Validity of Acoustic Measures Obtained Using Various Recording Methods Including Smartphones With and Without Headset Microphones.

PURPOSE: The goal of this study was to assess various recording methods, including combinations of high- versus low-cost microphones, recording interfaces, and smartphones in terms of their ability to produce commonly used time- and spectral-based voice measurements. METHOD: Twenty-four vowel samples representing a diversity of voice quality deviations and severities from a wide age range of male and female speakers were played via a head-and-thorax model and recorded using a high-cost, research standard GRAS 40AF (GRAS Sound & Vibration) microphone and amplification system. Additional recordings were made using various combinations of headset microphones (AKG C555 L [AKG Acoustics GmbH], Shure SM35-XLR [Shure Incorporated], AVID AE-36 [AVID Products, Inc.]) and audio interfaces (Focusrite Scarlett 2i2 [Focusrite Audio Engineering Ltd.] and PC, Focusrite and smartphone, smartphone via a TRRS adapter), as well as smartphones direct (Apple iPhone 13 Pro, Google Pixel 6) using their built-in microphones. The effect of background noise from four different room conditions was also evaluated. Vowel samples were analyzed for measures of fundamental frequency, perturbation, cepstral peak prominence, and spectral tilt (low vs. high spectral ratio). RESULTS: Results show that a wide variety of recording methods, including smartphones with and without a low-cost headset microphone, can effectively track the wide range of acoustic characteristics in a diverse set of typical and disordered voice samples. Although significant differences in acoustic measures of voice may be observed, the presence of extremely strong correlations (rs > .90) with the recording standard implies a strong linear relationship between the results of different methods that may be used to predict and adjust any observed differences in measurement results. CONCLUSION: Because handheld smartphone distance and positioning may be highly variable when used in actual clinical recording situations, smartphone + a low-cost headset microphone is recommended as an affordable recording method that controls mouth-to-microphone distance and positioning and allows both hands to be available for manipulation of the smartphone device.

Impact of BMI on Dyspnea and Need for Surgical Intervention in Bilateral Vocal Fold Immobility.

OBJECTIVE: Bilateral vocal fold paralysis (BVFP) and posterior glottic stenosis (PGS) are causes of bilateral vocal fold immobility (BVFI) and may cause shortness of breath, stridor, and need for surgical intervention. Although increased body mass index (BMI) is associated with restrictive breathing patterns in patients with normal upper airways, it is unclear how BMI impacts dyspnea and need for surgical intervention in BVFI patients. STUDY DESIGN: Retrospective cohort study. SETTING: Three tertiary academic centers in the United States. METHODS: Demographics, BMI, Dyspnea Index (DI), etiology, presence of tracheostomy and surgical intervention (dilation, tracheostomy, cordotomy, arytenoidectomy, open reconstruction) were collected. Primary outcomes included dyspnea measured by DI and need for surgery to improve airway. Linear regressions were performed to assess continuous outcomes. Mann-Whitney U-test was utilized to assess categorical outcomes. RESULTS: Among 121 patients, 52 presented with BVFP and 69 with PGS. Previous neck surgery was the most common cause of BVFI (40.2%). 44.3% of patients received a tracheostomy. Through multivariate linear regression, increased BMI was significantly associated with increased DI in the entire cohort (ß = .43, P = .016). Increased BMI was also associated with need for any surgical intervention (odds ratio [OR] = 1.07, 95% confidence interval [CI] = [1.01-1.13]) in the overall cohort. When stratifying our data, BMI was only significantly associated with DI in BVFP (ß = .496) and need for surgical intervention in PGS (OR = 1.11, 95% CI = [1.01-1.21]), although a positive trend was seen in all analyses. CONCLUSION: Increased BMI may correlate with worsening dyspnea symptoms and need for surgical intervention in patients with BVFI. Weight-loss-related counseling may benefit symptom management.

Voice as an AI Biomarker of Health-Introducing Audiomics.

