Screening patients with sensorineural hearing loss for vestibular schwannoma using a Bayesian classifier

Artificial Intelligence in Temporal Bone Imaging: A Systematic Review

OBJECTIVE

The human temporal bone comprises more than 30 identifiable anatomical components. With the demand for precise image interpretation in this complex region, the utilization of artificial intelligence (AI) applications is steadily increasing. This systematic review aims to highlight the current role of AI in temporal bone imaging.

DATA SOURCES

A Systematic Review of English Publications searching MEDLINE (PubMed), COCHRANE Library, and EMBASE.

REVIEW METHODS

The search algorithm employed consisted of key items such as 'artificial intelligence,' 'machine learning,' 'deep learning,' 'neural network,' 'temporal bone,' and 'vestibular schwannoma.' Additionally, manual retrieval was conducted to capture any studies potentially missed in our initial search. All abstracts and full texts were screened based on our inclusion and exclusion criteria.

RESULTS

A total of 72 studies were included. 95.8% were retrospective and 88.9% were based on internal databases. Approximately two-thirds involved an AI-to-human comparison. Computed tomography (CT) was the imaging modality in 54.2% of the studies, with vestibular schwannoma (VS) being the most frequent study item (37.5%). Fifty-eight out of 72 articles employed neural networks, with 72.2% using various types of convolutional neural network models. Quality assessment of the included publications yielded a mean score of 13.6 ± 2.5 on a 20-point scale based on the CONSORT-AI extension.

CONCLUSION

Current research data highlight AI's potential in enhancing diagnostic accuracy with faster results and decreased performance errors compared to those of clinicians, thus improving patient care. However, the shortcomings of the existing research, often marked by heterogeneity and variable quality, underscore the need for more standardized methodological approaches to ensure the consistency and reliability of future data.

LEVEL OF EVIDENCE

NA Laryngoscope, 2024.

Intelligent alert system for predicting invasive mechanical ventilation needs via noninvasive parameters: employing an integrated machine learning method with integration of multicenter databases

The use of invasive mechanical ventilation (IMV) is crucial in rescuing patients with respiratory dysfunction. Accurately predicting the demand for IMV is vital for clinical decision-making. However, current techniques are invasive and challenging to implement in pre-hospital and emergency rescue settings. To address this issue, a real-time prediction method utilizing only non-invasive parameters was developed to forecast IMV demand in this study. The model introduced the concept of real-time warning and leveraged the advantages of machine learning and integrated methods, achieving an AUC value of 0.935 (95% CI 0.933-0.937). The AUC value for the multi-center validation using the AmsterdamUMCdb database was 0.727, surpassing the performance of traditional risk adjustment algorithms (OSI(oxygenation saturation index): 0.608, P/F(oxygenation index): 0.558). Feature weight analysis demonstrated that BMI, Gcsverbal, and age significantly contributed to the model's decision-making. These findings highlight the substantial potential of a machine learning real-time dynamic warning model that solely relies on non-invasive parameters to predict IMV demand. Such a model can provide technical support for predicting the need for IMV in pre-hospital and disaster scenarios.

MRI Screening in Vestibular Schwannoma: A Deep Learning-based Analysis of Clinical and Audiometric Data

Objective

To find a more objective method of assessing which patients should be screened for a vestibular schwannoma (VS) with magnetic resonance imaging (MRI) using a deep-learning algorithm to assess clinical and audiometric data.

Materials and Methods

Clinical and audiometric data were collected for 592 patients who received an audiogram between January 2015 and 2020 at Duke University Health Center with and without VS confirmed by MRI. These data were analyzed using a deep learning-based analysis to determine if the need for MRI screening could be assessed more objectively with adequate sensitivity and specificity.

Results

Patients with VS showed slightly elevated, but not statistically significant, mean thresholds compared to those without. Tinnitus, gradual hearing loss, and aural fullness were more common in patients with VS. Of these, only the presence of tinnitus was statistically significant. Several machine learning algorithms were used to incorporate and model the collected clinical and audiometric data, but none were able to distinguish ears with and without confirmed VS. When tumor size was taken into account the analysis was still unable to distinguish a difference.

Conclusions

Using audiometric and clinical data, deep learning-based analyses failed to produce an adequately sensitive and specific model for the detection of patients with VS. This suggests that a specific pattern of audiometric asymmetry and clinical symptoms may not necessarily be predictive of the presence/absence of VS to a level that clinicians would be comfortable forgoing an MRI.

