Predictive models for cochlear implant outcomes: Performance, generalizability, and the impact of cohort size

Changes in participatory and societal outcomes during the waiting period for cochlear implantation - an observational study

INTRODUCTION

Various factors, including an aging population and expanding eligibility criteria, may increase the demand for cochlear implants (CIs), potentially resulting in longer waiting times. In most Dutch CI centers, the time between referral and surgery exceeds 6 months. Clinical experience suggests that during the waiting period for cochlear implantation, hearing and communication difficulties increase. Simultaneously, there is an interest in outcomes more closely aligned with patient values and needs, which resulted in the SMILE (Societal Merit of Interventions on hearing Loss Evaluation) study. This paper presents results on observed changes in societal and participatory outcomes during waiting time in participants with a time to CI surgery exceeding 6 months.

METHODS

SMILE is a prospective multi-center study including 232 individuals who were referred for unilateral CI. Continuous and nominal data from multiple questionnaires, sent immediately after referral and shortly before surgery, were analyzed by computing differences, Cohen's D, and odds ratios.

RESULTS

Of the total 232 participants, 102 had a time between inclusion and surgery exceeding 6 months. Of these, 89 had (partially) filled out surveys at both time points. Of all the domain scores 55% did not show differences between timepoints. All Cohen's D estimates were relatively small, ranging from - 0.298 to 0.388 for those outcomes that showed noteworthy changes.

CONCLUSION

Waiting time from referral to surgery, even though exceeding 6 months, was observed to not seriously affect non-clinically-prioritized patients in an adverse way. Future investigations should identify subgroups on tolerable waiting times regarding short- and long-term outcomes.

TRIAL REGISTRATION

Trial registration number at ClinicalTrials.gov: NCT05525221, 25-8-2022.

Variability in clinicians' prediction accuracy for outcomes of adult cochlear implant users

OBJECTIVE

The variability in outcomes among adult cochlear implant (CI) users poses challenges for clinicians in accurately predicting the benefits of the implant for individual candidates. This study aimed to investigate the accuracy and confidence of clinicians in predicting speech perception outcomes for adult CI users one-year post-implantation.

DESIGN

Participants were presented with comprehensive information on pre-implantation, one-month post-implantation, and six-month post-implantation data for 10 case studies. The cases encompassed a range of one-year post-implantation phoneme scores, from low performers (27%) to high performers (92%). Participants were tasked with predicting the speech perception outcomes for these cases one year after implantation.

STUDY SAMPLE

Forty-one clinicians completed the full outcome prediction survey.

RESULTS

Our findings revealed a significant over-prediction of low performance by clinicians. Interestingly, clinicians tended to predict average performance (73-76% phoneme score) even when provided with information suggesting lower-than-average performance. Most clinicians expressed confidence in their predictions, irrespective of their accuracy.

CONCLUSIONS

Identifying signs of low performance, particularly in the early post-implantation period, can enable clinicians to implement early interventions. Further research into accurate outcome prediction is essential for managing expectations, providing counselling, increasing CI adoption, and optimising clinical care for both high and low performers.

Cochlear implantation in patients with inner ear schwannomas: a systematic review and meta-analysis of audiological outcomes

PURPOSE

In patients with inner ear schwannomas (IES), reports on hearing rehabilitation with cochlear implants (CI) have increased over the past decade, most of which are case reports or small case series. The aim of this study is to systematically review the reported hearing results with CI in patients with IES considering the different audiologic outcome measures used in different countries.

METHODS

According to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guideline, a search of published literature was conducted. We included patients with IES (primary or with secondary extension from the internal auditory canal (IAC) to the inner ear, sporadic or NF2 related) undergoing cochlear implantation with or without tumour removal. The audiological results were divided into the categories "monosyllables", "disyllables", "multisyllabic words or numbers", and "sentences".

RESULTS

Predefined audiological outcome measures were available from 110 patients and 111 ears in 27 reports. The mean recognition scores for monosyllabic words with CI were 55% (SD: 24), for bisyllabic words 61% (SD: 36), for multisyllabic words and numbers 87% (SD: 25), and 71% (SD: 30) for sentences. Results from for multisyllabic words and numbers in general showed a tendency towards a ceiling effect. Possible risk factors for performance below average were higher complexity tumours (inner ear plus IAC/CPA), NF2, CI without tumour removal ("CI through tumour"), and sequential cochlear implantation after tumour removal (staged surgery).

CONCLUSION

Hearing loss in patients with inner ear schwannomas can be successfully rehabilitated with CI with above average speech performance in most cases. Cochlear implantation thus represents a valuable option for hearing rehabilitation also in patients with IES while at the same time maintaining the possibility of MRI follow-up. Further studies should investigate possible risk factors for poor performance. Audiological tests and outcome parameters should be reported in detail and ideally be harmonized to allow better comparison between languages.

