Classification of laryngeal diseases including laryngeal cancer, benign mucosal disease, and vocal cord paralysis by artificial intelligence using voice analysis.

Voice change is often the first sign of laryngeal cancer, leading to diagnosis through hospital laryngoscopy. Screening for laryngeal cancer solely based on voice could enhance early detection. However, identifying voice indicators specific to laryngeal cancer is challenging, especially when differentiating it from other laryngeal ailments. This study presents an artificial intelligence model designed to distinguish between healthy voices, laryngeal cancer voices, and those of the other laryngeal conditions. We gathered voice samples of individuals with laryngeal cancer, vocal cord paralysis, benign mucosal diseases, and healthy participants. Comprehensive testing was conducted to determine the best mel-frequency cepstral coefficient conversion and machine learning techniques, with results analyzed in-depth. In our tests, laryngeal diseases distinguishing from healthy voices achieved an accuracy of 0.85-0.97. However, when multiclass classification, accuracy ranged from 0.75 to 0.83. These findings highlight the challenges of artificial intelligence-driven voice-based diagnosis due to overlaps with benign conditions but also underscore its potential.

Optical Biopsy: Automated Classification of Airway Endoscopic Findings Using a Convolutional Neural Network.

OBJECTIVES/HYPOTHESIS: Create an autonomous computational system to classify endoscopy findings. STUDY DESIGN: Computational analysis of vocal fold images at an academic, tertiary-care laryngology practice. METHODS: A series of normal and abnormal vocal fold images were obtained from the image database of an academic tertiary care laryngology practice. The benign images included normals, nodules, papilloma, polyps, and webs. A separate set of carcinoma and leukoplakia images comprised a single malignant-premalignant class. All images were classified with their existing labels. Images were randomly withheld from each class for testing. The remaining images were used to train and validate a neural network for classifying vocal fold lesions. Two classifiers were developed. A multiclass system classified the five categories of benign lesions. A separate analysis was performed using a binary classifier trained to distinguish malignant-premalignant from benign lesions. RESULTS: Precision ranged from 71.7% (polyps) to 89.7% (papilloma), and recall ranged from 70.0% (papilloma) to 88.0% (nodules) for the benign classifier. Overall accuracy for the benign classifier was 80.8%. The binary classifier correctly identified 92.0% of the malignant-premalignant lesions with an overall accuracy of 93.0%. CONCLUSIONS: Autonomous classification of endoscopic images with artificial intelligence technology is possible. Better network implementations and larger datasets will continue to improve classifier accuracy. A clinically useful optical cancer screening system may require a multimodality approach that incorporates nonvisual spectra. LEVEL OF EVIDENCE: NA Laryngoscope, 132:S1-S8, 2022.

Advanced computing solutions for analysis of laryngeal disorders.

Clinical diagnosis of voice pathologies is performed by analyzing audio, color, shape, and vibration patterns of the laryngeal recordings which are taken with medical imaging devices such as video-laryngostroboscope, direct laryngoscopy, and high-speed videoendoscopes. This paper examines state-of-the-art methods and reveals open issues and problems of computing solutions for analysis and identification of laryngeal disorders. We propose a categorical representation of the most significant applications published so far in terms of their scopes, used methodologies, and achieved results. Laryngeal image/video analysis is discussed in four main categories: segmentation of vocal folds, classification of vocal fold disorders, vocal fold vibration analysis, and vocal fold image stitching. By this study, we reveal new opportunities and potentials of vision-based computerized solutions for evaluation, early diagnosis, and prevention of laryngeal disorders. Graphical abstract.

Vocal fold scars: a common classification proposal by the American Laryngological Association and European Laryngological Society.

PURPOSE: Vocal fold scar is one the most challenging benign laryngeal pathologies. The purpose of this paper is to propose a classification that will allow for a common description of this entity between laryngologists, prevent discrepancies in interpretation, allow for comparison of related studies, and offer a training tool for young laryngologists. METHODS/RESULTS: Based on the depth and laterality of scarring, we propose 4 types: type I, characterized by atrophy of lamina propria with/without affected epithelium; type II, where the epithelium, lamina propria, and muscle are affected; type III, where the scar is located on the anterior commissure; type IV, which includes extended scar formation in both anteroposterior and rostro-caudal axis with significant loss of vocal fold mass. CONCLUSION: We believe that our proposal is comprehensive and encompasses all existing iatrogenic and non-iatrogenic etiologies in a simple and concise manner. It also serves its purpose as a descriptive, comparative, and training tool.

