Download Bony Cochlear Nerve Canal Stenosis and Speech Discrimination in

The Laryngoscope
C 2014 The American Laryngological,
V
Rhinological and Otological Society, Inc.
Bony Cochlear Nerve Canal Stenosis and Speech Discrimination in
Pediatric Unilateral Hearing Loss
Patricia L. Purcell, MD; Ayaka J. Iwata, MD; Grace S. Phillips, MD; Angelisa M. Paladin, MD;
Kathleen C. Y. Sie, MD; David L. Horn, MD
Objectives/Hypothesis: To examine the relationship between bony cochlear nerve canal (BCNC) width, degree of hearing loss, and speech discrimination in children with unilateral sensorineural hearing loss (USNHL).
Study Design: Retrospective chart review (case-control study).
Methods: Audiometric database was cross-referenced with radiologic database at pediatric tertiary care facility to identify children with USNHL and temporal bone computed tomography. BCNC widths were measured independently by two radiologists blinded to affected ear. Regression analyses investigated associations among variables.
Results: One hundred and sixty children with USNHL had temporal bone imaging. Mean BCNC width was significantly
smaller in affected ears, P 5 0.0001. Narrower width was associated with more severe hearing loss, P 5 0.01. Among children
who had narrower cochlear nerve canals in affected ears compared to unaffected ears, smaller width was associated with
lower speech discrimination score, P 5 0.03. Increasing asymmetry in BCNC width between affected and unaffected ears was
associated with poorer discrimination scores, P 5 0.02. Among ears with asymmetrically smaller cochlear nerve canals, a 1mm reduction in cochlear canal width between the normal and affected ear was associated with 30.4% lower word recognition score percentage in the affected ear, P 5 <0.001.
Conclusion: There is a significant association between BCNC stenosis and impaired speech discrimination, independent
of degree of hearing loss. Further investigation is needed to determine whether BCNC stenosis is a poor prognostic factor for
auditory rehabilitation.
Key Words: Unilateral hearing loss, temporal bone, computed tomography, cochlear nerve deficiency.
Level of Evidence: 3b.
Laryngoscope, 125:1691–1696, 2015
INTRODUCTION
Congenital sensorineural hearing loss (SNHL) has
an estimated prevalence of approximately one to two per
1,000 newborns, although estimates vary and may be
higher.1 High-resolution computed tomography (CT)
imaging is often used to investigate the etiology of
SNHL due to its relatively low cost and excellent resolution of osseous temporal bone structures.2
Bony cochlear nerve canal (BCNC) stenosis is a
temporal bone anomaly associated with congenital
SNHL that is easily assessed on high-resolution CT. The
From the Department of Otolaryngology (P.L.P., K.C.Y.S., D.L.H.); the
Department of Speech and Hearing Sciences (D.L.H.); the Department of
Radiology (G.S.P., A.M.P.), University Of Washington; the Department of
Otolaryngology, Seattle Children’s Hospital (K.C.Y.S., D.L.H.), Seattle,
Washington; and the Department of Otolaryngology, Henry Ford Health
System (A.J.I.), Detroit, Michigan, U.S.A.
Editor’s Note: This Manuscript was accepted for publication
November 17, 2014.
Preliminary results from this study were presented as a poster for
American Society of Pediatric Otolaryngology at the Combined Otolaryngological Spring Meeting, Las Vegas, Nevada, U.S.A. May 14–18, 2014.
This study was supported by the Institutional National Research
Service Award for Research Training in Otolaryngology (grant
2T32DC000018) from the National Institute on Deafness and Other
Communication Disorders. The authors have no other funding, financial
relationships, or conflicts of interest to disclose.