Machine Learning in the Management of Lateral Skull Base Tumors: A Systematic Review

The application of machine learning (ML) techniques to otolaryngology remains a topic of interest and prevalence in the literature, though no previous articles have summarized the current state of ML application to management and the diagnosis of lateral skull base (LSB) tumors. Subsequently, we present a systematic overview of previous applications of ML techniques to the management of LSB tumors. Independent searches were conducted on PubMed and Web of Science between August 2020 and February 2021 to identify the literature pertaining to the use of ML techniques in LSB tumor surgery written in the English language. All articles were assessed in regard to their application task, ML methodology, and their outcomes. A total of 32 articles were examined. The number of articles involving applications of ML techniques to LSB tumor surgeries has significantly increased since the first article relevant to this field was published in 1994. The most commonly employed ML category was tree-based algorithms. Most articles were included in the category of surgical management (13; 40.6%), followed by those in disease classification (8; 25%). Overall, the application of ML techniques to the management of LSB tumor has evolved rapidly over the past two decades, and the anticipated growth in the future could significantly augment the surgical outcomes and management of LSB tumors.

Machine Learning for Vestibular Schwannoma Diagnosis Using Audiometrie Data Alone

OBJECTIVE

The aim of this study is to compare machine learning algorithms and established rule-based evaluations in screening audiograms for the purpose of diagnosing vestibular schwannomas. A secondary aim is to assess the performance of rule-based evaluations for predicting vestibular schwannomas using the largest dataset in the literature.

STUDY DESIGN

Retrospective case-control study.

SETTING

Tertiary referral center.

PATIENTS

Seven hundred sixty seven adult patients with confirmed vestibular schwannoma and a pretreatment audiogram on file and 2000 randomly selected adult controls with audiograms.

INTERVENTIONS

Audiometric data were analyzed using machine learning algorithms and standard rule-based criteria for defining asymmetric hearing loss.

MAIN OUTCOME MEASURES

The primary outcome is the ability to identify patients with vestibular schwannomas based on audiometric data alone, using machine learning algorithms and rule-based formulas. The secondary outcome is the application of conventional rule-based formulas to a larger dataset using advanced computational techniques.

RESULTS

The machine learning algorithms had mildly improved specificity in some fields compared with rule-based evaluations and had similar sensitivity to previous rule-based evaluations in diagnosis of vestibular schwannomas.

CONCLUSIONS

Machine learning algorithms perform similarly to rule-based evaluations in identifying patients with vestibular schwannomas based on audiometric data alone. Performance of established rule-based formulas was consistent with earlier performance metrics, when analyzed using a large dataset.

Influence of noise level and seniority in the workplace on the SAL, ELI and percentage of hearing loss indices in the diagnosis and prevention of hearing loss in the working population

INTRODUCTION

This research relates the most important work-related factors affecting the development of hearing loss to the main methods used as medical assessment criteria in the diagnosis of occupational deafness. These criteria are the Speech Average Loss Index (SAL), the Early Loss Index (ELI) and the Percentage of Hearing Loss, and are applied to data obtained from audiograms performed on workers in occupational medical examinations.

METHOD

Depending on the assessment method selected, these often return different results in grading an individual's hearing status and predicting how it will evolve. To address this problem, medical examinations (including audiograms) were carried out on a heterogeneous sample of 1,418 workers in Spain, from which demographic or personal data (gender, age, etc.), occupational data (noise level to which each individual is exposed, etc.) and other non-work-related factors (exposure to noise outside work, family history, etc.) were also gathered. Using Bayesian Networks, the conditional probability of an individual developing hearing loss was obtained taking into account all these factors and, specifically, noise level and length of service in the workplace. Sensitivity analyses were also carried out using the three scales (SAL, ELI and Percentage Hearing Loss Index), proving their suitability as tools the diagnosis and prediction of deafness. These networks were validated under the Receiver Operating Characteristic curve (ROC) criterion and in particular by the Area Under the Curve (AUC).

RESULTS

The results show that all three methods are deficient in so far as detecting preventive hearing problems related to noise in most workplaces.