A Grading System for Transnasal Flexible Laryngoscopy.

OBJECTIVE: Transnasal flexible laryngoscopy (TFL) is becoming more popular in laryngology clinical practice. There has not been any grading system for TFL to help the physician document and communicate the laryngeal view yet. In this study, we aimed to classify the laryngeal view based on the visualization of the glottic aperture with TFL performed on conscious patients. METHODS: The TFL videos of 920 randomized patients were evaluated by three blind observers experienced with laryngology. The laryngeal view, consisting of the basic anatomic landmarks of the glottis, arytenoids, and epiglottis, was examined, and the glottic aperture was classified with a five-point grading system. Interobserver agreements for the grading system scores were assessed by using the kappa (k) statistic. RESULTS: Nine hundred and twenty subjects were enrolled in the study. Six hundred and thirty-eight (69.3%) were men, and 282 (30.6%) were women, and the mean age was 40.13 ± 15.08 (18-89 years). The number of patients constituting grade 1 was 737 (80.1%), while grade 2a was 122 (13.2%), grade 2b was 32 (3.4%), grade 3 was 24 (2.6%), and finally, grade 4 was only 5 (0.5%). The k score was 0.945 (P < 0.001) between the ratings of observer 1 and observer 2, 0.933 (P < 0.001) between observer 1 and observer 3, and 0.91 (P < 0.001) between observer 2 and observer 3. CONCLUSION: This new grading system for the laryngeal view can help physicians assess the upper airways, and it can also help visualize how much of a glottic opening there is.

Visible Vessels of Vocal Folds: Can they have a Diagnostic Role?

BACKGROUND: Challenges in visual identification of laryngeal disorders lead researchers to investigate new opportunities to help clinical examination. This paper presents an efficient and simple method which extracts and assesses blood vessels on vocal fold tissue in order to serve medical diagnosis. METHODS: The proposed vessel segmentation approach has been designed in order to overcome difficulties raised by design specifications of videolaryngostroboscopy and anatomic structure of vocal fold vasculature. The limited number of medical studies on vocal fold vasculature point out that the direction of blood vessels and amount of vasculature are discriminative features for vocal fold disorders. Therefore, we extracted the features of vessels on the basis of these studies. We represent vessels as vascular vectors and suggest a vector field based measurement that quantifies the orientation pattern of blood vessels towards vocal fold pathologies. RESULTS: In order to demonstrate the relationship between vessel structure and vocal fold disorders, we performed classification of vocal fold disorders by using only vessel features. A binary tree of Support Vector Machine (SVM) has been exploited for classification. Average recall of proposed vessel extraction method was calculated as 0.82 while healthy, sulcus vocalis, laryngitis classification accuracy of 0.75 was achieved. CONCLUSION: Obtained success rates showed the efficiency of vocal fold vessels in serving as an indicator of laryngeal diseases.

[Research status and challenges of vocal cord leukoplakia].

Vocal cord leukoplakia is a clinical diagnosis defined as a whitish patch or a plaque on a mucosal surface. Because of the diversity of histopathological types, the complexity and unpredictability risks for malignant transformation, there are still many controversies about its histopathological classification, diagnosis and treatment. The aim of this article is to review the epidemiology, etiology, pathological classification, diagnosis, treatment and prognosis of vocal cord leukoplakia.

Prolonged Intubation Injuries of the Larynx: Endoscopic Diagnosis, Classification, And Treatment.

Laryngeal trauma from prolonged endotracheal intubation occurs in patients of all ages. Most changes are superficial and heal quickly. Injuries that are found consistently during intubation include nonspecific changes, edema, granulation tissue, ulceration, and othermiscellaneous injuries. In thispapersignificant, severe, and lasting trauma of the larynx has been classified on thebasis of theknown factors in pathogenesis, observations made atendoscopy, and photographic documentation. This classification has required introduction of new descriptive terminology: "tongues of granulation tissue," "ulcerated troughs," "healed furrows," and "healed fibrous nodule." During intubation the degree of injury can be precisely assessed under general anesthesia by using telescopes for image magnification, thus assisting adecision whether to continue intubation orperform tracheotomy to minimize long-term morbidity. Changes that are found after extubation result from granulation tissue, ulceration, ora combination of both and have been illustrated on flow charts; a knowledge and understanding of these sequelae allows them to be identified by both indirect and direct laryngoscopy so that treatment can be planned.

A new classification of vocal fold leukoplakia by morphological appearance guiding the treatment.

