"Seeing inside out": revealing the effectiveness of otoscopy training in virtual reality enhanced practical exams - a randomized controlled trial.

BACKGROUND: The study aimed to assess the impact of different training modalities on otoscopy performance during a practical exam using a high-fidelity simulator and to determine if objective evaluation of otoscopy is feasible using a simulator that records insertion depth and tympanic membrane coverage. METHODS: Participants were assigned to one of four groups: control and three intervention groups with varying training approaches. Participants received otoscopy training and then were assessed through a practical exam on a high-fidelity simulator that uses virtual reality to visualize the ear canal and middle ear. Performance was evaluated using a modified Objective Structured Assessment of Technical Skills checklist and Integrated Procedural Performance Instrument checklist. Insertion depth, tympanic membrane coverage, and correct diagnosis were recorded. Data were tested for normal distribution using the Shapiro-Wilk test. One-way ANOVA and, for non-normally distributed data, Kruskal-Wallis test combined with Dunn's test for multiple comparisons were used. Interrater reliability was assessed using Cohen's κ and Intraclass correlation coefficient. RESULTS: All groups rated their training sessions positively. Performance on the OSATS checklist was similar among groups. IPPI scores indicated comparable patient handling skills. The feedback group examined larger tympanic membrane areas and had higher rates of correct diagnosis. The correct insertion depth was rarely achieved by all participants. Interrater reliability for OSATS was strong. IPPI reliability showed good correlation. CONCLUSION: Regardless of training modality, participants perceived learning improvement and skill acquisition. Feedback improved examination performance, indicating simulator-guided training enhances skills. High-fidelity simulator usage in exams provides an objective assessment of performance.

Bone cement repair of malleus handle fractures: Intraoperative video and case report of two patients.

Isolated malleus fractures are a rare occurrence with few reported cases in the literature. Symptoms include sudden otalgia, hearing loss, tinnitus and aural fullness. Work-up and diagnosis are based on a combination of thorough anamnesis and careful otoscopic evaluation or high-resolution computer tomography. We present two cases of isolated malleus handle fractures who were diagnosed based on a combination of pneumatic otoscopy and tympanometry. Both fractures were surgically repaired using hydroxyapatite bone cement as showcased in the supplemental video material. Post-operative audiometry showed improvement in the pure-tone-average of both patients as well as normalisation of tympanometry. Isolated malleus fracture should be suspected in cases of sudden hearing loss and tinnitus following digital manipulation of the outer ear canal together with a conductive hearing loss with a mostly high-frequent air-bone-gap and hypercompliant tympanometry with hypermobility of the tympanic membrane on pneumatic insufflation. Surgical repair of the fracture using bone cement has good hearing outcomes and leads to improvement in auditory symptoms.

Development and Validation of an Automated Classifier to Diagnose Acute Otitis Media in Children.

Importance: Acute otitis media (AOM) is a frequently diagnosed illness in children, yet the accuracy of diagnosis has been consistently low. Multiple neural networks have been developed to recognize the presence of AOM with limited clinical application. Objective: To develop and internally validate an artificial intelligence decision-support tool to interpret videos of the tympanic membrane and enhance accuracy in the diagnosis of AOM. Design, Setting, and Participants: This diagnostic study analyzed otoscopic videos of the tympanic membrane captured using a smartphone during outpatient clinic visits at 2 sites in Pennsylvania between 2018 and 2023. Eligible participants included children who presented for sick visits or wellness visits. Exposure: Otoscopic examination. Main Outcomes and Measures: Using the otoscopic videos that were annotated by validated otoscopists, a deep residual-recurrent neural network was trained to predict both features of the tympanic membrane and the diagnosis of AOM vs no AOM. The accuracy of this network was compared with a second network trained using a decision tree approach. A noise quality filter was also trained to prompt users that the video segment acquired may not be adequate for diagnostic purposes. Results: Using 1151 videos from 635 children (majority younger than 3 years of age), the deep residual-recurrent neural network had almost identical diagnostic accuracy as the decision tree network. The finalized deep residual-recurrent neural network algorithm classified tympanic membrane videos into AOM vs no AOM categories with a sensitivity of 93.8% (95% CI, 92.6%-95.0%) and specificity of 93.5% (95% CI, 92.8%-94.3%) and the decision tree model had a sensitivity of 93.7% (95% CI, 92.4%-94.9%) and specificity of 93.3% (92.5%-94.1%). Of the tympanic membrane features outputted, bulging of the TM most closely aligned with the predicted diagnosis; bulging was present in 230 of 230 cases (100%) in which the diagnosis was predicted to be AOM in the test set. Conclusions and Relevance: These findings suggest that given its high accuracy, the algorithm and medical-grade application that facilitates image acquisition and quality filtering could reasonably be used in primary care or acute care settings to aid with automated diagnosis of AOM and decisions regarding treatment.