This Viewpoint discusses the need to create standards for audiomics to identify unique audio biomarkers of health and disease­now possible because of more efficient voice data analysis available through the use of artificial intelligence (AI)­and to improve patient care.

Current Practices in Voice Data Collection and Limitations to Voice AI Research: A National Survey.

INTRODUCTION: Accuracy and validity of voice AI algorithms rely on substantial quality voice data. Although commensurable amounts of voice data are captured daily in voice centers across North America, there is no standardized protocol for acoustic data management, which limits the usability of these datasets for voice artificial intelligence (AI) research. OBJECTIVE: The aim was to capture current practices of voice data collection, storage, analysis, and perceived limitations to collaborative voice research. METHODS: A 30-question online survey was developed with expert guidance from the voicecollab.ai members, an international collaborative of voice AI researchers. The survey was disseminated via REDCap to an estimated 200 practitioners at North American voice centers. Survey questions assessed respondents' current practices in terms of acoustic data collection, storage, and retrieval as well as limitations to collaborative voice research. RESULTS: Seventy-two respondents completed the survey of which 81.7% were laryngologists and 18.3% were speech language pathologists (SLPs). Eighteen percent of respondents reported seeing 40%-60% and 55% reported seeing >60 patients with voice disorders weekly (conservative estimate of over 4000 patients/week). Only 28% of respondents reported utilizing standardized protocols for collection and storage of acoustic data. Although, 87% of respondents conduct voice research, only 38% of respondents report doing so on a multi-institutional level. Perceived limitations to conducting collaborative voice research include lack of standardized methodology for collection (30%) and lack of human resources to prepare and label voice data adequately (55%). CONCLUSION: To conduct large-scale multi-institutional voice research with AI, there is a pertinent need for standardization of acoustic data management, as well as an infrastructure for secure and efficient data sharing. LEVEL OF EVIDENCE: 5 Laryngoscope, 134:1333-1339, 2024.

Decision-Making in the Treatment of Idiopathic Subglottic Stenosis: A Survey of Laryngologists.

OBJECTIVE: Idiopathic subglottic stenosis (iSGS) is a rare condition involving the subglottic larynx and upper trachea, commonly affecting Caucasian females between 30 and 50 years of age. Despite its homogeneous presentation, clinical management for iSGS is yet to be standardized, leading to variability in outcomes between predominant interventions. In recognition of the heterogenicity in iSGS treatment and the need to improve patient outcomes, this study aimed to survey laryngologists to understand the factors influencing clinical decision-making and the incorporation of new treatment modalities for iSGS. METHODS: An online survey was sent to 145 academic laryngologists. The survey assessed respondents' professional backgrounds, experience treating iSGS, treatment algorithms, and how various patient factors affect management. RESULTS: Of the 87 (60%) laryngologists who responded to the survey, the most common clinical assessments were tracheoscopy/bronchoscopy (96.8%) and pulmonary function tests (43.6%). Endoscopic dilation (ED) was the most common primary treatment offered (97.5%): 28.7% of surveyed laryngologists offer SISI as a primary treatment, and 74.7% perform SISI as a planned postoperative treatment. The most common SISI protocol was repeated injections every 4-6 weeks for a series of 1-3 total injections. Notably, 9.2% perform the Maddern procedure. Routine algorithms of care involving surgery were most often based on prior experience and prior patient outcomes (75.9%) and conversations with colleagues (64.4%). Only 31% report using the same protocol learned during their fellowship training. CONCLUSION: This survey highlights significant variation in the management of patients with iSGS. Understanding the factors that influence decision-making may lead to potential standardization in heterogeneous treatment approaches and may improve clinical outcomes. LEVEL OF EVIDENCE: 5 Laryngoscope, 134:865-872, 2024.

Validation of an AI-assisted Treatment Outcome Measure for Gender-Affirming Voice Care: Comparing AI Accuracy to Listener's Perception of Voice Femininity.