OBJECTIVE: To provide a reference for classification and treatment of vocal cords leukoplakia. METHODS: 640 cases of patients with vocal cords leukoplakia were divided into three groups based on the appearances. There were respectively 81 smooth flat lesions, 155 smooth hypertrophy lesions and 13 rough lesions were treated with conservative methods and 26, 153 and 212 cases were resected surgically for biopsy. RESULTS: A majority of smooth flat leukoplakia lesions were cured by non-operational methods and no atypical hyperplasia. Most of rough lesions were invalid and even progressed treated with conservative therapy and were severe dysplasia or canceration. Although 67.6% smooth hypertrophy lesions were cured or improved by conservative methods and 47% lesions had no or mild dysplasia, over 30% of lesions were still invalid or progressed and over 50% had moderate, severe dysplasia and canceration. By statistical analysis, the appearance of leukoplakia was concordant with pathological changes. CONCLUSIONS: It is suggested that smooth flat vocal cords leukoplakia could be treated with conservative methods and rough lesions should be resected operationally, while for smooth hypertrophy vocal cords leukoplakia could be first followed-up and then adopt measures based on the appearance changes.

Evaluation of vascular features of vocal cords proposed by the European Laryngological Society.

A newly proposed classification by the European Laryngological Society (ELS) of glottic lesions by narrow-band imaging (NBI) divides their vascular patterns into longitudinal and perpendicular ones. The latter are further subdivided into the wide and narrow patterns. The longitudinal, wide, and narrow patterns are characteristic of benign disease, papilloma, and malignancy, respectively. The aim of the study was to investigate the diagnostic effectiveness of the classification. Forty patients with glottic lesions underwent microlaryngoscopy. The vascular patterns of all vocal cords were defined with NBI. The affected vocal cords were histologically analysed and comprised the arm (A). Unaffected vocal cords were not histologically analysed but followed-up and comprised the arm (B) and were regarded as true negatives if no suspicious changes appeared during the follow-up. The vocal cords from the arm A were categorised into the benign and malignant group according to the histologic result. The ratio of vascular patterns was determined and the groups were statistically compared using the Chi-square test and Fisher's exact test. Perpendicular changes were observed in 36.6% (9/26) of benign diseases and in 100% (23/23) of cancer conditions (p < 0.001). Wide perpendicular changes appeared only in papillomas (6/6) while narrow ones mostly in malignancies (23/26) and also in benign conditions (3/26) (p < 0.001). The sensitivity, specificity, positive and negative predictive values, and accuracy were 100, 95, 88, 100 and 96%, respectively. The new ELS classification can be used effectively and safely to differentiate malignant from benign disease.

Immunolocalization of CD163+ Tumor-Associated Macrophages and Symmetric Proliferation of Ki-67 as Biomarkers to Differentiate New Different Grades of Laryngeal Dysplasia.

OBJECTIVES: To evaluate CD163+ tumor-associated macrophages (TAMs), Ki-67, and cyclin D1 to differentiate laryngeal dysplasia in the 2017 World Health Organization classification. METHODS: Immunohistochemistry for CD163, Ki-67, and cyclin D1 was performed using paraffin-embedded specimens. CD163+ TAMs infiltrating the epithelium were estimated. Ki-67 and cyclin D1 were evaluated in four parts of the epithelium-basal, parabasal, middle third, and upper third layers. RESULTS: In total, 133 specimens were analyzed, including low-grade dysplasia (n = 31), high-grade dysplasia (n = 49), carcinoma in situ (n = 23), and normal mucosa (n = 30). CD163+ TAMs infiltrating the epithelium were significantly higher in high-grade dysplasia than in low-grade dysplasia. In the basal layer, Ki-67+ and cyclin D1+ cells were overexpressed in high-grade dysplasia (P < .0001). The area under the curve was 0.958 for Ki-67 and 0.909 for CD163+ TAMs (P < .0001). CONCLUSIONS: CD163+ TAMs infiltrating the epithelium and Ki-67 overexpression in the basal layer may serve as biomarkers to differentiate low-grade dysplasia from high-grade dysplasia of the larynx. A symmetric proliferative pattern was observed during laryngeal carcinogenesis following Ki-67 overexpression.

Laryngeal Cysts in Adults: Simplifying Classification and Management.