Send correspondence to David L. Horn, MD, 4800 Sand Point Way
NE, Seattle, WA 98105. E-mail: [email protected]
DOI: 10.1002/lary.25087
Laryngoscope 125: July 2015
BCNC encases the cochlear nerve from the fundus of the
internal auditory canal (IAC) to the base of the cochlear
modiolus; therefore, it has been suggested that BCNC
stenosis serves as a marker of anatomic or functional
deficiency of the cochlear nerve.3,4 Several studies have
described an association between stenosis of the bony
cochlear nerve canal and congenital SNHL.5,6
Children with cochlear nerve deficiency based on
magnetic resonance imaging (MRI) have been found to
have hearing characteristics typical of auditory neuropathy,7 which is associated with poor speech discrimination
out of proportion to the degree of hearing loss. However,
the relationship between BCNC stenosis and speech discrimination has yet to be well characterized. If BCNC
stenosis is indeed a sign of cochlear nerve hypoplasia,
then we hypothesize that subjects with stenosis will
demonstrate poor speech discrimination independent of
the degree of hearing loss.
There are several reasons to study this question in
children with unilateral hearing loss. Recent studies
have found that children with unilateral SNHL have
higher rates of BCNC stenosis than children with bilateral hearing loss,8 and accurate description of the phenotype may facilitate further investigation of the etiology
and have ramifications for amplification approaches. In
addition, the subjects can serve as their own matched
control population when comparing measurements. The
objective of this investigation is to evaluate children
Purcell et al.: Bony Cochlear Nerve Canal Stenosis
1691
Fig. 1. Axial temporal bone computed tomography images of 8-year-old male with normal
BCNC (A) and stenotic BCNC (B), as designated by arrows. BCNC 5 bony cochlear
nerve canal.
with unilateral hearing loss to further confirm the correlation between BCNC stenosis and the severity of hearing loss and to determine whether BCNC width is
significantly correlated with speech discrimination independent of hearing threshold.
MATERIALS AND METHODS
This study received institutional review board approval at
Seattle Children’s Hospital (Seattle, Washington), a pediatric
tertiary care facility. The hospital’s audiogram database was
then queried to identify all children with unilateral SNHL from
January 2007 to July 2013. Unilateral hearing loss was defined
as bone-conduction pure tone average (PTA) thresholds 30 dB
SPL in the affected ear with thresholds 20 dB SPL in the normal ear. Children who had acquired hearing loss, such as temporal bone fracture, were excluded. Medical record numbers
were then cross-referenced to a radiologic database to identify
those children who had undergone high-resolution CT temporal
bone imaging with slice thickness of < 1 mm. For children with
multiple audiograms, the hearing threshold was calculated from
the audiogram performed closest to the date of the CT scan.
Measurements were made on axial images using a picture
archiving and communications system by two radiologists who
were both blinded to the affected ear. The plane of the axial
images was parallel to the infraorbitomeatal line. The images
were reconstructed by using a high spatial-resolution bone algorithm with individual magnification for the right and left temporal bones of approximately 103 (see Fig. 1). The width of the
canal was obtained at its midportion along the inner margin of
its bony walls, using a method similar to that described by Fatterpekar et al.9 The measurements were manually obtained by
using calipers and were calculated to the nearest 0.1 mm.
Measurement similarity between radiologists was examined using the Pearson correlation coefficient. The degree of
BCNC stenosis was determined using both the absolute width
of the BCNC in the affected ear and the difference in BCNC
width between normal and affected ears. The difference was calculated by subtracting the BCNC measurement of the abnormal
ear from the measurement of the normal ear. For example, subjects with equal BCNC widths between ears were calculated to
have a difference value equal to zero, whereas subjects with stenosis had a narrower BCNC width in the affected ear, and
therefore a positive difference score.
Analyses
After measurements from the two radiologists were averaged, mean BCNC widths of normal and affected ears were
compared using paired t tests. Receiver operating characteristic
(ROC) curves were generated by comparing the affected ears
Laryngoscope 125: July 2015
1692
to normal ears to determine the specificity and sensitivity for
measurement values. The BCNC cutoff width was investigated
as an etiologic factor for known hearing loss, not a screening
instrument. In order to determine the measurement below
which no ears had normal hearing, efforts were therefore made
to maximize the specificity rather than the sensitivity when
establishing cutpoint recommendations.
Linear regression was used to describe associations
between absolute measurement of BCNC width, PTA threshold,
and word recognition score (WRS) percentage. Stepwise regression modeling evaluated the significance of adjusting for hearing threshold and adjusting for the level above threshold at
which speech discrimination testing was performed. Analyses
also investigated whether there were any significant differences
in results with use of three- or four-frequency PTA thresholds
or variation in speech discrimination instrument types. Final
analysis models utilized three-frequency PTA thresholds (500,
1000, 2000 Hz) and included all children with available WRS
percentage, regardless of instrument type.