Explainable machine learning reveals the relationship between hearing thresholds and speech-in-noise recognition in listeners with normal audiograms

Some individuals complain of listening-in-noise difficulty despite having a normal audiogram. In this study, machine learning is applied to examine the extent to which hearing thresholds can predict speech-in-noise recognition among normal-hearing individuals. The specific goals were to (1) compare the performance of one standard (GAM, generalized additive model) and four machine learning models (ANN, artificial neural network; DNN, deep neural network; RF, random forest; XGBoost; eXtreme gradient boosting), and (2) examine the relative contribution of individual audiometric frequencies and demographic variables in predicting speech-in-noise recognition. Archival data included thresholds (0.25-16 kHz) and speech recognition thresholds (SRTs) from listeners with clinically normal audiograms (n = 764 participants or 1528 ears; age, 4-38 years old). Among the machine learning models, XGBoost performed significantly better than other methods (mean absolute error; MAE = 1.62 dB). ANN and RF yielded similar performances (MAE = 1.68 and 1.67 dB, respectively), whereas, surprisingly, DNN showed relatively poorer performance (MAE = 1.94 dB). The MAE for GAM was 1.61 dB. SHapley Additive exPlanations revealed that age, thresholds at 16 kHz, 12.5 kHz, etc., on the order of importance, contributed to SRT. These results suggest the importance of hearing in the extended high frequencies for predicting speech-in-noise recognition in listeners with normal audiograms.

Validation of a predictive model for speech discrimination after cochlear impIant provision

BACKGROUND

If sufficient speech discrimination is no longer achieved with conventional hearing systems, an audiological indication for a cochlear implant (CI) is given. However, there are no established target criteria for CI aftercare with regard to the level of speech comprehension to be achieved. The aim of this study is to validate an existing predictive model for speech comprehension after CI provision. This is applied to different patient groups.

MATERIALS AND METHODS

The prospective study included 124 postlingually deaf adults. The model is based on preoperative maximum monosyllabic recognition score, aided monosyllabic recognition score at 65 dBSPL, and age the time of implantation. The model was investigated with regard to prediction accuracy for monosyllabic recognition with CI after 6 months.

RESULTS

Mean speech discrimination improved from 10% with hearing aid to 65% with CI after 6 months, with a statistically significant improvement in 93% of cases. Deterioration of aided unilateral speech discrimination was not observed. The mean prediction error was 11.5 percentage points in the cases with preoperative scores better than zero and 23.2 percentage points in all other cases.

CONCLUSION

Cochlear implantation should also be considered in patients with moderately severe to severe hearing loss and insufficient speech discrimination with hearing aids. The model based on preoperatively measured data for predicting speech discrimination with CI can be used in preoperative consultation and in the context of postoperative quality assurance.

[Validation of a predictive model for speech discrimination after cochlear implant provision]

BACKGROUND

If sufficient speech discrimination is no longer achieved with conventional hearing systems, an audiological indication for a cochlear implant (CI) is given. However, there are no established target criteria for CI aftercare with regard to the level of speech comprehension to be achieved. The aim of this study is to validate an existing predictive model for speech comprehension after CI provision. This is applied to different patient groups.

MATERIALS AND METHODS

The prospective study included 124 postlingually deaf adults. The model is based on preoperative maximum monosyllabic recognition score, aided monosyllabic recognition score at 65 dB_SPL, and age the time of implantation. The model was investigated with regard to prediction accuracy for monosyllabic recognition with CI after 6 months.

RESULTS

Mean speech discrimination improved from 10% with hearing aid to 65% with CI after 6 months, with a statistically significant improvement in 93% of cases. Deterioration of aided unilateral speech discrimination was not observed. The mean prediction error was 11.5 percentage points in the cases with preoperative scores better than zero and 23.2 percentage points in all other cases.

CONCLUSION

Cochlear implantation should also be considered in patients with moderately severe to severe hearing loss and insufficient speech discrimination with hearing aids. The model based on preoperatively measured data for predicting speech discrimination with CI can be used in preoperative consultation and in the context of postoperative quality assurance.

Functional Brain Connections Identify Sensorineural Hearing Loss and Predict the Outcome of Cochlear Implantation

Identification of congenital sensorineural hearing loss (SNHL) and early intervention, especially by cochlear implantation (CI), are crucial for restoring hearing in patients. However, high accuracy diagnostics of SNHL and prognostic prediction of CI are lacking to date. To diagnose SNHL and predict the outcome of CI, we propose a method combining functional connections (FCs) measured by functional magnetic resonance imaging (fMRI) and machine learning. A total of 68 children with SNHL and 34 healthy controls (HC) of matched age and gender were recruited to construct classification models for SNHL and HC. A total of 52 children with SNHL that underwent CI were selected to establish a predictive model of the outcome measured by the category of auditory performance (CAP), and their resting-state fMRI images were acquired. After the dimensional reduction of FCs by kernel principal component analysis, three machine learning methods including the support vector machine, logistic regression, and k-nearest neighbor and their voting were used as the classifiers. A multiple logistic regression method was performed to predict the CAP of CI. The classification model of voting achieves an area under the curve of 0.84, which is higher than that of three single classifiers. The multiple logistic regression model predicts CAP after CI in SNHL with an average accuracy of 82.7%. These models may improve the identification of SNHL through fMRI images and prognosis prediction of CI in SNHL.