Objective Laryngeal cysts may occur at any mucosa-lined location within the larynx and account for 5% to 10% of nonmalignant laryngeal lesions. A number of proposed classifications for laryngeal cysts exist; however, no previously published classification aims to guide management. This review analyzes contemporary laryngeal cyst management and proposes a framework for the terminology and management of cystic lesions in the larynx. Data Sources PubMed/Medline. Review Methods A primary literature search of the entire Medline database was performed for all titles of publications pertaining to laryngeal cysts and reviewed for relevance. Full manuscripts were reviewed per the relevance of their titles and abstracts, and selection into this review was according to their clinical and scientific relevance. Conclusion Laryngeal cysts have been associated with rapid-onset epiglottitis, dyspnea, stridor, and death; therefore, they should not be considered of little significance. Symptoms are varied and nonspecific. Laryngoscopy is the primary initial diagnostic tool. Cross-sectional imaging may be required, and future use of endolaryngeal ultrasound and optical coherence tomography may revolutionize practice. Where possible, cysts should be completely excised, and there is growing evidence that a transoral approach is superior to transcervical excision for nearly all cysts. Histology provides definitive diagnosis, and oncocytic cysts require close follow-up. Implications for Practice A new classification system is proposed that increases clarity in terminology, with the aim of better preparing surgeons and authors for future advances in the understanding and management of laryngeal cysts.

Episodic Laryngeal Breathing Disorders: Literature Review and Proposal of Preliminary Theoretical Framework.

OBJECTIVE: The purposes of this literature review were (1) to identify and assess frameworks for clinical characterization of episodic laryngeal breathing disorders (ELBD) and their subtypes, (2) to integrate concepts from these frameworks into a novel theoretical paradigm, and (3) to provide a preliminary algorithm to classify clinical features of ELBD for future study of its clinical manifestations and underlying pathophysiological mechanisms. STUDY DESIGN: This is a literature review. METHODS: Peer-reviewed literature from 1983 to 2015 pertaining to models for ELBD was searched using Pubmed, Ovid, Proquest, Cochrane Database of Systematic Reviews, and Google Scholar. Theoretical models for ELBD were identified, evaluated, and integrated into a novel comprehensive framework. Consensus across three salient models provided a working definition and inclusionary criteria for ELBD within the new framework. Inconsistencies and discrepancies within the models provided an analytic platform for future research. RESULTS: Comparison among three conceptual models-(1) Irritable larynx syndrome, (2) Dichotomous triggers, and (3) Periodic occurrence of laryngeal obstruction-showed that the models uniformly consider ELBD to involve episodic laryngeal obstruction causing dyspnea. The models differed in their description of source of dyspnea, in their inclusion of corollary behaviors, in their inclusion of other laryngeal-based behaviors (eg, cough), and types of triggers. CONCLUSION: The proposed integrated theoretical framework for ELBD provides a preliminary systematic platform for the identification of key clinical feature patterns indicative of ELBD and associated clinical subgroups. This algorithmic paradigm should evolve with better understanding of this spectrum of disorders and its underlying pathophysiological mechanisms.

Accuracy of Acoustic Analysis Measurements in the Evaluation of Patients With Different Laryngeal Diagnoses.

OBJECTIVE: This study aims to investigate the accuracy of acoustic measures in discriminating between patients with different laryngeal diagnoses. STUDY DESIGN: The study design is descriptive, cross-sectional, and retrospective. METHODS: A total of 279 female patients participated in the research. Acoustic measures of the mean and standard deviation (SD) values of the fundamental frequency (F0), jitter, shimmer, and glottal to noise excitation (GNE) were extracted from the emission of the vowel /ε/. RESULTS: Isolated acoustic measures do not demonstrate adequate performance in discriminating patients with and without laryngeal alteration. The combination of GNE, SD of the F0, jitter, and shimmer improved the ability to classify patients with and without laryngeal alteration. In isolation, the SD of the F0, shimmer, and GNE presented acceptable performance in discriminating individuals with different laryngeal diagnoses. The combination of acoustic measurements caused discrete improvement in performance of the classifier to discriminate healthy larynx vs vocal polyp (SD of the F0, shimmer, and GNE), healthy larynx vs unilateral vocal fold paralysis (SD of the F0 and jitter), healthy larynx vs vocal nodules (SD of the F0 and jitter), healthy larynx vs sulcus vocalis (SD of the F0 and shimmer), and healthy larynx vs voice disorder due to gastroesophageal reflux (F0 mean, jitter, and shimmer). CONCLUSIONS: Isolated acoustic measures do not demonstrate adequate performance in discriminating patients with and without laryngeal alteration, although they present acceptable performance in classifying different laryngeal diagnoses. Combined acoustic measures present an acceptable capacity to discriminate between the presence and the absence of laryngeal alteration and to differentiate several laryngeal diagnoses.