CONCLUSIONS

The most restrictive methods for detecting possible cases of deafness are the SAL index and the Percentage Loss Index. The ELI index is the least restrictive of the three methods, but it is not able to discriminate the causes of hearing problems in an individual caused by exposure to noise, either by its intensity level or by the time of exposure to noise. Practical Applications: The use of the three methods in the field of occupational risk prevention is extremely limited and it seems reasonable to think that there is a need for the construction of new scales to correct or improve the existing ones.

Multiple Audiometric Analysis in the Screening of Vestibular Schwannoma

Introduction

Asymmetric sensorineural hearing loss is the main early symptom of retrocochlear lesions, hence its importance in screening for vestibular schwannomas. Currently, there is no consensus regarding its definition. The objective was to identify the audiometric pattern that would serve as a predictor for vestibular schwannoma in patients with asymmetric hearing loss.

Materials and methods

A cross-sectional study was conducted that included patients with asymmetric hearing loss attending a secondary care center and a tertiary care center. Clinical, audiometric and imaging (MRI with gadolinium) variables were collected. Asymmetric hearing loss was defined as a difference of 15 dB in one or more frequencies between both ears. The sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV) and accuracy of different audiometric patterns were analyzed.

Results

A total of 107 patients were studied and divided into two groups: group 1 without vestibular schwannoma (n=98); and group 2 with vestibular schwannoma (n=9). No significant difference in demographic characteristics or audiometric patterns was found in patients with and without vestibular schwannoma. The audiometric pattern with the best sensitivity as a screening test was a difference >20 dB in the 4,000 Hz frequency, with a sensitivity of 77.78%, specificity of 30.61%, PPV of 8.33%, NPV of 93.75% and accuracy of 34.50%.

Conclusion

The audiometric pattern with the best results was a difference >20 dB in the 4,000 Hz frequency range; however, patients with asymmetric hearing loss could not be differentiated from patients with retrocochlear lesions based only on audiometry. Asymmetrical hearing loss must be studied with MRI.

Imaging of the Vestibular Schwannoma: Diagnosis, Monitoring, and Treatment Planning

Appropriate imaging strategies for the detection, treatment planning, and posttreatment monitoring of vestibular schwannomas will be discussed. The typical and variant imaging appearances of vestibular schwannomas, as well as the imaging features that should prompt consideration of differential diagnoses, will be illustrated. Understanding the natural history of vestibular schwannomas, optimal measurement and definition of tumour growth helps the radiologist evaluate for the failure of conservative management and requirement for surgery or radiotherapy. In order to determine the success of conservative management, the radiologist is required to understand the natural history of vestibular schwannomas and how tumour growth is defined. Finally, the imaging features which help guide appropriate treatment with surgery or radiotherapy will be highlighted, and the expected posttreatment imaging changes will be described.

The use of magnetic resonance imaging in the investigation of patients with unilateral non-pulsatile tinnitus without asymmetrical hearing loss

BACKGROUND

Tinnitus is a common condition presenting to the ENT out-patient clinic. Vestibular schwannomas are benign cerebellopontine angle tumours that usually present with unilateral sensorineural hearing loss. Magnetic resonance imaging of the internal auditory meatus is the definitive investigation in their detection. The current recommendation is for unilateral tinnitus patients to undergo magnetic resonance imaging of the internal auditory meatus to exclude vestibular schwannoma.

OBJECTIVE

To evaluate magnetic resonance imaging in the investigation of patients with unilateral non-pulsatile tinnitus without asymmetrical hearing loss.

METHOD

A retrospective case series was conducted of all patients who underwent magnetic resonance imaging of the internal auditory meatus to investigate unilateral non-pulsatile tinnitus without asymmetrical hearing loss, from 1 January 2014 to 1 January 2019.

RESULTS

Of 2066 scans, 566 (27 per cent) were performed to investigate patients (335 female, 231 male) with unilateral non-pulsatile tinnitus without asymmetrical hearing loss. Three vestibular schwannomas were detected on imaging, and 134 incidental findings were discovered.

CONCLUSION

The detection rate of vestibular schwannoma in this group was just 0.3 per cent. This paper questions the utility of magnetic resonance imaging evaluation in these patients.