When Can Children Perform Valsalva and Toynbee Maneuvers? An Exploratory Study.

OBJECTIVES: Barochallenge-induced Eustachian tube dysfunction (ETD) is difficult to diagnose because the examination is often normal during clinical assessment. In adults, functional tympanometry testing, performed by asking the patient to Valsalva and Toynbee while measuring the pressure shift, can aid in the diagnosis of ETD. However, standardized values do not exist in children. We aim to determine the age at which children can perform these maneuvers and the normative values in this population. METHODS: Patients with a normal basic ear examination 4 years and older, presenting to the pediatric Otolaryngology clinic, were recruited. Otoscopy, baseline tympanometry, followed by Valsalva and Toynbee maneuvers were performed. Because there are no pediatric norms, we hypothesized that children would achieve the same minimum normal pressure shift as cited in the adult literature (+20 daPa or higher for Valsalva and -20 daPa or lower for Toynbee). The data were analyzed using receiver operating characteristic curves and logistic regression. RESULTS: One hundred sixty-eight children (276 ears) were assessed. Participants as young as 4 years old were able to perform a Valsalva and Toynbee. Age cut-offs at which children achieved adult norms were 12.5 years ( p = 0.016) and 8.5 years ( p = 0.071) for Valsalva and Toynbee maneuvers, respectively. Mean pressure shift ranged from +29 to -36 daPa, and males were 2.5 times more likely to achieve Toynbee compared with females ( p = 0.006). CONCLUSIONS: Functional tympanometry testing may be used to help diagnose barochallenge-induced ETD in older children.

Otoscopy-assisted intranasal cyst marsupialization for congenital dacryocystocele.

We describe 3 infants with congenital dacryocystocele resistant to conservative treatment who were treated with a novel, simple intranasal cyst marsupialization (ICM) technique. Otoscopy-guided ICM was performed by an otolaryngologist in the manner of otoscopic myringotomy for cases with nasal cyst distension. All 3 infants were treated successfully by a single surgical procedure under topical anesthesia in an office setting.

Tympanic membrane perforations cannot be reliably detected using computed tomography based on 15 cadaver dogs.

The integrity of the tympanic membrane is an important factor when deciding treatment and therapeutic recommendations for dogs with ear disease; however, otoscopic examination may be difficult to perform due to features of external ear canal disease or patient compliance. CT is useful for the evaluation of middle ear disease, including cases in which middle ear disease is detected incidentally. The tympanic membrane is detectable using CT, but anecdotally, apparent focal defects or discontinuities of the tympanic membrane are often seen in patients with and without ear disease. The purpose of this prospective, observer agreement study was to determine if perforations of the tympanic membrane are reliably detectable on CT. Fifteen cadaver dogs underwent CT and video otoscopy to verify the integrity of each tympanic membrane. Cadavers were randomly assigned to have the tympanic membranes left intact or to undergo a myringotomy on either the left, the right, or both sides. CT was performed immediately following the myringotomies. Four blinded evaluators evaluated the pre- and post-myringotomy scans for a total of 30 scans (60 tympanic membranes). Average accuracy was low (44%), and interobserver agreement for all four evaluators was fair. Although the tympanic membrane is visible on CT, perforations of the tympanic membrane are unlikely to be accurately detected or excluded. The appearance of an intact tympanic membrane or defect in the membrane on CT should not be used as criteria to guide clinical treatment recommendations based on this cadaver model.

