scholarly journals Uncovering the important acoustic features for detecting vocal fold paralysis with explainable machine learning

2020 ◽  
Author(s):  
Daniel M. Low ◽  
Gregory Randolph ◽  
Vishwanatha Rao ◽  
Satrajit S. Ghosh ◽  
Phillip C. Song
Keyword(s):  
Machine Learning ◽  
Vocal Fold ◽  
Characteristic Curve ◽  
Case Series ◽  
Vocal Fold Paralysis ◽  
Acoustic Parameter ◽  
Control Group ◽  
Acoustic Features ◽  
Level Of Evidence ◽  
And Task

AbstractObjectivesTo detect unilateral vocal fold paralysis (UVFP) from voice recordings using an explainable model of machine learning.Study DesignCase series - retrospective with a control group.MethodsPatients with confirmed UVFP through endoscopic examination (N=77) and controls with normal voices matched for age and sex (N=77) were included. Two tasks were used to elicit voice samples: reading the Rainbow Passage and sustaining phonation of the vowel /a/. The eighty-eight extended Geneva Minimalistic Acoustic Parameter Set (eGeMAPS) features were extracted as inputs for four machine learning models of differing complexity. Training and testing were performed using bootstrapped cross-validation. SHAP was used to identify important features.ResultsThe median Area Under the Receiver Operating Characteristic Curve (ROC AUC) score ranged from 0.79 to 0.87 depending on model and task. After removing redundant features for explainability, the highest median ROC AUC score was 0.84 using only 13 features for the vowel task and 0.87 using 39 features for the reading task. The most important features included intensity measures, mean MFCC1, mean F1 amplitude and frequency, and shimmer variability depending on model and task.ConclusionUsing the largest dataset studying UVFP to date, we achieve high performance from just a few seconds of voice recordings while discovering which acoustic features are important across models. Notably, we demonstrate that the models use different combinations of features to achieve similar effect sizes. Overall the categories of features related to vocal fold physiology were conserved across the models. Machine learning thus provides a mechanism to detect UVFP and contextualize the accuracy relative to both model architecture and pathophysiology.Level of EvidenceType 3

Application of a Computer Vision Tool for Automated Glottic Tracking to Vocal Fold Paralysis Patients

2021 ◽  
pp. 019459982198960
Author(s):  
Tiffany V. Wang ◽  
Nat Adamian ◽  
Phillip C. Song ◽  
Ramon A. Franco ◽  
Molly N. Huston ◽  
...  
Keyword(s):  
Vocal Fold ◽  
Assessment Tool ◽  
Medical Center ◽  
Characteristic Curve ◽  
Academic Medical Center ◽  
Vocal Fold Paralysis ◽  
Control Group ◽  
Large Patient ◽  
Patient Reported ◽  
Before And After

Objectives (1) Demonstrate true vocal fold (TVF) tracking software (AGATI [Automated Glottic Action Tracking by artificial Intelligence]) as a quantitative assessment of unilateral vocal fold paralysis (UVFP) in a large patient cohort. (2) Correlate patient-reported metrics with AGATI measurements of TVF anterior glottic angles, before and after procedural intervention. Study Design Retrospective cohort study. Setting Academic medical center. Methods AGATI was used to analyze videolaryngoscopy from healthy adults (n = 72) and patients with UVFP (n = 70). Minimum, 3rd percentile, 97th percentile, and maximum anterior glottic angles (AGAs) were computed for each patient. In patients with UVFP, patient-reported outcomes (Voice Handicap Index 10, Dyspnea Index, and Eating Assessment Tool 10) were assessed, before and after procedural intervention (injection or medialization laryngoplasty). A receiver operating characteristic curve for the logistic fit of paralysis vs control group was used to determine AGA cutoff values for defining UVFP. Results Mean (SD) 3rd percentile AGA (in degrees) was 2.67 (3.21) in control and 5.64 (5.42) in patients with UVFP ( P < .001); mean (SD) 97th percentile AGA was 57.08 (11.14) in control and 42.59 (12.37) in patients with UVFP ( P < .001). For patients with UVFP who underwent procedural intervention, the mean 97th percentile AGA decreased by 5 degrees from pre- to postprocedure ( P = .026). The difference between the 97th and 3rd percentile AGA predicted UVFP with 77% sensitivity and 92% specificity ( P < .0001). There was no correlation between AGA measurements and patient-reported outcome scores. Conclusions AGATI demonstrated a difference in AGA measurements between paralysis and control patients. AGATI can predict UVFP with 77% sensitivity and 92% specificity.