Linear spline regression10 was used to investigate association between WRS percentage and difference in BCNC width
between the normal and affected ears. Relative to linear regression, linear spline analysis is a more detailed modeling technique to allow for changes in the slope of association. A knot
can be set at points where relationship is expected to change
direction, such as at a maximum peak with a downtrending
association on either side. When plotting the association
between the WRS percentage and difference in BCNC width
between normal and affected ears, a knot was set a priori at the
BCNC difference of zero because it was anticipated that subjects with equal, or symmetric, BCNC measurements would
have highest the WRS percentage.
For all analyses, statistical significance was set at a P
value of 0.05. Analyses were performed using STATA version
13.1 (STATA, Inc, College Station, TX).
RESULTS
The audiometric database query identified 341 children with unilateral hearing loss. After the exclusion of
children with acquired hearing loss, there were 160 children who had unilateral sensorineural hearing loss
(USNHL) and available temporal bone CT imaging, for a
total of 320 ears. The mean age at hearing loss diagnosis
was 4.8 years (standard deviation [SD] 5 3.7) (see
Table I). The group was evenly divided by sex, with 80
females and 80 males. Of the 160 children, 70 had
undergone speech discrimination testing. The majority
of children included in the study, 84 of 160 (53%), had
been diagnosed with severe-profound unilateral SNHL,
but the majority of subjects who underwent speech
Purcell et al.: Bony Cochlear Nerve Canal Stenosis
TABLE I.
Ages of Children With USNHL at Time of Diagnosis, CT Imaging and Speech Discrimination Testing.
Mean Age in Years of All Children, (Range)
Age at diagnosis
Age at CT imaging
Age at speech discrimination testing
Mean Age in Years of Children With WRS (Range)
4.8 (birth to 16 years)
5.5 (birth to 16 years)
7.3 (2 months to 20 years)
–
8.1 (1 month to 20 years)
9.7 (3 to 21 years)
CT 5computed tomography; USNHL 5 unilateral sensorineural hearing loss; WRS 5 word recognition score.
discrimination testing had moderate HL, 50 of 70 (71%)
(see Table II).
All of the children had ear-specific PTAs, and all
had documentation of an attempt at speech reception
threshold testing, although 60 children were unable to
complete the testing due to the degree of hearing loss.
Of the 70 children who had WRS percentages, 57 were
tested using open-set instruments, seven using closedset instruments, and six did not have a description of
instrument type. The open-set instruments included the
following: 29 children were tested using Northwestern
University Auditory Test No. 6; 14 with Phonetically
Balanced Kindergarten; 11 with Central Institute for the
Deaf W-22; and three others underwent open-set testing,
but the instrument type was unspecified.
Measurements from the two radiologists were compared and found to be very highly correlated, Pearson
correlation coefficient 0.83, P value < 0.001. After confirming a strong correlation, the measurements from
both radiologists were averaged. Sample mean BCNC
width in the affected ears was 1.9 mm (SD 5 0.6) and in
normal ears was 2.1 mm (SD 5 0.32). The difference was
significant, P value < 0.0001.
ROC curves were generated by comparing BCNC
widths of the affected ears against those of normal ears
(see Fig. 2). Area under the curve (AUC) was determined
to be 0.58, 95% confidence interval (CI) (0.52 to 0.65).
Sensitivity and specificity cutpoints demonstrated the
greatest likelihood of correct classification of hearing
status between the affected and unaffected ears at
1.4 mm, which corresponded to specificity of 99% and
sensitivity of 17% (see Table III). The sensitivity of 17%
is the proportion of congenital USNHL attributable to
BCNC stenosis in this sample of patients using a cutpoint of 1.4 mm. The majority of affected ears did not
have BCNC stenosis, but SNHL due to alternative etiology. Using a 1.4-mm cutpoint, 27 (17%) of the affected
ears in this study would be diagnosed with BCNC stenosis. There were two normal hearing ears (1%) with a
BCNC width of 1.4 mm; no normal hearing ears were
smaller than 1.4 mm.