OBJECTIVES: There is currently a lack of objective treatment outcome measures for transgender individuals undergoing gender-affirming voice care. Recently, Bensoussan et al developed an AI model that is able to generate a voice femininity rating based on a short voice sample provided through a smartphone application. The purpose of this study was to examine the feasibility of using this model as a treatment outcome measure by comparing its performance to human listeners. Additionally, we examined the effect of two different training datasets on the model's accuracy and performance when presented with external data. METHODS: 100 voice recordings from 50 cisgender males and 50 cisgender females were retrospectively collected from patients presenting at a university voice clinic for reasons other than dysphonia. The recordings were evaluated by expert and naïve human listeners, who rated each voice based on how sure they were the voice belonged to a female speaker (% voice femininity [R]). Human ratings were compared to ratings generated by (1) the AI model trained on a high-quality low-quantity dataset (voices from the Perceptual Voice Quality Database) (PVQD model), and (2) the AI model trained on a low-quality high-quantity dataset (voices from the Mozilla Common Voice database) (Mozilla model). Ambiguity scores were calculated as the absolute value of the difference between the rating and certainty (0 or 100%). RESULTS: Both expert and naïve listeners achieved 100% accuracy in identifying voice gender based on a binary classification (female >50% voice femininity [R]). In comparison, the Mozilla-trained model achieved 92% accuracy and the previously published PVQD model achieved 84% accuracy in determining voice gender (female >50% AI voice femininity). While both AI models correlated with human ratings, the Mozilla-trained model showed a stronger correlation as well as lower overall rating ambiguity than the PVQD-trained model. The Mozilla model also appeared to handle pitch information in a similar way to human raters. CONCLUSIONS: The AI model predicted voice gender with high accuracy when compared to human listeners and has potential as a useful outcome measure for transgender individuals receiving gender-affirming voice training. The Mozilla-trained model performed better than the PVQD-trained model, indicating that for binary classification tasks, the quantity of data may influence accuracy more than the quality of the data used for training the voice AI models.

Prospective >12 Months Outcomes After Vocal Fold Injection Medialization With Silk Microparticle-Hyaluronic Acid Material.

OBJECTIVE: Vocal fold injection medialization (VFIM) is widely used as an initial treatment for unilateral vocal fold paralysis (UVFP). Current practices employ materials that share the limitation of temporary clinical effect from variable resorption rates. A novel silk protein microparticle-hyaluronic acid-based material (silk-HA) has demonstrated cellular infiltration and tissue deposition that may portend a durable medialization effect. We report on ≥12 months outcomes after VFIM with silk-HA. METHODS: Prospective open-label study of patients with UVFP that elected treatment with VFIM with silk-HA. Blinded experts rated laryngeal stroboscopic exams. RESULTS: Seventeen patients with UVFP underwent VFIM with silk-HA. Twelve of the 17 patients have ≥12 months follow-up. Seven patients demonstrated durable treatment benefit ≥12 months after injection with median improvement of 19 (p = 0.0156) in VHI-10. There was no significant change in VHI-10 between 1 and 12 months for these patients. Blinded ratings indicated that 5/7 patients with sustained improvements in VHI-10 exhibited complete or touch glottal closure at 12 months. Two of the seven patients exhibited a small (<1 mm) glottal gap at 12 months. Seven patients experienced initial benefit with later regression 3-4 months after injection. CONCLUSION: VFIM with silk-HA can offer durable improvement in voice-related outcomes for UVFP past 12 months. A subset of patients treated with silk-HA experienced early loss of effect around 3-4 months postinjection. Clinical factors predictive of sustained treatment response to silk-HA injection require further exploration. LEVEL OF EVIDENCE: 3 Laryngoscope, 2023.

Machine Learning Analysis of Physical Activity Data to Classify Postural Dysfunction.