Registration of preoperative temporal bone CT-scan to otoendoscopic video for augmented-reality based on convolutional neural networks.

PURPOSE: Patient-to-image registration is a preliminary step required in surgical navigation based on preoperative images. Human intervention and fiducial markers hamper this task as they are time-consuming and introduce potential errors. We aimed to develop a fully automatic 2D registration system for augmented reality in ear surgery. METHODS: CT-scans and corresponding oto-endoscopic videos were collected from 41 patients (58 ears) undergoing ear examination (vestibular schwannoma before surgery, profound hearing loss requiring cochlear implant, suspicion of perilymphatic fistula, contralateral ears in cases of unilateral chronic otitis media). Two to four images were selected from each case. For the training phase, data from patients (75% of the dataset) and 11 cadaveric specimens were used. Tympanic membranes and malleus handles were contoured on both video images and CT-scans by expert surgeons. The algorithm used a U-Net network for detecting the contours of the tympanic membrane and the malleus on both preoperative CT-scans and endoscopic video frames. Then, contours were processed and registered through an iterative closest point algorithm. Validation was performed on 4 cases and testing on 6 cases. Registration error was measured by overlaying both images and measuring the average and Hausdorff distances. RESULTS: The proposed registration method yielded a precision compatible with ear surgery with a 2D mean overlay error of 0.65 ± 0.60 mm for the incus and 0.48 ± 0.32 mm for the round window. The average Hausdorff distance for these 2 targets was 0.98 ± 0.60 mm and 0.78 ± 0.34 mm respectively. An outlier case with higher errors (2.3 mm and 1.5 mm average Hausdorff distance for incus and round window respectively) was observed in relation to a high discrepancy between the projection angle of the reconstructed CT-scan and the video image. The maximum duration for the overall process was 18 s. CONCLUSIONS: A fully automatic 2D registration method based on a convolutional neural network and applied to ear surgery was developed. The method did not rely on any external fiducial markers nor human intervention for landmark recognition. The method was fast and its precision was compatible with ear surgery.

Acute otitis media diagnosis in childhood: still a problem in 2023?

BACKGROUND: Diagnosis of acute otitis media (AOM) in children can be challenging, given that symptoms are often non-specific or absent, and that the direct observation of the tympanic membrane in its entirety through otoscopy can sometimes be difficult. The aim of this study is to assess the diagnostic concordance in detection of AOM episodes between primary care paediatricians and physicians especially trained in paediatric otoscopy, and to characterize the most misleading elements in diagnostic failure. METHODS: Consecutive clinical charts of children regularly followed for recurrent AOM (RAOM, i.e.: >3 episodes in 6 months or > 4 episodes in 1 year) at our Otitis Media paediatric outpatient clinic were retrospectively screened, in order to collect any diagnosis of AOM episode (and the related clinical findings/middle ear complaints) performed by primary care paediatricians/emergency room paediatricians. Diagnosis of AOM episode was validated by the same experienced physician (FF) in case of otoscopic relief of a bulging eardrum with at least one of the following: hyperaemia or yellow-like colour. The diagnostic concordance in detection of AOM episodes between primary care/emergency room paediatricians and our internal validator was expressed as the percentage of matching diagnosis. RESULTS: One hundred and thirty-four single AOM episodes occurring in 87 children (mean age: 26.9 +/- 18.9 months) were included in the analysis. Diagnostic concordance in detection of AOM episodes between primary care/emergency room paediatricians and our internal validator was reported in 72.4% of cases. The most common pitfall found in our study was the misleading diagnosis of AOM in case of hyperaemic tympanic membrane without bulging (32/37 out of non-validated diagnoses). CONCLUSIONS: AOM diagnosis still represents a relevant issue among paediatricians in our country, and the presence of tympanic membrane hyperaemia without concomitant bulging can be confusing.

