Abstract
OBJECTIVE
The goals of this study were to retrospectively review high-resolution CTs (HRCTs) of pediatric postmeningitic cochlear implant recipients and to correlate results with surgical findings.
METHODS
HRCTs of 20 children (11 months to 12 years old) who underwent implantation with multichannel devices were reviewed. Results were correlated with the degree of ossification observed at surgery.
RESULTS
Ninety percent of subjects required drilling of ossified bone within the basal turn at surgery. HRCT of the cochleas suggested ossification within the basal turn in 45% (50% sensitivity). Ossification of the lateral semicircular canal on HRCT was present in 72% (77% sensitivity). Five of 6 cases without radiographic evidence of ossification had positive findings at surgery.
CONCLUSION
Ossification is a common occurrence in postmeningitic deaf children. Ossification of the lateral semicircular canal on HRCT is a more sensitive measure for predicting ossification than evidence of cochlear involvement. Absence of ossification on HRCT is no guarantee of cochlear patency at the time of implantation.
Radiologic study of the temporal bone is of special importance in children deafened by meningitis who are being evaluated for cochlear implantation. Ossification of the cochlea is likely to develop in children deafened by meningitis, and it can render implantation technically more difficult and increase the likelihood of partial insertion of the electrode array. The presence of ossification in a recently deafened child is evidence that cochlear implantation should proceed on an urgent basis. The degree of ossification can influence the choice of ear of implantation, surgical approach, and cochlear implant system. It is also important for optimal counseling of families, not only in regard to the surgical procedure but also in regard to the potential for device-programming problems and short- or long-term lack of response to auditory stimuli.
A number of investigators have reviewed the issue of cochlear patency as determined by high-resolution CT (HRCT). 1 - 5 These reports usually focus on the apical, intermediate, and basal turns of the cochlea, as well as the round and oval windows. Radiologic classification systems proposed to quantify the degree of ossification also focus on these areas, which are of technical significance during implant surgery. 3 , 4 However, Muren and Bredberg 6 suggested that assessment of cochlear implant candidates deafened by meningitis should include radiographic analysis of the semicircular canals (SCCs) because their involvement may herald the development of cochlear ossification.
Since 1991, at the Children's Memorial Medical Center, we have considered the status of the SCCs in the radiologic evaluation of cochlear implant candidates. Our experience suggests that the presence of even isolated involvement of the SCCs, especially the lateral SCC (LSCC), may be evidence of ossification of clinical significance (Naidich MJ, et al. CT in the evaluation of children with postmeningitic sensorineural hearing loss for cochlear implant. AJNR Am J Neuroradiol. Submitted for publication). It was the purpose of this study to review our radiologic temporal bone findings and determine their relationship to the presence and extent of ossification found at surgery.
METHODS AND MATERIAL
All children who underwent implantation at Children's Memorial Medical Center between January 1991 and August 1998 who were deafened by meningitis and for whom preoperative HRCT scans were available were included in this study. Twenty children met these criteria: 12 boys and 8 girls, ranging in age from 11 months to 12 years, with a mean age of 4.7 years at time of implantation. Five children received the Nucleus 22, 12 the Clarion, 1 the Nucleus 24, and 1 the COMBI 40+ with split electrode array. Two patients had partial insertion of the active electrodes (Nucleus and Clarion). The right ear was implanted in 11 (55% of patients), and the left in 9 (45%). The infectious agents causing meningitis are listed in Table 1.
70236-1-fig1.png)
Axial HRCT of the basal turn of the cochlea:
Infectious agent causing meningitis
70236-1-table1.png)
For children who had more than 1 HRCT before implantation, the scan obtained closest to surgery was used. All HRCT studies were performed on a General Electric (Milwaukee, WI) High Speed Advantage (1994 to present) or 9800 (before 1994) CT scanner with a bone algorithm technique. Seventeen were performed with 1-mm slice thickness, and the 3 studies performed before 1994 had 1.5-mm slice thickness. All studies were performed in a true axial projection, but only 17 patients had coronal projections.