Additional structural temporal bone abnormalities
were also identified. Of the 27 patients with BCNC
width of 1.4 mm or smaller, seven also had narrow IAC
and two had cochlear dysplasia or malformation. Of the
18 children with enlarged vestibular aqueduct, none had
a BCNC width of 1.4 mm or smaller.
Stepwise linear regression evaluated the association
between BCNC width, PTA threshold, and WRS percentage. A univariate linear regression model that included
Laryngoscope 125: July 2015
all 160 children found the severity of stenosis to be
associated with degree of HL, P value 0.04; however, correlation was weak, r 5 20.15.
Multiple linear regression was performed to characterize the association between BCNC width and WRS
percentage among the 70 children with WRS results
while adjusting for PTA threshold. The level above the
threshold at which speech discrimination testing was
performed was explored as a potential covariate, but it
was not significant. Results were similar whether the
analysis adjusted for three-frequency or four-frequency
PTA threshold and whether it included only open versus
both open- and closed-set instruments for speech discrimination testing.
A multiple linear regression model utilizing an
absolute measurement of BCNC width found an association between BCNC width and WRS percentage that
approached significance, P value 0.06 (see Fig. 3) with a
higher degree of correlation, r 5 0.57, than univariate
model containing only the BCNC width and hearing
threshold.
Several children were noted to have unusually wide
BCNC measurements in their affected ears when compared to their unaffected ears. Specifically, there were
three children who had a bony cochlear nerve canal in
the affected ear that was greater than 0.6-mm wider
than the unaffected ear.
In order to better evaluate the effect of stenosis,
analysis was repeated using only the 91 children who
displayed smaller cochlear canal widths in their affected
ears compared to their unaffected ears. Among these
children, severity of stenosis was associated with degree
of HL, P value 0.03, but the correlation was again weak,
r 5 20.2.
Multiple linear regression was then performed to
characterize the association between BCNC width and
WRS percentage among the 37 children who had available WRS percentage and smaller cochlear nerve canals
in their affected ears. We identified an association
between BCNC width and WRS percentage that was significant, P value 0.03, with a strong correlation of
r 5 0.61.
In order to analyze all the children, including those
who did not have a smaller cochlear nerve canal in their
affected ears, we explored the hypothesis that BCNC
asymmetry is associated with WRS percentage. When
WRS percentage was replotted as a function of BCNC
difference score, speech perception appeared to fall off
with increasing asymmetry in either direction from zero.
A linear spline model was created by placing a single
Purcell et al.: Bony Cochlear Nerve Canal Stenosis
1693
TABLE II.
Degree of Hearing Loss and Mean BCNC Width of Children With USNHL.
Total
Degree of HL
Mild (30 to 40 dB)
Number of Children
With Imaging (%)
Mean BCNC Width in mm (SD),
Affected Ear
160 (100%)
1.9 (0.6)*
Mean BCNC Width
in mm (SD), Normal Ear
2.1 (0.32)*
Number of Children
With WRS (%)
70 (100%)
10 (6%)
1.93 (0.13)
1.95 (0.19)
7 (10%)
Moderate–moderately
severe (41 to 70 dB)
66 (41%)
2.06 (0.46)
2.08 (0.36)
50 (71%)
Severe–Profound (711 dB)
84 (53%)
1.78 (0.71)**
2.12 (0.3)**
13 (19%)
*Significant difference, P value 0.0001.
**Significant difference, P value <0.0001.
BCNC 5 bony cochlear nerve canal; SD 5 standard deviation; USNHL 5 Unilateral sensorineural hearing loss; WRS 5 word recognition score.
knot at the point where the difference in BCNC width
between the normal and abnormal ears was equal to
zero.
After adjustment for hearing threshold, the model
confirmed a peak in WRS percentage at difference measurement of zero, which is the point of equal measurement, or symmetry, between normal and affected ears
(see Fig. 4). A likelihood ratio test provided evidence for
an overall association between WRS and BCNC asymmetry, P value 0.021.