BACKGROUND: Machine learning (ML) analysis of biometric data in non-controlled environments is underexplored. OBJECTIVE: To evaluate whether ML analysis of physical activity data can be employed to classify whether individuals have postural dysfunction in middle-aged and older individuals. METHODS: A 1 week period of physical activity was measured by a waist-worn uni-axial accelerometer during the 2003-2004 National Health and Nutrition Examination Survey sampling period. Features of physical activity along with basic demographic information (42 variables) were paired with ML models to predict the success or failure of a standard 30 s modified Romberg test during which participants had their eyes closed and stood upon a 3-inch compliant surface. Model performance was evaluated by area under the receiver operating characteristic curve (AUC-ROC), balanced accuracy, and F1-score. RESULTS: The cohort was comprised of 1625 participants ≥40 years (median age 61, IQR 51-71). Approximately half (47%) were diagnosed with postural dysfunction having failed the binarized (pass/fail) scoring mechanism of the modified Romberg exam. Five ML models were trained on the classification task, achieving AUC values ranging from 0.67 to 0.73. The support vector machine (SVM) and a gradient-boosted model, XGBoost, achieved the highest AUC of 0.73 (SD 0.71-0.75). Age was the most important variable for SVM classification, followed by four features that evaluated accelerometer counts at various thresholds, including those delineating total, moderate, and moderate-vigorous activity. CONCLUSIONS: ML analysis of accelerometer-derived physical activity data to classify postural dysfunction in middle-aged and older individuals is feasible in real-world environments such as the home. LEVEL OF EVIDENCE: 3 Laryngoscope, 133:3529-3533, 2023.

Factors Associated With Dysphagia in Patients Undergoing Tracheal Resection.

Importance: Patients undergoing tracheal resection commonly experience dysphagia postoperatively, and the patient factors that predict severity and duration of symptoms are currently unclear. Objective: To determine the association of patient and surgical factors on postoperative dysphagia in adult patients undergoing tracheal resection. Design, Setting, and Participants: This was a retrospective cohort study of patients undergoing tracheal resection at 2 tertiary academic centers from February 2014 to May 2021. The centers included LAC+USC (Los Angeles County + University of Southern California) Medical Center and Keck Hospital of USC, both tertiary care academic institutions. Patients involved in the study underwent a tracheal or cricotracheal resection. Exposures: Tracheal or cricotracheal resection. Main Outcomes and Measures: The main outcome was dysphagia symptoms as measured by the functional oral intake scale (FOIS) on postoperative days (PODs) 3, 5, and 7, on the day of discharge, and at the 1-month follow-up visit. Demographics, medical comorbidities, and surgical factors were evaluated for association with FOIS scores at each time period using Kendall rank correlation and Cliff delta. Results: The study cohort consisted of 54 patients, with a mean (SD) age of 47 (15.7) years old, of whom 34 (63%) were male. Length of resection segment ranged from 2 to 6 cm, with a mean (SD) length of 3.8 (1.2) cm. The median (range) FOIS score was 4 (1-7) on PODs 3, 5, 7. On the day of discharge and at 1-month postoperative follow-up, the median (range) FOIS score was 5 (1-7) and 7 (1-7), respectively. Increasing patient age was moderately associated with decreasing FOIS scores at all measured time points (τ = -0.33; 95% CI, -0.51 to -0.15 on POD 3; τ = -0.38; 95% CI, -0.55 to -0.21 on POD 5; τ = -0.33; 95% CI, -0.58 to -0.08 on POD 7; τ = -0.22; 95% CI, -0.42 to -0.01 on day of discharge; and τ = -0.31; 95% CI, -0.53 to -0.09 at 1-month follow-up visit). History of neurological disease, including traumatic brain injury and intraoperative hyoid release, was not associated with FOIS score at any of the measured time points (δ = 0.03; 95% CI, -0.31 to 0.36 on POD 3; δ = 0.11; 95% CI, -0.28 to 0.47 on POD 5, δ = 0.3; 95% CI, -0.25 to 0.70 on POD 7; δ = 0.15; 95% CI, -0.24 to 0.51 on the day of discharge, and δ = 0.27; 95% CI, -0.05 to 0.53 at follow-up). Resection length was also not correlated with FOIS score with τ ranging from -0.04 to -0.23. Conclusions and Relevance: In this retrospective cohort study, most patients undergoing tracheal or cricotracheal resection experienced full resolution of dysphagia symptoms within the initial follow-up period. During preoperative patient selection and counseling, physicians should consider that older adult patients will experience greater severity of dysphagia throughout their postoperative course and delayed resolution of symptoms.