scholarly journals Using Innovative Acoustic Analysis to Predict the Postoperative Outcomes of Unilateral Vocal Fold Paralysis

2016 ◽  
Vol 2016 ◽  
pp. 1-9 ◽  
Author(s):  
Yung-An Tsou ◽  
Yi-Wen Liu ◽  
Wen-Dien Chang ◽  
Wei-Chen Chen ◽  
Hsiang-Chun Ke ◽  
...  
Keyword(s):  
Vocal Fold ◽  
Linear Prediction ◽  
Acoustic Analysis ◽  
Characteristic Curve ◽  
Voice Quality ◽  
Treatment Strategies ◽  
Vocal Fold Paralysis ◽  
Unilateral Vocal Fold Paralysis ◽  
Outcome Predictor ◽  
Laryngeal Framework Surgery

Objective.Autologous fat injection laryngoplasty is ineffective for some patients with iatrogenic vocal fold paralysis, and additional laryngeal framework surgery is often required. An acoustically measurable outcome predictor for lipoinjection laryngoplasty would assist phonosurgeons in formulating treatment strategies.Methods.Seventeen thyroid surgery patients with unilateral vocal fold paralysis participated in this study. All subjects underwent lipoinjection laryngoplasty to treat postsurgery vocal hoarseness. After treatment, patients were assigned to success and failure groups on the basis of voice improvement. Linear prediction analysis was used to construct a new voice quality indicator, the number of irregular peaks (NIrrP). It compared with the measures used in the Multi-Dimensional Voice Program (MDVP), such as jitter (frequency perturbation) and shimmer (perturbation of amplitude).Results.By comparing the [i] vowel produced by patients before the lipoinjection laryngoplasty (AUC = 0.98, 95% CI = 0.78–0.99), NIrrP was shown to be a more accurate predictor of long-term surgical outcomes than jitter (AUC = 0.73, 95% CI = 0.47–0.91) and shimmer (AUC = 0.63, 95% CI = 0.37–0.85), as identified by the receiver operating characteristic curve.Conclusions.NIrrP measured using the LP model could be a more accurate outcome predictor than the parameters used in the MDVP.

scholarly journals Evaluation of Diffusional Characteristics and Microstructure in Unilateral Vocal Fold Paralysis Using Diffusion Tensor Imaging

2019 ◽  
pp. 014556131987472
Author(s):  
Jie Cai ◽  
Logan Klein ◽  
Pinhong Wu ◽  
Azure Wilson ◽  
Xinlin Xu ◽  
...  
Keyword(s):  
Vocal Fold ◽  
Diffusion Tensor ◽  
Vocal Fold Paralysis ◽  
Cross Sectional Area ◽  
Control Group ◽  
Sectional Area ◽  
Unilateral Vocal Fold Paralysis ◽  
Cross Sectional ◽  
Injury Group ◽  
The Cross

Objective: To investigate the value of diffusion tensor imaging (DTI) in the evaluation of vocal fold tissue microstructure after recurrent laryngeal nerve (RLN) injury. Methods: Six canines were divided into 2 groups: a unilateral vocal fold paralysis group (n = 4) and a control group (n = 2). The RLN was cut in the unilateral vocal fold paralysis group, and no intervention was applied in the control group. After 4 months, the canines’ larynges were removed and placed in a small animal magnetic resonance imaging (MRI) system (9.4T BioSpec MRI; Bruker, Germany). After scanning, the vocal folds were isolated, sectioned, and stained. The slides were then analyzed for the cross-sectional area and muscle fiber density through feature extraction technology. Pearson correlation analysis was performed on the DTI scan and histological section extraction results. Results: In the vocal fold muscle layer, the fractional anisotropy (FA) of the unilateral RLN injury group was higher than that of the control group, and the Tensor Trace was lower than that of the control group. This difference was statistically significant, P < .05. In the lamina propria, the FA of the unilateral RLN injury group was lower than that of the control group, P > .05, and the Tensor Trace was lower than that of the control group, P < .05. The muscle fiber cross-sectional area of the RLN injury group was significantly smaller than the control group with statistical significance, P < .05, and the density of muscle fibers was lower, P < .05. The correlation coefficient between FA and the cross-sectional area was −0.838, P = .002, and .726; P = .017 between Tensor Trace and the cross-sectional area. Conclusion: Diffusion tensor imaging is an effective method to assess the changes in the microstructure of atrophic vocal fold muscle tissue after RLN injury.

