sano R, Teranishi M, Yamazaki M, et al. Contrast enhancement of the inner ear in magnetic resonance images taken at 10 minutes or 4 hours after intravenous gadolinium injection. Acta Otolaryngol 2012;132:241Y6. Otology & Neurotology, Vol. 35, No. 10, 2014
Copyright © 2014 Otology & Neurotology, Inc. Unauthorized reproduction of this article is prohibited.
e342
M. HAGIWARA ET AL.
3. Naganawa S, Kawai H, Sone M, Nakashima T. Increased sensitivity to low concentration gadolinium contrast by optimized heavily T2weighted 3D-FLAIR to visualize endolymphatic space. Magn Reson Med Sci 2010;9:73Y80. 4. Niyazov D, Andrews JC, Strelioff D, Sinha S, Lufkin R. Diagnosis and endolymphatic hydrops in vivo with magnetic resonance imaging. Otol Neurotol 2001;22:813Y7. 5. Landis JR, Koch GG. The measurement of observer agreement for categorical data. Biometrics 1977;33:159Y74. 6. Carfrae M, Holtzman A, Eames F, Parnes SM, Lupinetti A. 3 Tesla delayed contrast magnetic resonance imaging evaluation of Meniere’s disease. Laryngoscope 2008;118:501Y5. 7. Naganawa S, Komada T, Fukatsu H, Ishigaki T, Takizawa O. Observation of contrast enhancement in the cochlear fluid space of healthy subjects using a 3D-FLAIR sequence at 3 Tesla. Eur Radiol 2006;16:733Y7. 8. Nakashima T, Naganawa S, Teranishi M, et al. Endolymphatic hydrops revealed by intravenous gadolinium injection in patients with Meniere’s disease. Acta Otolaryngol 2010;130:338Y43. 9. Yamazaki M, Naganawa S, Tagaya M, et al. Comparison of contrast effect on the cochlear perilymph after intratympanic and intravenous gadolinium injection. AJNR Am J Neuroradiol 2012;33:773Y8. 10. Naganawa S, Yamazaki M, Kawai H, Bokura K, Sone M, Nakashima T. Imaging of endolymphatic and perilymphatic fluid after intravenous administration of single-dose gadodiamide. Magn Reson Med Sci 2012;11:145Y50. 11. Nakashima T, Naganawa S, Sugiura M, et al. Visualization of endolymphatic hydrops in patients with Meniere’s disease. Laryngoscope 2007;117:415Y20. 12. Fang ZM, Chen X, Gu X, et al. A new magnetic resonance imaging scoring system for perilymphatic space appearance after intratympanic gadolinium injection, and its clinical application. J Laryngol Otol 2012;126:454Y9. 13. Naganawa S, Sugiura M, Kawamura M, Fukatsu H, Sone M, Nakashima T. Imaging of endolymphatic and perilymphatic fluid at 3T after intratympanic administration of gadolinium-diethylene-triamine pentaacetic acid. AJNR Am J Neuroradiol 2008;29:724Y6.
14. Naganawa S, Satake H, Kawamura M, Fukatsu H, Sone M, Nakashima T. Separate visualization of endolymphatic space, perilymphatic space and bone by a single pulse sequence; 3D-inversion recovery imaging utilizing real reconstruction after intratympanic Gd-DTPA administration at 3 Tesla. Eur Radiol 2008;18:920Y4. 15. Nakashima T, Naganawa S, Katayama N, et al. Clinical significance of endolymphatic imaging after intratympanic gadolinium injection. Acta Otolaryngol Suppl 2009;129:9Y14. 16. Tagaya M, Teranishi M, Naganawa S, et al. 3 Tesla magnetic resonance imaging obtained 4 hours after intravenous gadolinium injection in patients with sudden deafness. Acta Otolaryngol 2010; 130:665Y9. 17. Naganawa S, Yamazaki M, Kawai H, Bokura K, Sone M, Nakashima T. Imaging of Meniere’s disease after intravenous administration of single-dose gadodiamide: utility of subtraction images with different inversion time. Magn Reson Med Sci 2012; 11:213Y9. 18. Naganawa S, Yamazaki M, Kawai H, Bokura K, Sone M, Nakashima T. Imaging of Meniere’s disease by subtraction of MR cisternography from positive perilymph image. Magn Reson Med Sci 2012;11:303Y9. 19. Tanigawa T, Tamaki T, Yamamuro O, et al. Visualization of endolymphatic hydrops after administration of a standard dose of an intravenous gadolinium-based contrast agent. Acta Otolaryngol 2011; 131:596Y601. 20. Nakashima T, Naganawa S, Pyykko I, et al. Grading of endolymphatic hydrops using magnetic resonance imaging. Acta Otolaryngol Suppl 2009;129:5Y8. 21. Pyykko I, Nakashima T, Yoshida T, Zou J, Naganawa S. Meniere’s disease: a reappraisal supported by a variable latency of symptoms and the MRI visualisation of endolymphatic hydrops. BMJ Open 2013;3:1Y10. 22. Yazawa Y, Kitahara M. Bilateral endolymphatic hydrops in Meniere’s disease: review of temporal bone autopsies. Ann Otol Rhinol Laryngol 1990;99(7 Pt 1):524Y8.
Otology & Neurotology, Vol. 35, No. 10, 2014
Copyright © 2014 Otology & Neurotology, Inc. Unauthorized reproduction of this article is prohibited.
