Auris Nasus Larynx 37 (2010) 669–675 www.elsevier.com/locate/anl
Relationship between high-resolution computed tomography densitometry and audiometry in otosclerosis Mei-mei Zhu a, Yan Sha b, Pei-yun Zhuang c,d, Aleksandra E. Olszewski d, Jia-qi Jiang a,d, Jiang-hong Xu a, Chen-mei Xu a, Bing Chen a,* a
Otology & Skull Base Surgery Department, Eye Ear Nose & Throat Hospital, Fudan University, Fenyang Road 83, Shanghai 200031, China b Department of Radiology, Eye Ear Nose & Throat Hospital, Fudan University, Shanghai, China c Department of Otorhinolaryngology, Zhongshan Hospital, Xiamen University, Xiamen, China d Division of Otolaryngology-Head and Neck Surgery, Department of Surgery, University of Wisconsin-Madison, Madison, WI, USA Received 3 September 2009; accepted 11 March 2010
Abstract Objective: The aim of this study is to evaluate the usefulness of high-resolution computed tomography (HRCT) densitometry in the diagnosis of otosclerosis and to investigate the relationship between CT densitometry and audiometry. Methods: HRCT findings and audiometry were compared among 34 patients (34 ears, the otosclerosis group) with surgically confirmed otosclerosis between January 2007 and December 2007 and 33 patients (33 opposite normal ears, the control group) with facial paralysis diagnosed at the same period of time. Seven regions of interest (ROI) were set manually around the otic capsule on the axial slice of 0.75-mmthick CT image. The mean CT values of these seven regions were measured. In each ROI, the mean CT value of the otosclerosis group and that of the control group were compared. Based on the CT findings, the ears with otosclerosis were classified into two groups: Group A showed no pathological CT findings; Group B showed low density around the cochlea. In the otosclerosis group, the relationship between the findings of CT and the results of audiometry was analyzed. Results: The mean CT values in the area posterior to the oval window and anterior to the oval window were significantly lower for the otosclerosis group compared with the control group (the former t = 2.030, p = 0.046; the latter Z = 4.979, p < 0.01). Group A consisted of 30 patients, 7 of which (23.33%) exhibited conductive hearing loss, and 23 of which (76.67%) exhibited mixed hearing loss; Group B had 4 patients, all with mixed hearing loss. For the otosclerosis group, the mean CT value in the area posterior to the oval window was positively correlated with the mean air conduction threshold (r = 0.4273, p = 0.0117) and with the mean air-bone gap (r = 0.3995, p = 0.0192). Conclusion: Quantitative evaluation of CT with slices less than 1 mm in thickness may provide important information for the diagnosis and assessment of otosclerosis which are unattainable through other methods. # 2010 Elsevier Ireland Ltd. All rights reserved. Keywords: Otosclerosis; HRCT; Densitometry; Audiometry
1. Introduction Otosclerosis, an osteodystrophy of the otic capsule which can results in progressive hearing loss, is one of the main causes of deafness in adult patients. The diagnosis of otosclerosis is mainly based on detailed clinical history, physical examination, audiometry (including tuning-fork testing and pure tone audiometry). Recently, computed * Corresponding author. Tel.: +86 21 64377134. E-mail address:
[email protected] (B. Chen).
tomography (CT), especially high-resolution CT (HRCT) with thin collimation, has proven to be a promising approach in detection of small otosclerosis lesions and the assessment of otosclerosis [1]. However, some otosclerosis patients may exhibit no pathological CT findings; in such cases, CT can only be used to exclude differential diagnoses. As otosclerosis patients may present normal CT appearing, other CT parameters are indispensable for evaluating otosclerosis. Studies have been done to determine the relationship between CT findings and audiometry, but most relied on qualitative analysis [2–6]. Some of these studies included quantitative
0385-8146/$ – see front matter # 2010 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.anl.2010.03.002
670
M.-m. Zhu et al. / Auris Nasus Larynx 37 (2010) 669–675
analysis which only focused on the correlation between the CT value and the bone conduction threshold at different frequencies [4,5]. However, the correlation between the CT value and the air conduction threshold and that between the CT value and the air-bone gap have yet to be investigated. To our knowledge, the correlation between CT values and mean thresholds of speech frequencies has never been studied. As there is no broad consensus on CT densitometry around the otic capsule in otosclerosis patients, a general description of CT densitometry would also be beneficial towards standardizing diagnoses. The purpose of our study is to determine the function of HRCT densitometry in the diagnosis of otosclerosis and to explore the relationship between the findings of CT densitometry and audiometry, including the correlation between the CT value and the mean air conduction threshold, the mean bone conduction threshold and the mean air-bone gap of speech frequencies.