Variations in Practices and Preferences of Vocal Fold Injection Materials: A National Survey.

BACKGROUND: Vocal fold injection augmentation (VFI) is a common procedure for the treatment of glottic insufficiency. Material options for VFI and decisions regarding material selection are not standardized and often based on clinician preference. OBJECTIVE: This study aims to understand the variations in provider preference and utilization of injectable materials for VFI. METHODS: A 40-question survey was sent to 158 academic laryngologists. Questions pertained to the type of injectable materials used including brand preferences and rationale for preferences. RESULTS: Ninety-seven of the 158 laryngologists contacted participated in the survey (61.4%). The most frequently used injectable materials were Hyaluronic Acid (HA)-based products. Carboxymethylcellulose (CMC)-based products were preferred for trial augmentation (57.2%), HA-based products were preferred for acute/subacute vocal fold paralysis, presbyphonia, and sulcus/scar (54.2%, 61.5%, 44.7%, 41.7% respectively), and Calcium Hydroxyapatite (CaHA)-based products were preferred for long-term paralysis (28.1%). CMC-based products were discontinued by 21.8% of participants, largely due to quick material resorption. 17.8% of participants discontinued HA-based products largely due to adverse events and 26.0% abandoned CaHA-based products mostly due to inflammatory properties causing vocal fold stiffness and material unpredictability. Over 30% of respondents reported wanting to reinitiate micronized alloderm Cymetra® as an available injectable. CONCLUSION: Our survey demonstrated that there are significant variations in practice and preferences in regard to injectable material selection for VFI. As there is limited data on the direct material comparison, understanding the rationale behind these variations is crucial to guide new providers in material selection and provide information to patients undergoing these procedures. LEVEL OF EVIDENCE: 5 Laryngoscope, 133:1176-1183, 2023.

Multi-institutional search for patient factors associated with adverse events following tracheotomy.

BACKGROUND: Tracheotomy is a common procedure for otolaryngologists. The risk of complications is difficult to predict. This study aims to identify measurable preoperative indicators associated with adverse events following tracheotomy. METHODS: The charts of adults undergoing tracheotomy for respiratory failure at one of four university-affiliated hospitals between 1/2012 and 8/2018 were reviewed. Complications were analyzed in the context of demographics, physiologic parameters, and comorbidities. RESULTS: Among 507 tracheotomies performed, the most common complications included infection, bleeding, and cardiac arrest. Mortality was 39 % in patients with pulmonary hypertension, 42 % in those with ejection fraction ≤ 40 and 32 % in those with abnormal right ventricular function, double the rates in patients without each of these findings. CONCLUSION: Many critically ill tracheotomy patients experience significant rates of adverse events. Risk factors for mortality include ejection fraction ≤ 40, pulmonary hypertension, and abnormal ventricular function. These should be considered for use in preoperative counseling.

Artificial Intelligence and Laryngeal Cancer: From Screening to Prognosis: A State of the Art Review.