Cardiothoracic Patients with Unilateral Vocal Fold Paralysis: Pneumonia Rates Following Injection Laryngoplasty

2020 ◽  
Vol 129 (11) ◽  
pp. 1129-1134 ◽  
Author(s):  
Jason H. Barnes ◽  
Diana M. Orbelo ◽  
Michael F. Armstrong ◽  
Semirra L. Bayan ◽  
Christine M. Lohse ◽  
...  
Keyword(s):  
Aspiration Pneumonia ◽  
Vocal Fold ◽  
Proportional Hazards ◽  
Retrospective Chart Review ◽  
Vocal Fold Paralysis ◽  
Cox Proportional Hazards ◽  
Cardiothoracic Surgery ◽  
Injection Laryngoplasty ◽  
Unilateral Vocal Fold Paralysis ◽  
Level Of Evidence

Objective: Recurrent laryngeal nerve injury is a potential complication of cardiothoracic surgery and cause of unilateral vocal fold paralysis (UVFP). Injection laryngoplasty (IL) is an intervention offered to patients with UVFP to alleviate symptoms including dysphagia, dysphonia and weak cough. There is no definitive evidence that IL prevents pneumonia. In this study, we compare rates of pneumonia in patients with UVFP secondary to cardiothoracic surgery who did or did not undergo IL. Methods: A retrospective chart review identified patients diagnosed with UVFP by an otolaryngologist using flexible laryngoscopy following cardiothoracic surgery from January 1, 2008 to December 31, 2017. Each subject was grouped by IL status and assessed for subsequent pneumonia within 6 months of their diagnosis of UVFP. The association of IL with pneumonia was evaluated using Cox proportional hazards regression. Results: Of 92 patients who met inclusion criteria, 35 (38%) underwent IL and 57 (62%) did not. Twenty patients developed pneumonia, four who had undergone IL and 16 who had not; 12 patients developed aspiration pneumonia including two having undergone IL and 10 who had not. Those who had IL were less likely to develop total pneumonia compared to those who had not (HR = 0.33, P = .045). The protective effect of IL was not as clearly sustained when measuring for aspiration pneumonia, specifically (HR = 0.34; P = .10). Discussion: Injection laryngoplasty may reduce the risk of pneumonia in patients with UVFP secondary to cardiothoracic surgery; however, further research is needed to quantify the potential protective nature of IL in this patient population. Level of evidence: 3 (A retrospective cohort study).

scholarly journals Determining the Patient Acceptable Symptomatic State for Patients Undergoing Arthroscopic Partial Meniscectomy in the Knee

2020 ◽  
Vol 48 (4) ◽  
pp. 847-852 ◽  
Author(s):  
Tim Dwyer ◽  
Thomas Zochowski ◽  
Darrell Ogilvie-Harris ◽  
John Theodoropoulos ◽  
Daniel Whelan ◽  
...  
Keyword(s):  
Roc Analysis ◽  
Characteristic Curve ◽  
Case Series ◽  
International Knee Documentation Committee ◽  
Partial Meniscectomy ◽  
Consecutive Series ◽  
Evaluation Tool ◽  
Arthroscopic Partial Meniscectomy ◽  
Level Of Evidence ◽  
Patient Reported

Background: Arthroscopic partial meniscectomy is one of the most common procedures in orthopaedic surgery. The patient acceptable symptomatic state (PASS), which defines a level of symptoms above which patients consider themselves well, remains to be well-defined in this population. Purpose: Using an anchor-based approach, our goal was to determine the 1-year PASS for the Knee injury and Osteoarthritis Outcome Score (KOOS), the International Knee Documentation Committee (IKDC) Subjective Knee Form, the Western Ontario Meniscal Evaluation Tool (WOMET), and the Marx Activity Scale (MAS) in patients who were treated with partial knee meniscectomy. Study Design: Case series; Level of evidence, 4. Methods: A consecutive series of patients with knee meniscal tears and a Kellgren-Lawrence grade of 0 to 2 treated with arthroscopic partial meniscectomy were eligible. The KOOS (0-100), IKDC (0-100), WOMET (0-100), and MAS (0-16) were administered at baseline and 12 months postoperatively. An external anchor question at 1 year postoperatively was used to determine PASS values. A receiver operating characteristic curve (ROC) analysis was used to determine the PASS value at which patients considered their status to be satisfactory. Results: The study included 110 patients (mean ± SD age, 53.8 ± 12.0 years), 57.3% were male, and the follow-up rate was 82%. In total, 70% of patients had an Outerbridge arthroscopic grade of 2 or lower. Based on ROC analysis, the 1-year postoperative PASS values (sensitivity, specificity) were 64.3 (47.8, 100.0) for KOOS Symptoms, 81.6 (71.6, 100.0) for KOOS Pain, 82.4 (82.1, 86.4) for KOOS Function in Daily Living, 71.0 (62.7, 81.8) for KOOS Function in Sport and Recreation, 51.0 (83.6, 95.5) for KOOS Knee-Related Quality of Life, 56.2 (82.1, 100.0) for IKDC, 58.5 (79.1, 100.0) for WOMET, and 7.0 (44.8, 68.2) for MAS. Baseline scores did not affect the PASS threshold across the different instruments. However, patients with higher baseline scores were more likely to achieve the PASS for the KOOS Symptoms (odds ratio [OR], 2.808; P = .047), IKDC (OR, 4.735; P = .006), and WOMET (OR, 2.985; P = .036). Age, sex, and cartilage status were not significantly related to the odds of achieving the PASS for any of the patient-reported outcome measures. Conclusion: These findings allow researchers and clinicians to determine whether partial meniscectomy is meaningful to patients at the individual level and will be helpful for responder analysis in future trials related to the treatment of meniscal abnormality.

