Pure tone audiogram classification using deep learning techniques.

OBJECTIVE: Pure tone audiometry has played a critical role in audiology as the initial diagnostic tool, offering vital insights for subsequent analyses. This study aims to develop a robust deep learning framework capable of accurately classifying audiograms across various commonly encountered tasks. DESIGN, SETTING, AND PARTICIPANTS: This single-centre retrospective study was conducted in accordance with the STROBE guidelines. A total of 12 518 audiograms were collected from 6259 patients aged between 4 and 96 years, who underwent pure tone audiometry testing between February 2018 and April 2022 at Tongji Hospital, Tongji Medical College, Wuhan, China. Three experienced audiologists independently annotated the audiograms, labelling the hearing loss in degrees, types and configurations of each audiogram. MAIN OUTCOME MEASURES: A deep learning framework was developed and utilised to classify audiograms across three tasks: determining the degrees of hearing loss, identifying the types of hearing loss, and categorising the configurations of audiograms. The classification performance was evaluated using four commonly used metrics: accuracy, precision, recall and F1-score. RESULTS: The deep learning method consistently outperformed alternative methods, including K-Nearest Neighbors, ExtraTrees, Random Forest, XGBoost, LightGBM, CatBoost and FastAI Net, across all three tasks. It achieved the highest accuracy rates, ranging from 96.75% to 99.85%. Precision values fell within the range of 88.93% to 98.41%, while recall values spanned from 89.25% to 98.38%. The F1-score also exhibited strong performance, ranging from 88.99% to 98.39%. CONCLUSIONS: This study demonstrated that a deep learning approach could accurately classify audiograms into their respective categories and could contribute to assisting doctors, particularly those lacking audiology expertise or experience, in better interpreting pure tone audiograms, enhancing diagnostic accuracy in primary care settings, and reducing the misdiagnosis rate of hearing conditions. In scenarios involving large-scale audiological data, the automated classification system could be used as a research tool to efficiently provide a comprehensive overview and statistical analysis. In the era of mobile audiometry, our deep learning framework can also help patients quickly and reliably understand their self-tested audiograms, potentially encouraging timely consultations with audiologists for further evaluation and intervention.

Field test of the Rapid Assessment of Hearing Loss survey protocol in Ntcheu district, Malawi.

Objective: (1) To test the feasibility of the Rapid Assessment of Hearing Loss (RAHL) survey protocol in Malawi (Ntcheu); (2) To estimate the prevalence and probable causes of hearing loss (adults 50+).Design: Cross-sectional population-based survey.Study sample: Clusters (n = 38) were selected using probability-proportionate-to-size-sampling. Within each cluster, 30 people aged 50+ were selected using compact-segment-sampling. All participants completed smartphone-based audiometry (hearTest). Prevalence was estimated using WHO definitions (PTA of thresholds 0.5, 1, 2, 4 kHz in the better ear of >25 dB HL (any) and >40 dB HL (≥moderate)). Otoscopy and questionnaire were used to assess probable causes. Participants with hearing loss and/or ear disease were asked about care-seeking and barriers.Results: Four teams completed the survey in 24 days. 1080 of 1153 (93.7%) participants were examined. The median time to complete the protocol was 24 min/participant. Prevalence of hearing loss was 35.9% (95% CI = 31.6-40.2) (any level); and 10.0% (95% CI = 7.9-12.5) (≥moderate). The majority was classified as probable sensorineural. Nearly one third of people (30.9%) needed diagnostic audiology services and possible hearing aid fitting. Hearing aid coverage was <1%. Lack of perceived need was a key barrier.Conclusion: The RAHL is simple, fast and provides information about the magnitude and probable causes of hearing loss to plan services.

Prevalence and nature of hearing loss in a cohort of children with sickle cell disease.