For stenotic ears, a 1-mm difference in BCNC
width between normal and abnormal ears was associated with 30.4 percentage point lower WRS, 95% CI
(246.5 to 214.3), P value < 0.001. Correlation was
quite strong, r 5 20.62. Similarly, children who had a
1-mm wider BCNC in the affected ear compared to the
normal hearing ear had a 37.9 percentage point lower
WRS, 95% CI (23.2 to 79.1). This portion of the spline
regression model did not reach significance, P value
0.07.
DISCUSSION
Cochlear nerve deficiency has been described as
occurring in up to 18% of ears with congenital SNHL.11
However, diagnosis of cochlear nerve deficiency requires
MRI, which may be more costly than CT and may
require general anesthesia in children of a young age at
some institutions. Clemmens et al. recently described
the correlation between BCNC stenosis seen on CT and
cochlear nerve deficiency demonstrated on MRI, suggesting that CT can be a reasonable test for cochlear nerve
deficiency in the workup of pediatric SNHL.12
CT imaging is of particularly high yield for children
with unilateral hearing loss. Previous studies have
reported that temporal bone anomalies occur in about
one-third of children with unilateral impairment,13 and
BCNC stenosis is among the most common anomalies
identified.14 There is also suggestion that, among children with stenotic BCNC, those with unilateral hearing
loss may be more likely to have cochlear nerve deficiency
in comparison to those with bilateral hearing loss.15
In our study, 17 percent of subjects met criteria for
stenosis using a cutpoint of 1.4 mm, which has been previously described16 and is slightly smaller than cutpoints
of 1.5 or 1.7 mm that have also been noted in the literature.17,18 Using a 1.4-mm cutpoint, only 1% of normalhearing ears in our study would be classified as stenotic.
Therefore, 1.4 mm appears to be a cutoff point below
which cochlear nerve dysfunction became almost a certainty in this population.
Previous studies have found that the severity of
BCNC stenosis correlates with the degree of hearing
TABLE III.
Sensitivity and Specificity for SNHL Associated With BCNC
Measurements.
BCNC Width, mm
Fig. 2. ROC analysis estimates sensitivity and specificity for diagnosing BCNC stenosis in a child with unilateral SNHL. A measurement of 1.4 mm corresponds with point A labeled on the curve.
This is the point that maximizes specificity, indicating that almost
all BCNC measurements as narrow, or narrower, will be associated with SNHL. BCNC 5 bony cochlear nerve canal;
ROC 5 receiver operating characteristic; SNHL 5 sensorineural
hearing loss. [Color figure can be viewed in the online issue,
which is available at www.laryngoscope.com.]
Laryngoscope 125: July 2015
1694
Sensitivity
Specificity
2.2 mm
74%
32%
2.0 mm
57%
54%
1.8 mm
1.6 mm
33%
20%
81%
96%
1.4 mm
17%
99%
BCNC 5 bony cochlear nerve canal; SNHL 5 sensorineural hearing
loss.
Purcell et al.: Bony Cochlear Nerve Canal Stenosis
Fig. 3. Figure represents a multiple linear regression model evaluating the association between BCNC width and WRS percentage
for affected ears among children with unilateral SNHL. The model
controls for hearing threshold, as noted by the four different lines
corresponding to progressively increasing PTA level. The association between BCNC width and WRS percentage approached significance (P value 0.064, with a moderate degree of correlation,
r 5 0.57). BCNC 5 bony cochlear nerve canal; PTA 5 pure tone
average; WRS 5 word recognition score.
loss.19 Our results are consistent with these findings;
however, perhaps more interestingly, our results support
the hypothesis of a stronger correlation between BCNC
stenosis and impaired speech discrimination, which is to
be expected if BCNC abnormalities serve as a marker
for cochlear nerve dysfunction. The results also suggest
that, in children with unilateral hearing impairment, an
alternative method for identifying a BCNC abnormality
could involve calculating a difference measurement
between normal and abnormal ears. At least one previous study has compared BCNC width among ears with
normal hearing and found no significant differences in
measurement.20
There are several limitations to this study. Because
it is retrospective, there was variability in age at testing,
degree of follow-up, consistency of radiological imaging,
and methods of audiometric evaluation. In addition,
audiometric test results may fluctuate over time. When
possible, we have attempted to control for sources of
potential variability in the analysis; in most cases, the
differences were found to not to significantly influence
final results. Another limitation relates to the small
number of children with unusually wide BCNC in the
affected ear. It was not feasible to analyze them alone;
however, the spline model was developed in order to
incorporate them into the investigation.