Unilateral True Vocal Fold Synkinesis Presenting with Airway Obstruction

2009 ◽  
Vol 118 (8) ◽  
pp. 587-591 ◽  
Author(s):  
Roya Azadarmaki ◽  
Natasha Mirza ◽  
Ahmed M. S. Soliman
Keyword(s):  
Botulinum Toxin ◽  
Airway Obstruction ◽  
Vocal Fold ◽  
Botulinum Toxin Injection ◽  
Case Series ◽  
Vocal Fold Paralysis ◽  
Unilateral Vocal Fold Paralysis ◽  
Nerve Section ◽  
Breathing Therapy ◽  
Vocal Fold Motion

Objectives We present a case series of 10 patients with unilateral true vocal fold paralysis who presented with airway obstruction. Methods A retrospective review of the authors' patients at 2 institutions with unilateral true vocal fold motion impairment was carried out over a 10-year period. Of these, 10 patients were identified who presented with stridor and dyspnea as a result of synkinesis. Six cases were a result of thyroidectomy, 1 case resulted from recurrent laryngeal nerve section for spasmodic dysphonia, 1 case occurred after anterior cervical diskectomy and fusion, and in 2 cases no cause was identified. Results Three patients underwent tracheotomy. Two patients underwent partial arytenoidectomy. Seven patients underwent botulinum toxin injection; 2 were treated with breathing therapy, and in 1 case breathing therapy was recommended. Seven patients underwent treatment with more than 1 method. Conclusions Unilateral vocal fold paralysis may present with airway obstruction as a result of synkinesis. Treatment should be incremental and starts with breathing therapy and botulinum toxin injection. Partial arytenoidectomy or tracheotomy may be necessary for refractory cases.

scholarly journals OVERLAPPING OF FUNCTIONAL ESOPHAGEAL DISORDERS AND IRRITABLE BOWEL SYNDROME, IN MUSICIANS AND ATHLETES

2018 ◽  
Vol 24 (1) ◽  
pp. 69-72
Author(s):  
Sebastian Nedelcut ◽  
Daniel-Corneliu Leucuta ◽  
Dan Lucian Dumitrascu
Keyword(s):  
Associated Factors ◽  
Functional Gastrointestinal Disorders ◽  
Case Series ◽  
Cross Sectional Study ◽  
Control Group ◽  
Level Of Evidence ◽  
Cross Sectional ◽  
Functional Heartburn ◽  
Irritable Bowel Disease ◽  
Irritable Bowel

ABSTRACT Introduction: Functional gastrointestinal disorders (FGIDs) are the most common disorders in the general population. These disorders can overlap, decreasing the quality of life. Objective: We analyzed the prevalence of functional esophageal disorders (FED) and irritable bowel disease (IBS), and their overlapping and associated factors in musicians and athletes. Methods: A cross-sectional study was conducted using FGID and associated factors questionnaires administered to four groups: instrumentalists, singers, athletes, and a control group of healthy volunteers. Results: Of the 161 subjects, 62 (38.51%) had only FED, 76 (47.2%) had only IBS, and 23 (14.29%) had FED-IBS overlap. Subjects with FED-IBS overlap had more severe symptoms of IBS, especially hard and lumpy stools and constipation, compared to those with IBS alone. IBS subtype was more frequent in the overlap group, while not specified IBS type was less frequent. Regarding FED, we found that subjects with FED-IBS overlap had more functional heartburn and less functional dysphagia symptoms. There was a higher risk of overlap in instrumentalists and smokers. Conclusions: FED and IBS are frequently encountered in musicians and athletes. Subjects with FED-IBS overlap presented more frequent and severe symptoms. Instrumentalists and smokers are at higher risk of overlap. Level of Evidence IV; Case series.