Remote paediatric ear examination comparing video-otoscopy and still otoscopy clinician rated outcomes.

OBJECTIVE: Telemedicine, particularly real time video-otoscopy in rural and remote Australia holds great potential in assessing and managing otology conditions. There is good evidence of store and forward images for assessment, however limited evidence exists for the use of real-time video-otoscopy. The objective of this study was to assess the validity of using real time video-otoscopy, compared to standard store and forward still image otoscopy, in a paediatric population. METHOD: Fifty-two paediatric tympanic membranes in 27 patients were examined and photographed by a telehealth facilitator with prior otoscope training. This occurred at two rural Western Australian health centre sites. These images were stored and forwarded to a tertiary paediatric hospital for otolaryngology department assessment on the day of real-time video-otoscopy consultation. During this consultation the same twenty-seven patients underwent real-time video-otoscopy assessment, which was recorded. Across six domains including, image quality, focus, light, cerumen amount, field of view and tympanic membrane landmarks, real-time video-otoscopy was compared against still image capture. The recording of each real-time video-otoscopy and still image tympanic membrane was assessed by two otology specialists for the ability to diagnose each as either normal or abnormal. An inter-rater reliability agreement was then calculated. RESULTS: There was greater image adequacy across five of the six domains for real time video-otoscopy compared to standard store and forward otoscopy images. Substantial agreement in diagnosing each tympanic membrane as either normal or abnormal between each rater was evident. CONCLUSION: This study supports the use of real time video-otoscopy during telemedicine consultation. With greater image quality, focus, light, field of view and identification of tympanic membrane landmarks video-otoscopy compared to still images has broad clinical applications. This includes primary assessment of the tympanic membrane and post operative follow-up clinical settings. Video-otoscopy offers a promising new way to over-come barriers in delivering ear health care in rural populations.

Diagnostic accuracy of the video otoscope in tympanic membrane perforation.

OBJECTIVE: The video otoscope has already proven to be useful for the diagnosis of several pathologies, so the objective of this study was to evaluate the diagnostic accuracy of the video otoscope in cases of tympanic membrane perforation. METHODS: This is a diagnostic accuracy study performed at the hearing health division of a tertiary-level referral hospital. Patients older than 8 years of age who had any symptom that could be related to perforation (otalgia, otorrhea, tinnitus, and/or hypoacusis) were invited to participate in the study. Participants were evaluated by three different diagnostic methods (otomicroscope, conventional otoscope, and video otoscope) performed by three different evaluators in a blind fashion. The microscope was considered the reference standard. RESULTS: 176 patients were evaluated, totaling 352 tympanic membranes. Twenty-seven tympanic membrane perforations were diagnosed by the microscope, a prevalence of 7.7%. The video otoscope showed a sensitivity of 85.2% (95% CI 81.5%‒88.9%), specificity of 98.1% (95% CI 96.7%‒99.5%) and accuracy of 97.1% (95% CI 95.4 %-98.8 %). The conventional otoscope showed a sensitivity of 96.3% (95% CI 94.3-98.3), specificity of 98.8% (95% CI 97.7-99.9) and accuracy of 98.6% (95% CI 97.4-99.8). The Kappa value between the microscope and the video otoscope was 0.8 and between the microscope and the conventional otoscope was 0.9. Regarding the participants' perception, 53.4% (p< 0.001) considered the video otoscope as the best method for understanding the tympanic membrane condition presented by them. CONCLUSIONS: The video otoscope showed relevant sensitivity and specificity for clinical practice in the diagnosis of tympanic membrane perforation. Moreover, this is an equipment that can facilitate the patient's understanding of the otologic pathology presented by him/her. In this regard, this method may be important for better patient compliance, requiring further studies to evaluate this hypothesis. LEVEL OF EVIDENCE: Is this diagnostic or monitoring test accurate? (Diagnosis)-Level 2 (Individual cross-sectional studies with consistently applied reference standard and blinding).