BACKGROUND: Sickle cell disease (SCD) may cause injury to any organ, including the auditory system. Although the association of SCD and hearing loss has been described, the nature of this complication is unknown. We sought to establish the prevalence and nature of hearing loss in a referred cohort of children with SCD and to identify correlating disease- or treatment-associated factors. PROCEDURE: We conducted a retrospective review of patients with SCD < 22 years of age who had hearing evaluations between August 1990 and December 2014. Demographics, audiograms, and disease and treatment variables were analyzed. RESULTS: Two hundred and ten audiograms among 81 patients were reviewed, and 189 were evaluable. Seventy-two children constituted the referred cohort. Fourteen (19.4%) had hearing loss documented on at least one audiogram. Seven (9.7%) patients had only conductive hearing loss, and the loss persisted for up to 10.3 years. The median age of first identification was eight years. Six (8.3%) patients had hearing loss that was at least partially sensorineural. One patient's hearing loss was ambiguous. All sensorineural hearing losses were unilateral and 4/6 patients had prior documented normal hearing, indicating acquired loss. No correlations were identified. CONCLUSIONS: Both conductive and sensorineural hearing losses are more prevalent in our study population than those observed in the general pediatric population. In children with SCD, sensorineural hearing loss appears to be acquired and unilateral. Conductive hearing loss was identified in older children and can persist. Serial screening is needed for early detection and more prompt intervention in this population.

Cochlear implant outcomes in the elderly: a uni- and multivariate analyses of prognostic factors.

PURPOSE: To assess preoperative features that could predict the audiological outcome after cochlear implantation in the elderly, in terms of pure tone audiometry, speech audiometry, and speech perception performance. METHODS: All available records of patients with cochlear implants aged 65 or more at the time of their implantation at our Institution were reviewed (50 patients, mean age 70.76 ± 4.03 years), recording preoperative clinical features. Pure tone audiometry, speech audiometry, and speech perception performance 1 year after cochlear implant activation and fitting were used as outcome measures. RESULTS: No statistically significant association emerged between clinical features and pure tone audiometry. On univariate analysis, progressive sensorineural hearing loss of unknown origin was associated with a better outcome in terms of speech audiometry and speech perception performance (p = 0.035 and p = 0.033, respectively). On multivariate analysis, progressive sensorineural hearing loss retained its independent prognostic significance in terms of speech perception performance (p = 0.042). The discriminatory power of a two-variable panel (age and etiology of hearing loss) featured an AUC (ROC) of 0.738 (an acceptable discriminatory power according to the Hosmer-Lemeshow scale). CONCLUSIONS: A progressive sensorineural hearing loss of unknown origin was associated with a better outcome in terms of speech perception in the elderly in our case study. Further features that can predict audiological outcome achievable with cochlear implants in the elderly are desirable to perform adequate counselling and rehabilitation programs.

Ototoxicity (cochleotoxicity) classifications: A review.

OBJECTIVE: Drug-mediated ototoxicity, specifically cochleotoxicity, is a concern for patients receiving medications for the treatment of serious illness. A number of classification schemes exist, most of which are based on pure-tone audiometry, in order to assist non-audiological/non-otological specialists in the identification and monitoring of iatrogenic hearing loss. This review identifies the primary classification systems used in cochleototoxicity monitoring. By bringing together classifications published in discipline-specific literature, the paper aims to increase awareness of their relative strengths and limitations in the assessment and monitoring of ototoxic hearing loss and to indicate how future classification systems may improve upon the status-quo. DESIGN: Literature review. STUDY SAMPLE: PubMed identified 4878 articles containing the search term ototox*. RESULTS: A systematic search identified 13 key classification systems. Cochleotoxicity classification systems can be divided into those which focus on hearing change from a baseline audiogram and those that focus on the functional impact of the hearing loss. CONCLUSIONS: Common weaknesses of these grading scales included a lack of sensitivity to small adverse changes in hearing thresholds, a lack of high-frequency audiometry (>8 kHz), and lack of indication of which changes are likely to be clinically significant for communication and quality of life.

Genotype phenotype correlations for hearing impairment: approaches to management.

Hearing impairment is an extremely heterogeneous disorder, with both environmental as well as genetic causes. This review describes the known genes involved in non-syndromic hearing impairment and their genotype-phenotype correlations where possible. Furthermore, some of the more frequent syndromic forms of hearing impairment are described, in particular where they overlap with the non-syndromic forms. Given the heterogeneity of the disorder, together with the indistinguishable phenotypes for many of the genes, it is suggested that testing for mutations is performed using massive parallel sequencing techniques, either by a large targeted set of genes or by an exome wide analysis.

The ICF Core Sets for hearing loss--researcher perspective. Part I: Systematic review of outcome measures identified in audiological research.

