Predicting Performance in Adult Cochlear Implantation

点击翻译按钮获取中文摘要

Cochlear implants (Cl) can restore audibility, speech understanding, and communication abilities for adults with hearing loss receiving limited benefit from hearing aids, thereby improving quality of life. Despite the success of Cls, there remains wide variability in performance related to both patient and treatment factors. Biographical/audiological variables such as age, age at onset, duration, etiology, and severity of hearing loss as well as duration of hearing aid and Cl use together with surgical factors have long been used to predict speech perception performance. These variables account for only 10-20% of the variability in speech perception test scores in quiet, implying they are poor biomarkers of auditory system suitability for Cl stimulation. Recent studies demonstrate that Cl performance in background noise is worse in the elderly, in part, related to age-dependent central processing differences. We propose that in adult Cl candidates, accurate assessment of both the cochlear-neural and central auditory substrates can predict performance. Predicting performance has relevance to patient counseling and shared decision making (between clinician and patient), design and recommendations for auditory rehabilitation, consideration for device mapping and troubleshooting, and patient stratification in future clinical trials. Electrocochleography (ECochG) is an acoustically-evoked electrophysiological method to assess the cochlear-neural substrate; responses are present in >95% of Cl recipients, and can account for up to ~50% of open set speech perception scores in quiet. However, the ECochG-total response (ECochG-TR), is necessary but not sufficient to predict speech perception in noise. Adding a preoperative cognitive screener to ECochG-TR improves the prediction model (ΔR2=0.26; R2mode1=0.60). Thus, a good cochlear-neural substrate and cognitive function are both needed for speech understanding in noise. We will use biographical/audiological variables together with preoperative, transtympanic (ttECochG) and/or intraoperative (iECochG) ECochG and surgical factors, to develop clinically useful preoperative and/or postoperative (pre-activation) speech perception prediction models. Aim 1: Determine if preoperative ttECochG-TR is a valid measure of cochlear-neural substrate integrity as it relates to Cl stimulation, compared to the validated iECochG-TR. To this end, we will determine the strength of correlation between ttECochG and: (1a) iECochG-TR and (1 b) implant ear speech perception measures in quiet (6-month). Aim 2: Develop prediction models for implant ear speech perception scores in noise (Primary) and quiet (Secondary) after 6 months of Cl use. Models will include baseline demographic/audiologic variables, cognitive measures plus: (2a) Preoperative ttECochG-TR in the clinic; (2b) Postoperative (pre-activation) iECochG-TR and surgical factors from post-implant CT imaging. Aim 3: Establish the generalizability (external validation) of the prediction models in a geographically separate location (Vanderbilt University) from Washington University in St. Louis as patient and surgical factors can vary across sites.