Feasibility of a Video Otoscope for Diagnosis of Otologic Pathology in the Pediatric Emergency Department.

OBJECTIVES: Performing pediatric otoscopy can be difficult secondary to patient compliance, which potentiates misdiagnosis and inaccurate treatment of acute otitis media. This study used a convenience sample to assess the feasibility of using a video otoscope for the examination of tympanic membranes in children presenting to a pediatric emergency department. METHODS: We obtained otoscopic videos using the JEDMED Horus + HD Video Otoscope. Participants were randomized to video or standard otoscopy, and a physician completed their bilateral ear examinations. In the video group, physicians reviewed otoscope videos with the patient's caregiver. The caregiver and physician completed separate surveys using a 5-point Likert Scale regarding perceptions of the otoscopic examination. A second physician reviewed each otoscopic video. RESULTS: We enrolled 213 participants in 2 groups (standard otoscopy, n = 94; video otoscopy, n = 119). We used Wilcoxon rank sum, Fisher exact test, and descriptive statistics to compare results across groups. For physicians, there were no statistically significant differences between groups with ease of device use, quality of otoscopic view, or diagnosis. There was moderate agreement between physician video otoscopic view satisfaction and slight agreement between physician video otologic diagnosis. Estimates of length of time to complete the ear examinations were longer more often for the video otoscope compared with standard for both caregivers (OR, 2.00; 95% confidence interval, 1.10-3.70; P = 0.02) and physicians (OR, 3.08; 95% confidence interval, 1.67-5.78; P < 0.01). There were no statistically significant differences between video and standard otoscopy with regard to caregiver perception of comfort, cooperation, satisfaction, or diagnosis understanding. CONCLUSIONS: Caregivers perceive that video otoscopy and standard otoscopy are comparable in comfort, cooperation, examination satisfaction, and diagnosis understanding. Physicians made a wider range of more subtle diagnoses with the video otoscope. However, examination length of time may limit the JEDMED Horus + HD Video Otoscope's feasibility in a busy pediatric emergency department.

Novel ENT live telehealth and live video-otoscopy clinics in remote Australia: outcomes and comparisons to traditional clinic models.

BACKGROUND: Coronavirus disease 2019 challenged the delivery of healthcare in Australia, disproportionately impacting vulnerable patients, including Aboriginal and/or Torres Strait Islander peoples and those living in remote regions. The otolaryngology service provided to remote Western Australia adapted to these barriers by altering clinical consultations to a digital model. METHODS: A review was undertaken of patients in regional Western Australia. Demographics and clinical outcomes from 20 live telehealth clinics were retrospectively reviewed and compared to 16 face-to-face clinics. RESULTS: The demographics of patients reviewed in both live telehealth and face-to-face clinics were similar, except for a larger proportion of Aboriginal and/or Torres Strait Islander patients utilising telehealth. The outcomes of patients reviewed through each model of care were comparable. Live video-otoscopy provided diagnostic quality images in 92 per cent of cases. CONCLUSION: The findings of our review suggest that, despite its limitations, a large proportion of ENT patients may be safely assessed through a live telehealth model.

Rigid normograde rhinoscopy-assisted traction-avulsion removal of small middle ear polyps from the auditory tube in five cats.