Association between Upper Respiratory Infection and Idiopathic Unilateral Vocal Fold Paralysis

2018 ◽  
Vol 127 (10) ◽  
pp. 667-671 ◽  
Author(s):  
Neel K. Bhatt ◽  
Patrik Pipkorn ◽  
Randal C. Paniello
Keyword(s):  
Respiratory Infection ◽  
Vocal Fold ◽  
Case Series ◽  
Vocal Fold Paralysis ◽  
Unilateral Vocal Fold Paralysis ◽  
Viral Etiology ◽  
Upper Respiratory Infection ◽  
Future Studies ◽  
Presenting Symptoms ◽  
Upper Respiratory

Introduction: Unilateral vocal fold paralysis (UVFP) without an identifiable cause is termed idiopathic unilateral vocal fold paralysis (IUVFP). Some authors have postulated that select cases of IUVFP have a viral etiology, but the causality has not been established. We set out to review institutional cases of IUVFP and determine if there is a correlation between upper respiratory infection symptoms and presentation of IUVFP. Methods: Cases of IUVFP were reviewed over a 10-year period (2002-2012). The history was investigated to review presenting symptoms. We specifically reviewed for symptoms of upper respiratory infection at the onset of UVFP and tallied the frequency. Symptoms included sore throat, laryngitis, cough, influenza, bronchitis, pneumonia, otalgia, and sinusitis. The seasonal onset (if possible) was determined based on the history provided from the initial consultation. Study Design: Case series. Results: Overall, 107 patients presented with IUVFP; 35.5% of patients reported symptoms of upper respiratory infection at the onset of UVFP. Among these individuals, pharyngitis/laryngitis was the most common presenting symptom; 34.2% reported cough. In total, 40.0% of patients with IUVFP reported an onset of symptoms between December and February. Conclusions: This study suggests that symptoms of upper respiratory infection frequently occur with the presentation of IUVFP. The onset of symptoms tended to occur between December and February. The mechanism of viral-mediated UVFP has not been established. Future studies to explore this pathophysiology are needed.

C02 Laser-assisted Phonosurgery for Benign Glottic Lesions

2008 ◽  
Vol 139 (2_suppl) ◽  
pp. P48-P48
Author(s):  
Marcel Geyer ◽  
Gian-Peppino Ledda ◽  
Neil C Tan ◽  
Roberto Puxeddu Consultant
Keyword(s):  
Co2 Laser ◽  
Vocal Fold ◽  
Voice Quality ◽  
Case Series ◽  
Functional Results ◽  
Disease Recurrence ◽  
Wilcoxon Test ◽  
Control Group ◽  
Retrospective Case ◽  
False Vocal

Objective (1) To determine glottic function following carbon dioxide (CO2) laser-assisted phonosurgery of benign laryngeal disease. (2) To evaluate post-operative glottic morphology and disease recurrence rate. Methods Comparative retrospective case series of patients with benign glottic pathology treated by laser-assisted phonosurgery over 10 years. 235 consecutive patients had pre- and postoperative data collected from objective laryngeal examination, videostroboscopy recording of vocal fold mucosal wave movement, electroacoustic voice analysis of fundamental frequency, jitter, shimmer, and harmonics to noise ratio, as well as perceptual voice evaluation. This data was also compared to that of 20 healthy volunteers. (Statistical analysis: Wilcoxon test and Mann-Whitney test respectively). Definitive voice and morphologic evaluation was completed after 6 months. Results Evaluation of the pre- and postoperative functional results demonstrated a statistically significant improvement in all spectrographically analysed objective voice parameters (p<0.001), with a restored voice quality as good as the control group (p<0.001). Postoperative morphological analysis using videostroboscopic examination confirmed 3 recurrences of granuloma and 1 of Reinke's oedema. Recurrence was estimated objectively if the lesion was analogous to the original pathology. Glottic closure was complete in 96.5% of cases. False vocal fold adduction was normal in 88.5% of cases, with forced hyper-adduction being present in 11.5%. Phonatory vibration was cord-cord type in 100% of cases. Conclusions Our study demonstrates a statistically significant improvement in all acoustic parameters recorded. Postoperative vocal fold function and mucosal wave morphology were largely restored. CO2 laser-assisted voice restoration for benign glottic disease is effective.

Sign in / Sign up

Export Citation Format

Share Document