It may be worthwhile to confirm whether enlargement of the BCNC in the affected ear of unilateral SNHL
patients represents an inner ear anomaly. The relationship detected in this investigation may have been carried
Laryngoscope 125: July 2015
by a small number of outliers and needs to be followed
up by further research. Bulbous IAC is a radiographic
finding described in previous studies,21 and case series of
children with SNHL and abnormal widening of the IAC
have been described.22 However, an association between
the enlarged BCNC and SNHL has yet to be established.
Currently, children with BCNC stenosis are not
managed differently than children with SNHL without
stenosis. Because children with poor speech discrimination tend to experience limited benefit from amplification
of the involved ear, a child with BCNC stenosis might be
a better candidate for a contralateral routing of signals
hearing system.
In addition, BCNC stenosis or MRI finding of cochlear nerve deficiency is controversial as a contraindication to cochlear implantation in patients with bilateral
hearing loss. Papsin investigated cochlear implantation
outcomes among children with various cochleovestibular
anomalies and found that children with narrowing of
the IAC and/or cochlear canal performed more poorly
than all other groups.23 However, there have been case
reports reporting some benefit to implantation in children with cochlear nerve deficiency.24 Although cochlear
implantation is not commonly performed for unilateral
hearing loss, future investigations could explore whether
BCNC stenosis is a poor prognostic factor for auditory
rehabilitation.
Fig. 4. Figure represents a linear spline regression model evaluating the association between the difference in BCNC width
between normal and affected ear and WRS percentage among
children with unilateral sensorineural hearing loss. At a difference
score of zero, there is no difference in BCNC width between the
two ears. Positive values indicate BCNC width of affected ear
smaller than normal ear; negative values indicate BCNC width of
affected larger than normal ear. The model controls for hearing
threshold. Asymmetry of BCNC width in either direction is significantly associated with poorer speech discrimination, regardless of
hearing threshold, P value 0.021. BCNC 5 bony cochlear nerve
canal; PTA 5 pure tone average; WRS 5 word recognition score.
Purcell et al.: Bony Cochlear Nerve Canal Stenosis
1695
CONCLUSION
BCNC stenosis is significantly associated with the
impairment of speech perception in children with unilateral hearing loss. Such a result is important when considering the options for management, and families
should be counseled accordingly. Future investigation of
the relationship between BCNC abnormalities and aided
speech perception is needed.
Acknowledgments
The authors would like to acknowledge Kenneth M. Rice,
PhD; Scott S. Coggeshall; and Jing Fan of the University of
Washington Biostatistics Department for providing statistical support for this project.
BIBLIOGRAPHY
1. Morton CC, Nance WE. Newborn hearing screening–a silent revolution.
NEJM 2006;354:2151–2164.
2. Pagarkar W, Gunny R, Saunders DE, Yung W, Rajput K. The bony cochlear nerve canal in children with absent or hypoplastic cochlear nerves.
Int J Pediatr Otorhinolaryngol 2011;75:764–773.
3. Miyasaka M, Nosaka S, Morimoto N, Taiji H, Masaki H. CT and MR imaging for pediatric cochlear implantation: emphasis on the relationship
between the cochlear nerve canal and the cochlear nerve. Pediatr Radiol
2010;40:1509–1516.
4. Yan F, Li J, Xian J, Wang Z, Mo L. The cochlear nerve canal and internal
auditory canal in children with normal cochlea but cochlear nerve deficiency. Acta Radiol 2013;54:292–298.
5. Adunka OF, Jewells V, Buchman CA. Value of computed tomography in
the evaluation of children with cochlear nerve deficiency. Otol Neurotol
2007;28:597–604.
6. Wilkins A, Prabhu SP, Huang L, Ogando PB, Kenna MA. Frequent association of cochlear nerve canal stenosis with pediatric sensorineural hearing loss. Arch Otolaryngol Head Neck Surg 2012;138:383–388.