CASE SERIES SUMMARY: This study aims to describe a novel minimally invasive technique for the removal of small middle ear polyps from the auditory tube openings in cats. Five cats with clinical signs of otitis externa and/or otitis media, and/or upper respiratory tract inflammation were included. All cats underwent pharyngolaryngoscopy under anaesthesia, CT scan of the head, neck and thoracic cavity, video-otoscopic examination, retrograde nasopharyngoscopy and normograde rhinoscopy. The five cats described in this study were all found to have significant respiratory tract inflammation (rhinitis, sinusitis, nasopharyngitis, otitis media) with small polypous protrusions from the auditory tube openings. A normograde rhinoscopy-assisted traction-avulsion (RATA) removal of these small polyps was performed in all cases without complications. The rostral nasopharynx was visualised by a unilaterally normograde advanced rigid endoscope passing the choana, and polyps were removed using a grasping forceps introduced in the contralateral nostril. Telephone follow-up revealed clear improvement in all cases. One of the cases was re-evaluated with a CT scan and endoscopy 4 weeks after treatment. The CT scan showed a significant improvement with no abnormalities in both external ear canals and air opacity in both tympanic bullae. Video-endoscopic examination revealed intact tympanic membranes with mild chronic abnormalities and patent auditory tube openings upon normograde rhinoscopy. RELEVANCE AND NOVEL INFORMATION: Rigid normograde RATA is a novel, minimally invasive and effective technique for the removal of small middle ear polyps from auditory tube openings in cats with otitis media.

Evaluating the generalizability of deep learning image classification algorithms to detect middle ear disease using otoscopy.

To evaluate the generalizability of artificial intelligence (AI) algorithms that use deep learning methods to identify middle ear disease from otoscopic images, between internal to external performance. 1842 otoscopic images were collected from three independent sources: (a) Van, Turkey, (b) Santiago, Chile, and (c) Ohio, USA. Diagnostic categories consisted of (i) normal or (ii) abnormal. Deep learning methods were used to develop models to evaluate internal and external performance, using area under the curve (AUC) estimates. A pooled assessment was performed by combining all cohorts together with fivefold cross validation. AI-otoscopy algorithms achieved high internal performance (mean AUC: 0.95, 95%CI: 0.80-1.00). However, performance was reduced when tested on external otoscopic images not used for training (mean AUC: 0.76, 95%CI: 0.61-0.91). Overall, external performance was significantly lower than internal performance (mean difference in AUC: -0.19, p ≤ 0.04). Combining cohorts achieved a substantial pooled performance (AUC: 0.96, standard error: 0.01). Internally applied algorithms for otoscopy performed well to identify middle ear disease from otoscopy images. However, external performance was reduced when applied to new test cohorts. Further efforts are required to explore data augmentation and pre-processing techniques that might improve external performance and develop a robust, generalizable algorithm for real-world clinical applications.

Audiologic Measures in an Indigenous Community with A2ML1- and FUT2-Related Otitis Media.

Background: Many indigenous peoples are at elevated risk for otitis media, however there is limited information on hearing loss due to OM in these communities. An Indigenous Filipino community that has previously been described with an elevated prevalence of OM that is due to rare A2ML1 variants and a common FUT2 variant underwent additional phenological testing. In this study, we describe the audiologic profiles in A2ML1- and FUT2-related otitis media and the validity of otoscopy and genotyping for A2ML1 and FUT2 variants in screening for otitis media and hearing loss. Method: We analyzed A2ML1 and FUT2 genotypes together with demographic, otologic and audiologic data from tympanometry and hearing level assessments of 109 indigenous individuals. Results: We confirmed previous findings of a spectrum of nonsyndromic otitis media as associated with A2ML1 variants. A2ML1 and FUT2 variants were associated with high-frequency hearing loss at 4000 Hz. As expected, young age was associated with flat tympanograms, and eardrum perforations due to chronic otitis media were associated with severe-to-profound hearing loss across frequencies. Adding A2ML1 or FUT2 genotypes improved the validity of otoscopy as a screening test to rule out moderate-to-profound hearing loss. Conclusion: Continued multi-disciplinary management and audiologic follow-up using tympanometry and screening audiometry are needed to document and treat otitis media and prevent permanent hearing loss in the indigenous community.

Comparative validity of three simulation platforms for objective assessment of otoscopy skills.