The following structures of each ear were evaluated in the axial and coronal projections: the basal, intermediate, and apical turns of the cochlea; the round window (RW); the LSCC, posterior SCC (PSCC), and superior SCC (SSCC); and the vestibule. A grading system (Figs 1 and 2) was used to denote the degree of involvement of ossification within each of these structures: 0, normal appearance; 1, sclerosis (or increased density) without narrowing; 2, sclerosis (or increased density) with narrowing; and 3, partial or complete obliteration. The studies were read by 2 experienced neurora-diologists (S.E.B., C.D.) who were blinded to the surgical results, with a concurrence of 95%. Discrepancies in interpretation were resolved by a consensus reading.
The ear that was less involved by ossification was chosen for implantation. If there was no difference between ears, the parent chose the ear for implantation. A canal wall up mastoidectomy with facial recess approach was performed. The RW niche was drilled away to allow direct visualization of the membrane. A cochleostomy was created anterior to the RW.
The patient who received the COMBI 40+ with split electrode array had a modified procedure that involved creation of a superior tunnel, as described by Muren and Bredberg. 6 During surgery, the extent of ossification and/or fibrosis of the RW (normal, thickened, ossified) and the basal turn of the cochlea (millimeters drilled) was assessed, and this was routinely recorded on the patient chart.
70236-1-fig2.png)
Axial HRCT of the LSCC:
Surgical findings
70236-1-table2.png)
∗Diffuse narrowing noted distally.
DATA ANALYSIS
Data were entered into a statistical program (SPSS 8.0) for analysis. Descriptive statistics, including cross-tabulation tables, were produced. However, because of the small cell sizes in many of the bivariate analyses, inferential statistics were not always performed. For correlational analyses, the nonparametric Spearman rho (ρ) was computed and tested for statistical significance; significance was set at P < 0.05, 2-tailed.
RESULTS
The mean age at which the children had meningitis was 2.8 years (SD 2.6; range 1.6 months to 8.5 years; median 1.8 years). The mean time between meningitis and the HRCT was 21.7 months (SD 30.9; range 1-116 months; median 8.8 months). The mean time that elapsed between the HRCT and surgery was 1.6 months (SD 0.9; range 3 days to 4.2 months; median 1.6 months).
CT findings: Axial views
70236-1-table3.png)
CT findings: Coronal views (n = 17)
70236-1-table4.png)
Surgical findings are presented in Table 2. Fifteen subjects (75%) were noted to have thickened or ossified RW membranes. Eighteen subjects (90%) had some degree of ossification that required additional drilling within the basal turn of the cochlea. Four of the 5 with a normal-appearing RW membranes required drilling of ossified bone from within the basal turn. No significant correlation was found between the time from meningitis to implantation and the degree of ossification noted at surgery (ρ = 0.11).
The HRCT findings for the implanted ear are presented in Tables 3 and 4. There was no relationship between the radiographic status of the RW and surgical findings. Of the 11 cases in which the RW was reported to be normal on both axial and coronal views, only 1 was normal at surgery. Of the 6 RW cases graded as abnormal, 2 were judged to be normal at surgery.
Of the 17 patients in whom both axial and coronal views were available, 8 were graded as having no evidence of ossification of the cochlear turns; however, 6 (75%) of these required drilling of ossified bone within the basal turn, 2 of which required extensive drilling (≥8 mm). Of the 9 remaining subjects with radiographic evidence of ossification of the cochlea, all had ossification noted at implantation. The 3 patients in whom only axial views were available were all graded as having no evidence of cochlear ossification; however, in each case ossification was present at surgery. In total, of 11 patients who were without evidence of ossification of the cochlear turns on HRCT, only 2 (18%) had absence of ossification at the time of implantation.
Of all vestibular structures evaluated on HRCT, the LSCC most frequently demonstrated evidence of ossification. There was no evidence of ossification of the SSCC or PSCC when the LSCC appeared normal. Of the 18 subjects requiring drilling in the basal turn, 13 (72%) had abnormal LSCC findings, whereas only 4 (22%) had abnormal findings for the SSCC and PSCC. On the basis of axial views, 12 subjects showed no evidence of ossification involving the basal turn of the cochlea, but half of these did have a positive finding for the LSCC, 5 of whom had ossification at surgery.