2. Materials and methods 2.1. Patients Investigated were 34 patients (34 ears, the otosclerosis group) with surgically confirmed otosclerosis and 33 patients (33 opposite normal ears, the control group) with peripheral facial paralysis found between January 2007 and December 2007 in the Eye Ear Nose & Throat Hospital. Each otosclerosis patient underwent surgery on only one ear in 2007. Thus, for the otosclerosis group, the selected side for the study was the ear which was operated on. The etiologies for the control group of peripheral facial paralysis patients included Bell’s palsy, temporal bone trauma and congenital cholesteatoma. All the patients were examined using HRCT and audiometry. In order to avoid deviation in CT values due to age difference, only patients between the ages of 10 and 60 years were included in both groups. In the otosclerosis group, patients with congenital malformation, delayed ossification of the otic capsule, decalcified disease of the temporal bone, and previous surgery in the middle or inner ear for any disease including otosclerosis were excluded. The control group had no previous history of ear disease and there was no conductive hearing loss in this group. For the control group, the presence of abnormal CT or audiometry findings was regarded as an exclusion criterion. The otosclerosis group consisted of 10 males and 24 females (average age, 36 years; age range, 19– 57 years). In the control group, 20 males and 13 females (average age 37 years; age range 13–59 years) were included (Table 1). All patients underwent preoperative HRCT examination and pure tone audiometry within 1 week prior to their preoperative CT examination. 2.2. CT examination All CT images were obtained using a ten-detector-row helical CT scanner (Semens, SOMATOM Sensation 10)
Table 1 Subject summary. Otosclerosis, n (%)
Control, n (%)
x2
p
Age (year) 10 20 30 40 50 Total
1 3 13 10 7 34
2 10 8 6 7 33
6.280
0.179
Ear Left Right
14 (41.18) 20 (58.82)
18 (54.55) 15 (45.45)
1.200
0.273
Sex Male Female
10 (29.41) 24 (70.59)
20 (60.61) 13 (39.39)
6.590
0.010a
(2.94) (8.82) (38.24) (29.41) (20.59) (100.00)
(6.06) (30.31) (24.24) (18.18) (21.21) (100.00)
Pearson chi-square test. a p = 0.010.
with 6.0 0.75-mm collimation and 0.75-mm slice thickness. The scan conditions were 120 kV, 180 mA, 0.75 mm/s, 512 512 matrix, and the field of view (FOV) was 22 cm. The window width was 4000 Hounsfield units (HU) and the window level was 700 HU. In the CT workstation, the stapes, vestibule, cochlea and internal auditory canal were depicted in the low axial section [7]. Seven circular regions of interest (ROIs) were manually set around the otic capsule in the low axial section (Fig. 1, Table 2). The area of each ROI was 1 mm2 and the mean CT value in each ROI was measured automatically in the CT workstation. As both the ROI 1 and ROI 2 appeared more likely to be affected by the partial volume effect than other ROIs, our methods for locating
Fig. 1. Example of the ROI setting. In the low axial section, seven ROIs are manually placed. The area of each ROI is 1 mm2 and the mean CT value in each ROI can be measured automatically using the tool in the CT workstation.
M.-m. Zhu et al. / Auris Nasus Larynx 37 (2010) 669–675 Table 2 Position of ROIs. Number
Position
1
Posterior to the oval window (posterior to the intersection of stapes posterior arch and stapes footplate) Anterior to the oval window (anterior to the intersection of stapes anterior arch and stapes footplate) Lateral to the middle turn of the cochlea Lateral to the part between apex of the cochlea and the middle turn of the cochlea Apex of the cochlea Anteromedial to the middle turn of the cochlea Anteromedial to the basal turn of the cochlea
2 3 4 5 6 7
ROI indicates region of interest.