Evaluation of pure-tone audiometric protocols in vestibular schwannoma screening

The objective was to evaluate the pure-tone audiogram-based screening protocols in VS diagnostics. We retrospectively analyzed presenting symptoms, pure tone audiometry and MRI finding from 246 VS patients and 442 controls were collected to test screening protocols (AAO-HNS, AMCLASS-A/B, Charing Cross, Cueva, DOH, Nashville, Oxford, Rule3000, Schlauch, Seattle, Sunderland) for sensitivity and specificity. Results were pooled with data from five other studies, and analysis of sensitivity, specificity and positive likelihood ratio (LR+) for each protocol was performed. Our results show that protocols with significantly higher sensitivity (AMCLASS-A/B, Nashville) show also significantly lowest specificity, and tend to have low association (positive likelihood ratio, LR+) to the VS. The highest LR+ was found for protocols AAO-HNS, Rule3000 and Seattle. In conclusions, knowing their properties, screening protocols are simple decision-making tools in VS diagnostic. To use the advantage of the highest sensitivity, protocols AMCLASS-A + B or Nashville can be of choice. For more reasonable approach, applying the protocols with high LR+ (AAO-HNS, Rule3000, Seattle) may reduce the overall number of MRI scans at expense of only few primarily undiagnosed VS.

Asymmetric sensorineural hearing loss and vestibular schwannoma: when to image?

PURPOSE OF REVIEW

We review the literature on the use of audiometric protocols in the guidance of when to obtain MRI for detection of vestibular schwannoma. This discussion will focus on the sensitivity, specificity, and cost-effective analysis of audiometric criteria of asymmetric sensorineural hearing loss (ASNHL) when used to decide when MRI scans should be performed.

RECENT FINDINGS

The sensitivity for detecting vestibular schwannomas when invoking published audiometric protocols for triggering MRI acquisition in ASNHL ranged from 50 to 100%. Specificity of these protocols ranged from 23 to 83%. Such audiometric protocols are efficient, achieving sensitivity and specificity at these rates while reducing the screening rate to 18 to 35%. The reduced procurement of MRI while using such audiometric protocols is associated with annual cost savings of between 23 and 82%. While no definitive recommendations can be made from this review, some audiometric protocols offer a better balance of sensitivity and specificity than others.

SUMMARY

Audiometric protocols for triggering MRI acquisition in ASNHL for evaluation of vestibular schwannoma can be both sensitive and specific. These are competitive measures, and so no protocol is both 100% sensitive and specific. Such protocols become less effective in populations with increased incidence of noise-induced hearing loss. Invocation of such audiometric protocols can considerably reduce the annual cost of MRI evaluation for vestibular schwannomas.

Audiologist-led screening of acoustic neuromas in patients with asymmetrical sensorineural hearing loss and/or unilateral tinnitus: our experience in 1126 patients

OBJECTIVE

To determine whether patients within an otolaryngology department presenting with asymmetrical sensorineural hearing loss and/or unilateral tinnitus can be safely and cost-efficiently screened for acoustic neuroma by audiologists as a first or only point of contact.

METHODS

A prospective case series and cost analysis were conducted at a tertiary referral centre. Between April 2013 and March 2017, 1126 adult patients presented to the audiology department with asymmetrical sensorineural hearing loss and/or unilateral tinnitus. All were screened for acoustic neuroma with magnetic resonance imaging, based on pre-determined criteria. The main outcome measure was the presence of acoustic neuroma or other pathology on magnetic resonance imaging.

RESULTS

Twenty-five patients (2.22 per cent) were found to have an acoustic neuroma (size range: 3-20 mm) and were referred to the otolaryngologist for further assessment. The remaining patients were appropriately managed and discharged by the audiologists without ENT input. This resulted in an overall cost saving of £164 850.

CONCLUSION

Patients with asymmetrical sensorineural hearing loss and/or unilateral tinnitus can be safely screened for acoustic neuroma and independently managed by audiologists as a first or only point of contact, resulting in considerable departmental cost savings.

Three-Year Ear, Nose, and Throat Cross-sectional Analysis of Audiometric Protocols for Magnetic Resonance Imaging Screening of Acoustic Tumors

Objectives. (1) Evaluate audiometric protocols and recommend protocols with best sensitivity and specificity for magnetic resonance imaging (MRI) screening of acoustic tumors; (2) determine clinical risks (false negative) of missing acoustic tumors and potential wastes in screening (false positive) nonacoustic tumors or radiologically “normal” cases; and (3) identify the decibel difference and range of frequencies compared by the best-performing protocols.

Study Design. Cross-sectional study with chart review.

Setting. Ear, nose, and throat (ENT); audiology; and radiology departments in a tertiary-care hospital.