The HRCT results were analyzed to determine whether findings on all 8 of the structures graded were predictive of the extent of ossification. Grades were totaled across the 8 structures for the axial and coronal findings, as well as across all 16 measures. No statistically significant correlations were found between severity of HRCT findings and the length of ossification drilled at surgery. In 6 patients in whom all structures in the implanted ear were graded 0, only 1 had absence of ossification of the basal turn; 4 of the remaining 5 had extensive ossification (≥8 mm) (Table 5). Only 1 of these 4 patients with extensive ossification had acquired meningitis years before implantation.
Table 6 summarizes results by showing the true-positive findings (sensitivity) for each of a number of radiologic measures, including across all measures. The LSCC was the most sensitive of the 8 sites, and a total across all sites did not improve sensitivity above that based on LSCC alone. If axial and coronal views are combined across all measures, the sensitivity of cochlear ossification was 50% (9/18), and the sensitivity of LSCC ossification was 77% (13/18).
DISCUSSION
In our series, 90% of postmeningitic deaf children had ossification requiring drilling of the inferior segment of the basal turn of the cochlea for placement of a cochlear implant. The finding of ossification of the RW on HRCT was not clinically useful for predicting surgical findings. HRCT of the turns of the cochlea suggested ossification within the basal turn in 45% of cases and had a sensitivity of 50%. HRCT of the LSCC suggested the presence of ossification in 70% of cases and had a sensitivity of 77%. In fact, of the 8 structures graded on HRCT, the LSCC was found to be the single most sensitive radiologic measure for the presence of cochlear ossification. One patient in whom no drilling was required had a finding of ossification isolated to the LSCC (grade 2), and this patient did have an ossified RW membrane at surgery.
Cases without HRCT evidence of ossification
70236-1-table5.png)
∗Only axial views available.
True-positive radiologic findings (sensitivity) ∗
70236-1-table6.png)
∗Gold standard is drilling in basal turn required at surgery.
†At least 1 positive finding on any of the 8 measures.
Although evaluation of the SCCs improved the accuracy of HRCT in predicting the presence of ossification, false-negative results did occur. Five cases had grade 0 preoperative HRCTs on all structures but required drilling of the basal turn (Table 5). In 2 of these, it is possible that lack of coronal views diminished the sensitivity of HRCT. Interestingly, this group with false-negative results included children deafened only 3 and 4 months before implantation but in whom extensive ossification was encountered. Therefore absence of HRCT evidence of ossification, even in children recently deafened by meningitis, is no guarantee of a fully patent cochlea.
The possibility of a false-negative HRCT and the lack of predictive value of HRCT in determining the extent of ossification that may be found at surgery need to be considered in the counseling and evaluation process before implantation. We approach all children with recently acquired postmeningitic deafness as requiring urgent evaluation for cochlear implantation, general health permitting. Our criteria for age at implantation and length of hearing aid trial are not rigidly applied to this special population.
Newer MRI technology and protocols can permit increased visualization of the membranous labyrinth.7-9 As a result, MRI may prove to be better at predicting the presence and extent of clinically significant ossification. However, MRI does not supplant CT because it cannot provide adequate information about the bony anatomy of the temporal bone that is needed before surgery. We agree with Frau et al 1 that routine use of MRI in the assessment of cochlear implant candidates is not necessary. In our experience, preoperative HRCT alone is usually adequate if the surgeon is prepared to address the issue of extensive ossification. Before surgery, we counsel all families of children with postmeningitic deafness about this possibility.
Several approaches to electrode insertion in the ossified cochlea have been described. 10 - 12 Early in our series, a partial insertion was performed in the rare circumstance in which drilling a tunnel through ossified bone within the inferior segment of the basal turn did not lead to a patent cochlea. More recently, we have begun using the COMBI 40+ (Med El Corp, Research Triangle Park, NC) with a modified split electrode array in cases in which ossification does not permit full electrode insertion subsequent to drilling of an inferior segment tunnel within the basal turn of the cochlea.
CONCLUSION
Ossification of the cochlea is a common occurrence in children deafened by meningitis and may in some cases progress rapidly. For this reason, the implant surgeon should be prepared to encounter ossification of varying degrees at surgery. In our series, the sensitivity of HRCT in identifying ossification of the cochlea was enhanced by evaluation of the SCCs. Patients with ossification of the LSCC were likely to have ossification within the basal turn, even if the cochlea appeared normal on HRCT.
We would like to thank Karen Berliner, Ph.D. for her advice and assistance in the statistical analysis of the data.