7. Buchman CA, Roush PA, Teagle HF, Brown CJ, Zdanski CJ, Grose JH.
Auditory neuropathy characteristics in children with cochlear nerve
deficiency. Ear Hear 2006;27:399–408.
8. Nakano A, Arimoto Y, Matsunaga T. Cochlear nerve deficiency and associated clinical features in patients with bilateral and unilateral hearing
loss. Otol Neurotol 2013;34:554–558.
Laryngoscope 125: July 2015
1696
9. Fatterpekar GM, Mukherji SK, Alley J, Lin Y, Castillo M. Hypoplasia of
the bony canal for the cochlear nerve in patients with congenital sensorineural hearing loss: initial observations. Radiology 2000;215:243–246.
10. Boucher KM, Slattery ML, Berry TD, Quesenberry C, Anderson K. Statistical methods in epidemiology: a comparison of statistical methods to
analyze dose-response and trend analysis in epidemiologic studies.
J Clin Epidemiol 1998;51:1223–1233.
11. McClay JE, Booth TN, Parry DA, Johnson R, Roland P. Evaluation of pediatric sensorineural hearing loss with magnetic resonance imaging. Arch
Otolaryngol Head Neck Surg 2008;134:945–952.
12. Clemmens CS, Guidi J, Caroff A, et al. Unilateral cochlear nerve deficiency in children. Otolaryngol Head Neck Surg 2013;149:318–325.
13. Song JJ, Choi HG, Oh SH, Chang SO, Kim CS, Lee JH. Unilateral sensorineural hearing loss in children: the importance of temporal bone computed tomography and audiometric follow-up. Otol Neurotol 2009;30:
604–608.
14. Masuda S, Usui S, Matsunaga T. High prevalence of inner ear and/or
internal auditory canal malformations in children with unilateral sensorineural hearing loss. Int J Pediatr Otorhinolaryngol 2013;77:228–
232.
15. Cho SW, Kang SI, Park SJ, et al. Clinical characteristics of patients with
narrow bony cochlear nerve canal: is the bilateral case just a duplicate
of the unilateral case? Laryngoscope 2013;123:1996–2000.
16. Stjernholm C, Muren C. Dimensions of the cochlear nerve canal: a radioanatomic investigation. Acta Otolaryngol 2002;122:43–48.
17. Hidenobu T, Morimoto N, Matsunaga T. Unilateral cochlear nerve hypoplasia in children with mild to moderate hearing loss. Acta Otolaryngol
2012;132:1160–1167.
18. Kono T. Computed tomographic features of the bony canal of the cochlear
nerve in pediatric patients with unilateral sensorineural hearing loss.
Radiat Med 2008;26:115–119.
19. Yi JS, Lim HW, Kang BC, Park SY, Park HJ, Lee KS. Proportion of bony
cochlear nerve canal anomalies in unilateral sensorineural hearing loss
in children. Int J Pediatr Otorhinolaryngol 2013;77:530–533.
20. Jang JH, Kim J-H, Yoo JC, et al. Implication of bony cochlear nerve canal
on hearing in patients with congenital unilateral sensorineural hearing
loss. Audiol Neurotol 2012;17:282–289.
21. McClay JE, Tandy R, Grundfast K, Choi S, Vezina G, Zalzal G, Willner A.
Major and minor temporal bone abnormalities in children with and
without congenital sensorineural hearing loss. Arch Otolaryngol Head
Neck Surg 2002;128:664–671.
22. Santos S, Dominguez MJ, Cervera J, Suarez A, Bueno A, Bartolome M,
Lopez R. Hearing loss and enlarged internal auditory canal in children.
Acta Otorrinolaringol Esp 2014;65:93–101.
23. Papsin BC. Cochlear implantation in children with anomalous cochleovestibular anomaly. Laryngoscope 2005;115(suppl 106):1–26.
24. Vincenti V, Ormitti F, Ventura E, Guida M, Piccinini A, Pasanisi E. Cochlear implantation in children with cochlear nerve deficiency. Int J
Pediatr Otorhinolaryngol 2014;78:912–917.
Purcell et al.: Bony Cochlear Nerve Canal Stenosis