Developmental Auditory and Speech-Language Performance in Pediatric Cochlear Implantation Recipients with Stable White Matter Lesions

Abstract

To analyze the association between stable asymptomatic white matter lesions (WMLs) and the cochlear implantation (CI) effect in congenitally deaf children, 43 CI children with stable asymptomatic WMLs determined via preoperative assessments and 86 peers with normal white matter were included. Outcome measurements included closed-set Mandarin Chinese (tone, disyllable, and sentence) recognition tests; categories of auditory performance (CAPs); and speech intelligibility rating (SIR) scales at 1, 12, and 24 months post-CI. Generalized estimating equation (GEE) models were used to analyze the association between WML and outcomes. In the WML group (control group), median CAP and SIR scores were 5 (5) and 4 (4) with mean rates of tone, disyllable, and sentence recognition of 84.8% (89.0%), 87.9% (89.7%), and 85.8% (88.0%) at 24 months post-CI, respectively. Auditory and speech performance improved significantly with implant use. Compared to their peers in the control group, for the participants with stable asymptomatic WMLs, auditory and speech abilities were not significantly different (p > 0.05). Stable asymptomatic WMLs might not be associated with poor auditory and speech intelligibility post-CI, which indicates that it is feasible to use comprehensive assessments to screen suitable candidates with WMLs who are likely to present with a good prognosis.

Keywords: CAP; SIR; cochlear implantation; pediatric; speech recognition; white matter lesions.

Figure 1

Axial, T2-FlAIR cerebral MRI scans six months apart of the individuals. Short black arrows point to lesions in periventricular white matter. Black arrows point to lesions in deep white matter. (A1) The first MRI scan for a 18-month-old boy showed extensive WMLs (PVH = 3, DWMH = 2, Fazekas score = 5). (A2) The second scan for the same child (24 months of age) showed improvement in WML with the Fazekas score declining (PVH = 2, DWMH = 2, Fazekas score = 4). (B1,B2) Two MRI scans of a 31-month-old (date of the second MRI scan) boy showed no obvious change in diffuse WMLs (PVH = 3, DWMH = 3, Fazekas score = 6).

Figure 2

Axial, T2-FlAIR cerebral MRI scans with annotated Fazekas scores. (A) Multifocal WMLs in periventricular white matter (short black arrows, PVH = 1, Fazekas score = 1). (B) Multifocal WMLs in deep white matter (black arrows) of the frontal and parietal lobes (DWMH = 2, Fazekas score = 2). (C) Extensive WMLs in periventricular white matter (short black arrows) and deep white matter (black arrows) of the frontal and parietal lobes (PVH = 3, DWMH = 2, Fazekas score = 5). (D) Diffuse WMLs in periventricular white matter (short black arrows) and deep white matter (black arrows) of whole brain areas (PVH = 2, DWMH = 3, Fazekas score = 5).

Figure 3

Axial, T2-FlAIR cerebral MRI scans with annotated involved areas. (A) The frontal and occipital lobes involvement (black arrows). (B) Periventricular white matter involvement (short black arrows). (C) The frontal and parietal lobes involvement (black arrows) together with periventricular white matter involvement (short black arrows). (D) The temporal, frontal and parietal lobes involvement (black arrows).

Figure 4

Association of WML with Mandarin Chinese tone (A), disyllable (B), and short sentence (C) recognition outcomes. The point estimate and 95% confidence interval (95% CI) for the coefficients with WML are presented for an unadjusted model (model 1), a model aiming to analyze the interaction between WML and time (model 2), and another model adjusted for covariates (model 3). The independent variables in model 2 were WML, time after device activation, and the interaction between WML and time after device activation. Model 3 was adjusted for time, age at implantation, parental literacy, residual hearing, experience of rehabilitation pre-CI, and MMAIS/IT-MMAIS score pre-CI.

Figure 5

Association of WML with CAP (A) and SIR (B) outcomes. The point estimate and 95% confidence interval (95% CI) for the odds ratios (ORs) with WML are presented for an unadjusted model (model 1), a model aiming to analyze the interaction between WML and time (model 2), and another model adjusted for covariates (model 3). The independent variables in model 2 were WML, time after device activation, and the interaction between WML and time after device activation. Model 3 was adjusted for time, age at implantation, parental literacy, residual hearing, experience of rehabilitation pre-CI, and MMAIS/IT-MMAIS score pre-CI.