OBJECTIVE: This state of the art review aims to examine contemporary advances in applications of artificial intelligence (AI) to the screening, detection, management, and prognostication of laryngeal cancer (LC). DATA SOURCES: Four bibliographic databases were searched: PubMed, EMBASE, Cochrane, and IEEE. REVIEW METHODS: A structured review of the current literature (up to January 2022) was performed. Search terms related to topics of AI in LC were identified and queried by 2 independent reviewers. Citations of selected studies and review articles were also evaluated to ensure comprehensiveness. CONCLUSIONS: AI applications in LC have encompassed a variety of data modalities, including radiomics, genomics, acoustics, clinical data, and videomics, to support screening, diagnosis, therapeutic decision making, and prognosis. However, most studies remain at the proof-of-concept level, as AI algorithms are trained on single-institution databases with limited data sets and a single data modality. IMPLICATIONS FOR PRACTICE: AI algorithms in LC will need to be trained on large multi-institutional data sets and integrate multimodal data for optimal performance and clinical utility from screening to prognosis. Out of the data types reviewed, genomics has the most potential to provide generalizable models thanks to available large multi-institutional open access genomic data sets. Voice acoustic data represent an inexpensive and accurate biomarker, which is easy and noninvasive to capture, offering a unique opportunity for screening and monitoring of LA, especially in low-resource settings.

A Novel Grading System for Supraglottic Stenosis Based on Morphology and Functional Status.

OBJECTIVE: Currently, no classification system exists to grade the severity of supraglottic stenosis. The aim of this investigation was to (1) develop a novel grading system for supraglottic stenosis that can both enhance communication between providers and relay information about patient functional status and (2) determine the reliability of the grading system. METHODS: A retrospective analysis of patients with supraglottic stenosis at three institutions from 2010-2021 was conducted. After demographic data were collected, two focus group meetings of five laryngologists were held to develop a grading system based on functional status and morphology of stenosis seen on laryngoscopy. Three laryngologists then used the grading system to rate 20 case examples of supraglottic stenosis. Quadratic-weighted kappa coefficients were calculated to assess inter-rater and intra-rater reliabilities of the novel grading system. RESULTS: Twenty-eight patients were included. Epiglottic and arytenoid fixation were morphological features associated with worse functional outcomes such as requiring a G-tube or a tracheostomy, respectively. Inter-rater reliability was substantial to almost perfect (Kw = 0.79-0.81) and intra-rater reliability was almost perfect for all raters (0.88-1.0) when using the novel grading system. CONCLUSION: A grading system for supraglottic stenosis has been proposed with strong inter-rater and intra-rater reliabilities. The proposed system has the advantage of being descriptive of both patient functionality and morphology of the stenosis. LEVEL OF EVIDENCE: 3-According to the Oxford Center for Evidence-Based Medicine 2011 level of evidence guidelines, this non-randomized retrospective cohort study is classified as level 3 evidence Laryngoscope, 133:1442-1447, 2023.

Gauging Stakeholder Perspectives on Artificial Intelligence in Gender-Affirming Voice Care.

BACKGROUND: As the main objective outcome measure used in gender-affirming voice care (GAVC), fundamental frequency (f0 ) often fails to accurately reflect patient perceptions of their voice. Our team developed an artificial intelligence (AI) program that provides an alternative objective outcome measure that has the potential to more accurately align with patient perceptions. OBJECTIVE: To gauge stakeholder receptivity to the use of AI in GAVC before employing a novel outcome measure in transgender and nonbinary communities. METHODS: This prospective qualitative study used online focus groups composed of speech-language pathologists (SLPs), transgender men (TGM), transgender women (TGW), and nonbinary (NB) individuals. Participant age, race, gender, and geographic location were recorded. Each cohort participated in a series of two focus group sessions. The first session focused on participant experiences in GAVC, whereas the second ascertained participant perspectives on the use of AI in GAVC. Transcripts of each discussion were coded using Nvivo to perform inductive thematic analysis. RESULTS: Seven SLPs, seven TGW, three TGM, and two NB individuals (mean [range] age, 35.5 [26-48] years) participated. Transgender and nonbinary participants were generally amenable to the technology, whereas SLPs were more hesitant about its use. Positive findings included appreciation for AI as an objective outcome measure and enthusiasm for its potential to longitudinally track progress. Hesitations concerned the actionability of using the AI and unease about the black box nature of the AI's analysis. CONCLUSION: Transgender and NB individuals were receptive to the use of AI technology in GAVC, whereas SLPs were more apprehensive about using AI. LEVEL OF EVIDENCE: NA Laryngoscope, 133:1698-1705, 2023.