Subjects and Methods. Three-year cohort (2006-2009) of 1751 ENT patients underwent MRI screening and pure-tone audiometry indicating sensorineural hearing loss. Audiometric protocols were ranked by highest sensitivity to acoustic tumors, specificity A to nonacoustic tumors, and specificity B to “radiologically normal” cases.

Results. No audiometric protocols achieved 100% sensitivity or specificity rates. Only 2 protocols achieved ≥90% sensitivity: the AMCLASS-A-Urben protocol (93.16%) and the Mangham protocol (91.58%). Eleven of 15 protocols for specificity A and 12 of 15 protocols for specificity B achieved ≥50%. Clinical risks ranged from 6.84% to 18.95%, whereas potential wastes ranged from 33.56% to 68.37% for specificity A and 31.76% to 66.86% for specificity B. Interaural difference parameters indicating highest mean sensitivity were on the order of ≥10 dB, ≥15 dB, and ≥20 dB. For frequency comparison parameters, “2 or more adjacent frequency” and “single-frequency” comparison indicated higher mean sensitivity than the “averaged multifrequency” comparison. Mean specificity showed an opposite pattern.

Conclusions. For optimum sensitivity, the Mangham protocol is preferred (sensitivity, 91.58%; specificity A, 44.23%; specificity B, 44.91%), which proposes a ≥10-dB interaural difference, averaging 1 to 8 kHz. For optimum specificity, the American Academy of Otolaryngology–Head and Neck Surgery protocol is preferred (sensitivity, 87.37%; specificity A, 65.38%; specificity B, 66.04%), which proposes ≥15 dB between ears, averaging 0.5 to 3 kHz.

Should all patients referred for magnetic resonance imaging scans of their internal auditory meatus be followed up in ENT clinics?

Magnetic resonance imaging (MRI) is recognised as the "Gold Standard" investigation for symptoms pertaining to the inner ear and detection of retro-cochlear pathology. There is still no accurate clinical predictor for cerebellopontine angle lesions and increasingly more normal scans are being performed. With constantly increasing demands on ENT outpatient clinics, our aim was to investigate whether all patients referred for MRI of their internal auditory meatus (IAM) require follow-up in ENT clinics. A retrospective study was carried out in a tertiary referral centre referring patients for MRI IAM from ENT clinics and neurotology clinics on 153 patients referred for MRI IAM performed over a 4-month period. The MRI reports and the case notes of the patients were reviewed. MRI results and patient symptoms with patient follow-up schedule and follow-up situation for at least 6 months post-scan were compared. There were two patient groups, those referred from main outpatients (81) and those referred by the audiological physicians (72). Of the total number of scans, 101 were reported as normal, 45 had incidental findings, and 7 showed pathology of the cerebellopontine angle. The presenting complaints of the patients, the scan results and the follow-up since were compared. Six months later 63% of those referred from outpatients were no longer being followed up in clinic. We suggest that 56.8% of patients referred for MRI IAM do not require ENT follow-up. Their symptoms and concerns could be dealt with at the first consultation where onward referral could be made if necessary. Once checked by the requesting clinician, uncomplicated scan results could be sent to the patients and general practitioners by post or email. This would reduce the burden on general otolaryngology outpatient clinics and improve resource utilisation.

Vestibular schwannoma: when to look for it?

OBJECTIVES

(1) To compare audiometric parameters in patients with vestibular schwannoma and in those with asymmetric hearing loss from other causes; and (2) to assess proposed screening criteria by comparing published protocols.

METHODS

Audiometric data from 199 vestibular schwannoma patients and 225 non-tumour patients were compared. Eight screening protocols were tested on these 424 patients.

RESULTS

Vestibular schwannoma and non-tumour patients with little or no hearing loss in the unaffected ear were inseparable; however, vestibular schwannoma patients with hearing loss in the unaffected ear had greater audiometric asymmetry, compared with non-tumour patients with the same pattern of hearing loss. The sensitivity of screening protocols varied from 73 to 100 per cent; parallelism was observed between sensitivity and screening rate.

CONCLUSION

As regards vestibular schwannoma screening protocols, the best compromise between sensitivity and screening rate was offered by a criterion comprising either: (1) > or =20 dB asymmetry at two neighbouring frequencies, or unilateral tinnitus, or (2) > or =15 dB asymmetry at two frequencies between 2 and 8 kHz.