OBJECTIVE: To review the literature in order to identify outcome measures used in research on adults with hearing loss (HL) as part of the ICF Core Sets development project, and to describe study and population characteristics of the reviewed studies. DESIGN: A systematic review methodology was applied using multiple databases. A comprehensive search was conducted and two search pools were created, pool I and pool II. STUDY SAMPLE: The study population included adults (≥ 18 years of age) with HL and oral language as the primary mode of communication. RESULTS: 122 studies were included. Outcome measures were distinguished by 'instrument type', and 10 types were identified. In total, 246 (pool I) and 122 (pool II) different measures were identified, and only approximately 20% were extracted twice or more. Most measures were related to speech recognition. Fifty-one different questionnaires were identified. Many studies used small sample sizes, and the sex of participants was not revealed in several studies. CONCLUSION: The low prevalence of identified measures reflects a lack of consensus regarding the optimal outcome measures to use in audiology. Reflections and discussions are made in relation to small sample sizes and the lack of sex differentiation/descriptions within the included articles.

The ICF Core Sets for hearing loss: researcher perspective, Part II: Linking outcome measures to the International Classification of Functioning, Disability and Health (ICF).

OBJECTIVE: To link outcome measures used in audiological research to the ICF classification and thereby describe audiological research from the ICF perspective. DESIGN: Through a peer-reviewed or a joint linking procedure, link outcome measures to the ICF classification system using standardized ICF linking rules. Additional linking rules were developed in combination with the established rules to overcome difficulties when connecting audiological data to ICF. Absolute and relative frequencies of ICF categories were reported. STUDY SAMPLE: The identified outcome measures from the previous study (Part I) constituted the empirical material. RESULTS: In total, 285 ICF categories were identified. The most prevalent categories were related to listening, hearing functions, auditory perceptions, emotions and the physical environment, such as noise and hearing aids. Categories related to communication showed lower relative frequencies, as did categories related to the social and attitudinal environment. CONCLUSIONS: Based on the linked outcome measures, communication as a research topic is subordinated to other research topics. The same conclusion can be drawn for research targeting the social and attitudinal environment of adults with HL. Difficulties in the linking procedure were highlighted and discussed, and suggestions for future revisions of the ICF from the audiological perspective were described.

Easy as 1-2-3: Development and Evaluation of a Simple yet Valid Audiogram-Classification System.

Almost since the inception of the modern-day electroacoustic audiometer a century ago the results of pure-tone audiometry have been characterized by an audiogram. For almost as many years, clinicians and researchers have sought ways to distill the volume and complexity of information on the audiogram. Commonly used approaches have made use of pure-tone averages (PTAs) for various frequency ranges with the PTA for 500, 1000, 2000 and 4000 Hz (PTA4) being the most widely used for the categorization of hearing loss severity. Here, a three-digit triad is proposed as a single-number summary of not only the severity, but also the configuration and bilateral symmetry of the hearing loss. Each digit in the triad ranges from 0 to 9, increasing as the level of the pure-tone hearing threshold level (HTL) increases from a range of optimal hearing (< 10 dB Hearing Level; HL) to complete hearing loss (≥ 90 dB HL). Each digit also represents a different frequency region of the audiogram proceeding from left to right as: (Low, L) PTA for 500, 1000, and 2000 Hz; (Center, C) PTA for 3000, 4000 and 6000 Hz; and (High, H) HTL at 8000 Hz. This LCH Triad audiogram-classification system is evaluated using a large United States (U.S.) national dataset (N = 8,795) from adults 20 to 80 + years of age and two large clinical datasets totaling 8,254 adults covering a similar age range. Its ability to capture variations in hearing function was found to be superior to that of the widely used PTA4.

Automated hearing loss type classification based on pure tone audiometry data.

Hearing problems are commonly diagnosed with the use of tonal audiometry, which measures a patient's hearing threshold in both air and bone conduction at various frequencies. Results of audiometry tests, usually represented graphically in the form of an audiogram, need to be interpreted by a professional audiologist in order to determine the exact type of hearing loss and administer proper treatment. However, the small number of professionals in the field can severely delay proper diagnosis. The presented work proposes a neural network solution for classification of tonal audiometry data. The solution, based on the Bidirectional Long Short-Term Memory architecture, has been devised and evaluated for classifying audiometry results into four classes, representing normal hearing, conductive hearing loss, mixed hearing loss, and sensorineural hearing loss. The network was trained using 15,046 test results analysed and categorised by professional audiologists. The proposed model achieves 99.33% classification accuracy on datasets outside of training. In clinical application, the model allows general practitioners to independently classify tonal audiometry results for patient referral. In addition, the proposed solution provides audiologists and otolaryngologists with access to an AI decision support system that has the potential to reduce their burden, improve diagnostic accuracy, and minimise human error.