OBJECTIVES: To assess the face, construct and content validity of three different platforms for otoscopy skills assessment, using a traditional otoscope with manikin, digital otoscope (Tympahealth) with manikin, and traditional otoscope with a low-cost model ear (SimEar). DESIGN: Prospective mixed methods study. SETTING: Tertiary hospital. PARTICIPANTS: Postgraduate trainees and expert assessors. MAIN OUTCOME MEASURES: Face and Content validity based on expert assessor ranking on each model and their feedback from semi-structured interviews. Construct validity based on Objective Structured Clinical Examination scores. RESULTS: Each platform differed in face, construct and content validity scores, with no one platform consistently outperforming others. Three main themes were identified during thematic analysis of expert assessor interviews: ability to assess what is seen, anatomical reality, and ease of use. The low-cost model showed greatest potential, where modification to include a silicone ear could lead to high validity with marginal increase in cost. CONCLUSION: Several modalities for assessing otoscopy skills exist, each with advantages and disadvantages. Modifications to a low-cost model, for use with either a traditional or digital otoscope, could prove to be the best model.

Value of simple otoscopy in diagnosing otitis media with effusion in children.

AIMS: To explore the value of otoscopy in diagnosing OME when performed by otorhinolaryngology, pediatrics, and primary care physicians; to evaluate the interobserver and intraobserver agreement of interpretation of otoscopy images. MATERIAL AND METHODS: A cross-sectional study using an anonymous mailed survey was used. We presented pre-recorded otoscopy images of pediatric patients to otorhinolaryngology, pediatrics, and primary care physicians (ten volunteer specialists and residents from each medical specialty). All participants had to answer "yes" or "no" if they considered that the image corresponded or not to an OME case, respectively. We considered that the images were positive for OME whenever the respective tympanogram was type B. RESULTS: Thirty-one otoscopy images and 1860 responses provided by sixty physicians were analyzed. The accuracy of otoscopy in diagnosing OME was highest in the Otolaryngologists group (mean 74.8%), with the worst rate observed in the primary care residents group (mean 51.3%). Overall sensitivity, specificity, and positive predictive value of otoscopy for diagnosing OME were significantly higher when performed by otorhinolaryngologists (75.8%, 72.8%, 66.8%, respectively). Fleiss' kappa showed that interobserver agreement was globally weak within each group of specialties, with overall better interobserver agreement observed among otorhinolaryngologists (κ = 0.30; 95% CI 0.27-0.32). CONCLUSION: According to our data, simple otoscopy as a single diagnostic method in pediatric OME is insufficient, even for otorhinolaryngologists. Current recommendations must be followed to improve diagnostic accuracy.

Machine Learning in Diagnosing Middle Ear Disorders Using Tympanic Membrane Images: A Meta-Analysis.

OBJECTIVE: To systematically evaluate the development of Machine Learning (ML) models and compare their diagnostic accuracy for the classification of Middle Ear Disorders (MED) using Tympanic Membrane (TM) images. METHODS: PubMed, EMBASE, CINAHL, and CENTRAL were searched up until November 30, 2021. Studies on the development of ML approaches for diagnosing MED using TM images were selected according to the inclusion criteria. PRISMA guidelines were followed with study design, analysis method, and outcomes extracted. Sensitivity, specificity, and area under the curve (AUC) were used to summarize the performance metrics of the meta-analysis. Risk of Bias was assessed using the Quality Assessment of Diagnostic Accuracy Studies-2 tool in combination with the Prediction Model Risk of Bias Assessment Tool. RESULTS: Sixteen studies were included, encompassing 20254 TM images (7025 normal TM and 13229 MED). The sample size ranged from 45 to 6066 per study. The accuracy of the 25 included ML approaches ranged from 76.00% to 98.26%. Eleven studies (68.8%) were rated as having a low risk of bias, with the reference standard as the major domain of high risk of bias (37.5%). Sensitivity and specificity were 93% (95% CI, 90%-95%) and 85% (95% CI, 82%-88%), respectively. The AUC of total TM images was 94% (95% CI, 91%-96%). The greater AUC was found using otoendoscopic images than otoscopic images. CONCLUSIONS: ML approaches perform robustly in distinguishing between normal ears and MED, however, it is proposed that a standardized TM image acquisition and annotation protocol should be developed. LEVEL OF EVIDENCE: NA Laryngoscope, 133:732-741, 2023.