The Effect of Subglottic Steroid injection on Blood Glucose in a Cohort of Patients With Subglottic Stenosis.

OBJECTIVE: The objective of this work was to assess the effect of a single intralesional subglottic steroid injection on fasting blood glucose over 7 days in a cohort of patients with subglottic stenosis. METHODS: A prospective cohort study of patients undergoing intralesional steroid injections at a tertiary academic center. Patients had baseline bloodwork performed, including fasting blood glucose (FBG), and hemoglobin A1C levels. Demographic data and risk factors were collected. Fasting capillary glucose (FCG) was measured using a capillary glucometer and performed by patients daily from days 0 to 7 after a single injection of Triamcinolone into the subglottic airway. Data were analyzed using descriptive and comparative statistics. RESULTS: Eleven patients were enrolled, and 10 completed data collection over 7 days. All were female, with a mean age of 52.6 years (SD 17.5). Two patients were diabetic (non-insulin dependent). There was a statistically significant increase in FCG on day 1 post-injection (mean = 122.4 mg/dl compared to 100.7 mg/dl) that normalized for all patients within 24-72 h. The mean increase in FCG was 21.5% (SD 22.5%) of the initial value for the cohort. The diabetic group had statistically significant higher glucose values on day 1 compared to the non-diabetic group (146.5 mg/dl compared to 117.0 mg/dl). CONCLUSION: A single subglottic steroid injection appears to cause a transient increase in FCG 1 day post injection, which resolves after 24-72 h and can be more pronounced in diabetic patients. LEVEL OF EVIDENCE: 4 Laryngoscope, 133:1590-1594, 2023.

Intractable Bleeding Following an Awake Bedside Injection Laryngoplasty in a Patient on KVAD ECMO.

INTRODUCTION: Awake injection laryngoplasty (IL) is becoming increasingly utilized in the inpatient setting, especially as a therapeutic option for patients with vocal fold immobility immediately following cardiothoracic surgery. While prior studies consistently demonstrate complication rates below 3%, significant bleeding has not been reported as a major complication in any awake IL case series. The objective of this report is to highlight a case of intractable bleeding following awake inpatient bedside IL in a patient on KVAD (Koji Takeda Ventricular Assist Device) extracorporeal membrane oxygenation (ECMO). METHODS: Case Report. RESULTS: A 24-year-old female admitted to the cardiac ICU for asystole was placed on KVAD ECMO for heart failure of unknown etiology. She was extubated and listed for cardiac transplant. On postoperative day 14, she underwent a left vocal fold injection at bedside to treat fold paralysis with a large glottic gap causing her complete aphonia, dysphagia, and chronic aspiration. Seven hours post procedure, the patient had to be reintubated due to intractable bleeding. A direct laryngoscopy was performed at bedside and continuous trickle of blood from the injection site at the superior posterior lateral surface of the vocal fold was stopped using a combination of epi-pledgets and hemostatic matrix. The same procedure had to be performed again due to further bleeding 2 days later and permanent hemostasis was achieved. During the week post injection, the patient required transfusion of 5 units of pRBC's. One month later the patient underwent successful orthotopic heart transplantation and was transferred from the ICU to a stepdown unit, and then a rehabilitation unit. No further IL hemorrhage occurred. CONCLUSION: Although a few studies have discussed the safety of IL in patients receiving anticoagulation, this case report demonstrates intractable bleeding requiring intubation and intervention to achieve hemostasis in a patient on KVAD ECMO. This report highlights the importance of weighing the risks and benefits of vocal fold injection in this patient population.