Dentistry's intrinsic link to provision of services for persons with disabilities.

This article provides the dentist with a background on disabilities, education, practice and provision of services related to statutory and regulatory requirements, particularly the Americans with Disabilities Act anrid the California Fair Employment and Housing Act.

Genetics of hearing loss: where are we standing now?

Hearing loss (HL) is the most common sensory impairment and is caused by a broad range of inherited to environmental causes. Inherited HL consists 50-60% of all HL cases. The inherited form of HL is further classified to different categories. More than 300 syndromes and 40 genes have been identified to result in different levels of HL. Although several diagnostic or screening tests have been developed, yet there are controversies around their use.

[Hearing impairment and language development]. / Periphere Hörstörungen und Spracherwerb.

A child has the first 4 years of life in which to develop speech. The first 8-10 months see the ability to distinguish sounds; later comes further linguistic ability. A hearing impairment more than 25 dB in the better ear in the speech frequencies (500-4 000 Hz) for more than 3 months has clear consequences, especially in the first 4 years of life. Therefore early diagnosis and treatment with rehabilitation of hearing loss is important for proper speech development.

Definitions of types of hearing impairment: a discussion paper.

OBJECTIVE: To develop a clear and meaningful set of definitions of types of hearing impairment. METHOD: A critical analysis was made of previous definitions together with a consideration of current knowledge of functions and dysfunctions of the auditory system. RESULTS AND CONCLUSIONS: A coherent set of definitions compatible with the contemporary understanding of auditory disorders and their perceptual effects is proposed.

The burden and management of infants with life-long and irreversible hearing impairment in Nigeria.

If left untreated or treated late and inappropriately, sensorineural hearing impairment has several adverse, irreversible and life-long health, psycho-social and economic consequences for the affected infants and their families. This seminar paper sets out to describe the burden, current practices and management options for sensorineural or permanent hearing impairment in early infancy for primary care physicians in Nigeria against the backdrop of their role as primary or first contact for all health consultations. Available studies suggest that up to 2.7% or 162,000 of the 6 million infants born annually may have permanent hearing impairment. The underlying aetiological factors may not be determined in the majority of the infants thus limiting the effectiveness of any primary prevention initiatives. However, the affected infants can be detected accurately with objective screening technologies such as otoacoustic emissions and auditory brainstem response in hospital or community settings. The ethical and scientific rationale for this intervention has been well established. Barring the challenge of parental follow-up default which is not insurmountable, affected infants and their families can be supported to establish appropriate auditory-based communication, avoid potentially harmful traditional therapies and child neglect commonly associated with childhood hearing impairment. Primary care providers in private and public practice have a crucial role in guiding parents to seek timely and appropriate services from ear care providers and child development specialists to ensure optimal child growth and developmental outcomes. On-going parental commitment to and active participation in the selected intervention programmes are essential to satisfactory long-term outcomes.

Frequency and causes of hearing impairment in tertiary care center.

OBJECTIVE: To determine the frequency and causes of hearing impairment in patients attending the out patient department of Civil Hospital Karachi. METHODS: This study was carried out at the ENT department of Civil Hospital Karachi from January 1998 to October 2000. Six hundred patients were randomly selected from the out patient department. Patients of all age groups and both genders were included, who complained of hearing impairment in one or both ears. Detailed history, complete E.N.T. examination and relevant audiological, radiological investigations were carried out. RESULTS: Frequency of conductive hearing loss was 50%, sensorineural hearing loss 20% and mixed hearing loss in 30%. CSOM remains the most common cause in this study. In majority of cases of sensorineural hearing loss, prenatal and perinatal factors predominate. Consanguinity was the most frequent factor in this study in sensorineural type of hearing loss. CONCLUSION: Conductive hearing loss is more prevalent than sensorineural hearing loss by a ratio of 2.5:1. Otitis media remains the commonest cause in this study i.e. 76%. Prevention of prenatal and postnatal factors can decrease the frequency of sensorineural hearing loss.