Contralateral Vocal Fold Reactive Lesions: Nomenclature, Treatment Choice, and Outcome.

OBJECTIVE: Contralateral reactive lesions (RLs) represent a distinct entity among benign bilateral vocal fold (VF) lesions. Lack of uniform nomenclature and a myriad of surgical options have hampered attempts to develop treatment guidelines. The objective of this study is to better define RLs and their prognosis, through the development of a standard nomenclature, with an aim to guide treatment and delineate the role of phonosurgery. STUDY DESIGN: Case series with chart review. SETTING: Tertiary care center. SUBJECTS AND METHODS: Analysis was performed on patients with Current Procedural Terminology code 31545. Operative reports with a primary lesion and contralateral RL were included. Outcomes included the Voice Handicap Index-10 (VHI-10) and GRBAS (grade, roughness, breathiness, asthenia, and strain) scale, lesion persistence/recurrence, mucosal wave, and edge character based on blinded videostroboscopy review. RESULTS: A nomenclature was developed based on intraoperative RLs (n = 30), defined by lesion consistency (fibrous or polypoid) and relationship to normal VF edge (gradual or steep). Reactive lesion treatment included no intervention, excision, potassium titanyl phosphate laser, steroid injection, or a combination thereof. Observations included the following: inconsistent treatment modalities were employed, excision of RLs did not yield better outcomes, fibrous RLs were more likely to persist and polypoid lesions more likely to recur, gradual lesions were more likely to remain disease free, and most treatments showed improved mucosal wave, VHI-10, and GRBAS. CONCLUSIONS: Reactive lesions have not been well classified, and treatments are based on subjective intraoperative decision making with unpredictable outcomes. The nomenclature proposed will allow for a better definition of the RL and provide a framework for future research to identify optimal treatment.

Classification of vocal fold leukoplakia by clinical scoring.

BACKGROUND: Vocal cord leukoplakia comprises a variety of lesions. The purpose of this study was to stratify vocal leukoplakias before surgery. METHODS: Patients with an initial diagnosis of vocal leukoplakia who underwent surgical excision at a tertiary referral center in Taiwan were recruited for this study. Their clinical records, including age, sex, preoperative laryngoscopic images in the office setting, and final pathology reports were collected and analyzed. RESULTS: Patient age (p = .010), nonhomogenous lesion texture (p = .001), and existence of hyperemia (p = .014) were identified as independent factors predicting malignancy. A predictive formula was established accordingly. The model showed an excellent discrimination role by receiver operating characteristic curve analysis (area under the curve = 0.86; p < .001). CONCLUSION: This study confirmed the value of a scoring system based on laryngoscopic characteristics and patient age for predicting the histologic results in vocal leukoplakia. It is helpful for classifying vocal leukoplakia and pretreatment planning. © 2016 Wiley Periodicals, Inc. Head Neck 38: E1998-E2003, 2016.

Characterization Methods for the Detection of Multiple Voice Disorders: Neurological, Functional, and Laryngeal Diseases.

This paper evaluates the accuracy of different characterization methods for the automatic detection of multiple speech disorders. The speech impairments considered include dysphonia in people with Parkinson's disease (PD), dysphonia diagnosed in patients with different laryngeal pathologies (LP), and hypernasality in children with cleft lip and palate (CLP). Four different methods are applied to analyze the voice signals including noise content measures, spectral-cepstral modeling, nonlinear features, and measurements to quantify the stability of the fundamental frequency. These measures are tested in six databases: three with recordings of PD patients, two with patients with LP, and one with children with CLP. The abnormal vibration of the vocal folds observed in PD patients and in people with LP is modeled using the stability measures with accuracies ranging from 81% to 99% depending on the pathology. The spectral-cepstral features are used in this paper to model the voice spectrum with special emphasis around the first two formants. These measures exhibit accuracies ranging from 95% to 99% in the automatic detection of hypernasal voices, which confirms the presence of changes in the speech spectrum due to hypernasality. Noise measures suitably discriminate between dysphonic and healthy voices in both databases with speakers suffering from LP. The results obtained in this study suggest that it is not suitable to use every kind of features to model all of the voice pathologies; conversely, it is necessary to study the physiology of each impairment to choose the most appropriate set of features.

Office-based botulinum toxin injections.

Botulinum toxin injections are an effective treatment option for several laryngeal disorders. This article reviews the indications, procedural techniques, potential complications, and outcomes of Botox injections for laryngeal disorders.