Development of Cushing Syndrome After Serial In-office Subglottic Steroid Injections.

In-office subglottic intralesional steroid injections (SILSI) have gained popularity as an adjunct to operating room dilation in the treatment of subglottic stenosis. They are generally thought to have a low risk profile for development of systemic side effects. Here, we present a case of a 55 year old woman who developed symptoms of Cushing syndrome after receiving SILSI, including weight gain, striae, dorsal hump and alopecia. This case illustrates that despite the localized nature of SILSI, there is still a risk of developing systemic effects as a result of the treatment. Laryngoscope, 132:942-943, 2022.

A Comparison of an Artificial Intelligence Tool to Fundamental Frequency as an Outcome Measure in People Seeking a More Feminine Voice.

OBJECTIVES/HYPOTHESIS: An artificial intelligence (AI) tool was developed using audio clips of cis-male and cis-female voices based on spectral analysis to assess %probability of a voice being perceived as female (%Prob♀). This program was validated with 92% accuracy in cisgender speakers. The aim of the study was to assess the relationship of fo on %Prob♀ by a validated AI tool in a cohort of trans females who underwent intervention to feminize their voice with behavioral modification and/or surgery. STUDY DESIGN: Cohort study. METHODS: Fundamental frequency (fo ) from prolonged vowel sounds (fo /a/) and fo from spontaneous speech (fo -sp) were measured using the Kay Pentax Computerized Speech Lab (Montvale, NJ) in trans females postintervention. The same voice samples were analyzed by the AI tool for %Prob♀. Chi-square analysis and regression models were performed accepting >50% Prob♀ as female voice. RESULTS: Forty-two patients were available for analysis after intervention. fo -sp post-treatment was positively correlated with %Prob♀ (R = 0.645 [P < .001]). Chi-square analysis showed a significant association between AI %Prob♀ >50% for the speech samples and fo -sp >160 Hz (P < .01). Sixteen of 42 patients reached an fo -sp >160 Hz. Of these, the AI program only perceived nine patients as female (>50 %Prob♀). CONCLUSION: Patients with fo -sp >160 Hz after feminization treatments are not necessarily perceived as having a high probability of being female by a validated AI tool. AI may represent a useful outcome measurement tool for patients undergoing gender affirming voice care. LEVEL OF EVIDENCE: 3 Laryngoscope, 131:2567-2571, 2021.

Deep Learning for Voice Gender Identification: Proof-of-concept for Gender-Affirming Voice Care.

OBJECTIVES/HYPOTHESIS: The need for gender-affirming voice care has been increasing in the transgender population in the last decade. Currently, objective treatment outcome measurements are lacking to assess the success of these interventions. This study uses neural network models to predict binary gender from short audio samples of "male" and "female" voices. This preliminary work is a proof-of-concept for further work to develop an AI-assisted treatment outcome measure for gender-affirming voice care. STUDY DESIGN: Retrospective cohort study. METHODS: Two hundred seventy-eight voices from male and female speakers from the Perceptual Voice Qualities Database were used to train a deep neural network to classify voices as male or female. Each audio sample was mapped to the frequency domain using Mel spectrograms. To optimize model performance, we performed 10-fold cross validation of the entire dataset. The dataset was split into 80% training, 10% validation, and 10% test. RESULTS: Overall accuracy of 92% was obtained, both when considering the accuracy per spectrum and per patient metric. The accuracy of the model was higher for recognizing female voices (F1 score of 0.94) compared to male voices (F1 score of 0.87). CONCLUSIONS: This proof of concept study shows promising performance for further development of an AI-assisted tool to provide objective treatment outcome measurements for gender affirming voice care. LEVEL OF EVIDENCE: 3 Laryngoscope, 131:E1611-E1615, 2021.