these two ROIs were specifically aimed to minimize the likelihood of this effect. To decrease the possibility of the partial volume effect, ROI 1 and ROI 2 were located in the bony parts posterior or anterior to the intersections of the stapes posterior or anterior arch and stapes footplate. Also, the areas of the included bony parts for these ROIs were larger than 1 mm2. Based on the CT findings, the ears with otosclerosis were classified into two groups: Group A showed no pathological CT findings; Group B manifested low density around the cochlea (‘‘double-ring sign’’) (Fig. 2). All CT images were read by four experienced radiologists in the Eye Ear Nose & Throat Hospital. 2.3. Pure tone audiometry The audiometric test was conducted in the isolation booth using the audiometer VIRTUAL M322, headset Telephonics
671
296D200-1 and bone vibrator RADIOEAR B-71. Air conduction was recorded at 0.125, 0.25, 0.5, 1, 2, 4 and 8 kHz and the maximum air conduction values for each frequency were 85, 105, 110, 110, 110, 110 and 105 decibels hearing level (dB HL), respectively. Bone conduction was recorded at 0.25, 0.5, 1, 2 and 4 kHz and the maximum bone conduction values for each frequency were 35, 50, 65, 65 and 60 dB HL, respectively. The mean air conduction threshold and the mean bone conduction threshold were computed as the averages of the thresholds at the speech frequencies of 0.5, 1, 2, 4 kHz. The mean air-bone gap was defined as the mean air conduction threshold minus the mean bone conduction threshold. Conductive hearing loss (CHL) was diagnosed if the mean air-bone gap was more than 15 dB HL and the mean bone conduction threshold was less than 25 dB HL. If the mean air-bone gap was less than 15 dB HL and the mean bone conduction threshold was more than 25 dB HL, sensorineural hearing loss (SNHL) was diagnosed. When the mean air-bone gap was more than 15 dB HL and the mean bone conduction threshold was more than 25 dB HL, the diagnosis was mixed hearing loss (MHL). 2.4. Statistical analysis Qualitative data were described using percentages of total subjects in each group (%) and quantitative data were described using mean standard deviation (SD). The Pearson chi-square test was used to evaluate the comparability between the two groups by age, left or right ear, and sex. The statistical analysis of the CT densitometry results in each ROI between the otosclerosis group and the control group was performed using either a Student’s t-test or Wilcoxon rank-sum test according to the type of variables studied. Pearson linear correlation analysis was applied to evaluate the linear relationship between CT densitometry (HU) and hearing loss (dB HL) in each ROI for the otosclerosis group. The results were considered significant when the p value was less than 0.05.
3. Results
Fig. 2. A 52-year-old man with cochlear otosclerosis. The axial CT section shows the demineralization around the cochlea with obvious ‘‘double-ring sign’’ (arrow).
No significant differences in the number of patients were observed between the otosclerosis group and the control group when analyzing the groups according to age and left or right ear (the former x2 = 6.280, p = 0.179; the latter x2 = 1.200, p = 0.273). There was a significant difference between the two groups when organizing the groups by sex (x2 = 6.590, p = 0.010) (Table 1). The mean CT values in ROI 1 and ROI 2 were significantly lower for the otosclerosis group compared to the control group (the former t = 2.030, p = 0.046; the latter Z = 4.979, p < 0.01). No significant differences were found between subjects in the two groups when comparing CT values of the remaining ROIs (Figs. 3 and 4, Table 3).
672
M.-m. Zhu et al. / Auris Nasus Larynx 37 (2010) 669–675 Table 3 The mean CT value for each ROI (HU, mean SD). ROI a
1 2b 3 4 5 6 7
Otosclerosis (n=34)
Control (n=33)
p
1098.05 212.89 1007.63 253.00 1834.68 261.94 1879.62 181.20 1753.31 319.90 1983.62 165.72 1920.24 155.60
1201.53 204.01 1396.15 302.65 1933.34 153.16 1897.20 123.12 1811.21 183.93 1976.26 67.09 1942.02 84.30
0.046 0.05 >0.05 >0.05 >0.05 >0.05
ROI indicates region of interest; HU, Hounsfield unit; SD, standard deviation. a Student’s t-test t = 2.030, p = 0.046. b Wilcoxon rank-sum test Z = 4.979, p
