Deep, Nicholas

Abstract Purpose Failed stapes surgery presents a diagnostic challenge requiring high-resolution imaging to guide revision. Energy-integrating detector CT (EID-CT) is widely used but limited by spatial resolution and metal artifact. Photon-counting CT (PCCT) offers improved spatial resolution and reduced noise, with potential advantages in temporal bone imaging. The purpose of this study was to compare PCCT and EID-CT in identifying etiologies of failed primary stapes surgery, with emphasis on diagnostic accuracy, interobserver agreement, and interpreter confidence Methods In this retrospective study, 38 patients with failed primary stapes surgery who underwent successful revision surgery and pre-revision CT were included; 18 underwent PCCT and 20 underwent EID-CT. Three expert reviewers independently determined the cause of surgical failure, compared against the reference standard of intraoperative findings at revision. Diagnostic accuracy, interobserver agreement (Fleiss kappa), and reader confidence were compared between modalities. Results Overall interobserver agreement was substantially higher with PCCT (κ = 0.75; 95% CI, 0.60–0.90) than EID-CT (κ = 0.44; 95% CI, 0.30–0.59), with improved agreement observed across etiologies of failure; confidence intervals were non-overlapping for overall agreement and ossicular discontinuity. Consensus diagnostic accuracy was comparable between modalities (88.9% PCCT vs 85% EID-CT; p = 1.00), as was interpreter confidence (4.33 PCCT vs 4.17 EID-CT; p = 0.29). Conclusion In this small cohort, PCCT and EID-CT demonstrated similar diagnostic accuracy and reader confidence, whereas PCCT demonstrated significantly higher interobserver agreement, suggesting more reproducible assessment of failed stapes surgery and a potential role in supporting revision surgical planning.

Abstract Objectives To address the gap between artificial intelligence (AI) model development and clinical implementation by developing a machine-learning model to predict cochlear implant (CI) eligibility based on routine hearing tests, while illustrating 3 practical lessons: reformulation for clinical utility, selection of problem-specific metrics, and handling input data variability. Materials and Methods Data were extracted from adult patients who underwent behavioral audiometric and subsequent CI candidacy testing (Arizona Biomedical and Consonant-Nucleus-Consonant tests) at Mayo Clinic between 2011 and 2023. Regression models were initially developed to predict AzBio and CNC scores using features from routine audiograms. Models were then reformulated into binary classification tasks using clinical thresholds (AzBio ≤60%; CNC <50%) and reported sensitivity and positive predictive value (PPV). Input data variability was assessed by analyzing test-retest differences in isophonemes scores. Results Binary classification models achieved sensitivity of 90.4% with PPV 80.2% at CNC 50/AzBio 60, outperforming referral heuristics (60/60 and 75/40). Applying the models to 50 700 historical audiograms identified 525 patients at PPV 90% (sensitivity 58%), 396 at PPV 94% (sensitivity 44%), and 1582 at PPV 71% (sensitivity 94%). Variability analyses indicated that attainable precision is bounded by noise in key predictors. Conclusion This study demonstrates how AI models can be translated into clinically actionable tools by reformulating regression outputs, selecting metrics aligned with clinical priorities, and addressing data variability. These strategies provide a scalable framework for implementing AI in health care and improving decision-making in diverse clinical contexts.

Objectives To compare electrical stimulation and speech perception in vestibular schwannoma (VS) patients across treatment modalities and standard cochlear implant (CI) patients. Study Design Retrospective review. Setting Tertiary academic center. Patients Sixty-seven CI ears consisting of 23 standard CI controls and 44 VS patients with ipsilateral CI. VS patients were separated by treatment modality: 24 microsurgical resection, 15 radiation therapy, and 5 observation. Eleven VS patients with CI did not receive auditory sensation from electrical stimulation. Interventions VS microsurgical resection, radiation therapy, and observation. Main Outcome Measures Threshold (T-levels), comfort (C-levels), consonant-nucleus-consonant (CNC) word score, and AzBio sentence recognition scores. Results Patients who underwent microsurgical resection required significantly higher (p < 0.001) T-levels and C-levels compared with the standard CI group. Stimulation levels in the VS radiation therapy and observation groups were higher but not significantly different compared with standard CI controls. Overall rate of CI nonstimulation across all VS patients was 25%: 38% for microsurgery and 13% for radiation. Average CNC score for microsurgery group was 18% compared with 44% for radiation, 55% for observation, and 68.5% for standard CI controls. CNC word and AzBio scores were significantly lower (p < 0.001) for the microsurgery group compared with standard CI group. Conclusions Resection may negatively affect auditory nerve function, leading to higher stimulation levels, increased risk of nonstimulation, and poorer postoperative CI speech outcomes compared with observation or radiosurgery. When medically appropriate, nonsurgical VS management may be used to allow for optimization of CI programming and postoperative hearing outcomes.

Abstract Background Vestibular migraine (VM) is one of the common causes of episodic dizziness, but it is underdiagnosed and poorly understood. Previous research suggests that otolith reflex pathway performance is often impaired in this patient group, leading to altered perception of roll plane stimuli. Clinically, this perception can be measured with subjective visual vertical (SVV) testing. Purpose The aim of this study is to compare static SVV performance (absolute mean SVV tilt, variance) in a cohort of patients diagnosed with VM to results obtained from clinically derived normative data. Study Design Retrospective case review. Study Sample Ninety-four consecutive patients between 18 and 65 years of age diagnosed with VM were included in this comparison to clinically derived normative data. Data Collection and Analysis Retrospective chart review was completed. Demographic data, symptom report, and vestibular laboratory results were documented. SVV performance was documented in terms of absolute mean SVV tilt and response variance. Results Abnormal mean SVV tilt was described in 54% (n = 51) of patients with VM. Including abnormal response variance increased those identified with abnormal presentation to 67% (n = 63). Laboratory findings were insignificant for semicircular canal function, but of those with abnormal ocular vestibular myogenic potential results (n = 30), 77% (n = 23) demonstrated both abnormal SVV and utriculo-ocular reflex performance. There were no associations noted for SVV performance and demographic or other self-report variables. Conclusion Absolute mean SVV tilt and response variance are often abnormal in patients diagnosed with VM. These findings support theories suggesting atypical intralabyrinthine integration within the vestibular nuclei and cerebellar nodular pathways.