RESUMO
OBJECTIVES: Vocal process granuloma (VPG) is a chronic condition resulting from a mucoperichondrial injury of vocal process. Initial conservative treatment typically involves vocal hygiene education and antireflux medication. Treatment challenges arise with refractory cases. Outcomes of second-line treatments such as surgical excision and botulinum toxin injections remain inconsistent. Thus, we propose this study to investigate the effectiveness of intralesional steroid injections for refractory VPG. METHODS: We conducted a retrospective review of 23 patients with VPG who showed no improvement after 3 months of proton pump inhibitors. These patients underwent one to three courses of monthly in-office intralesional steroid injections as a second-line therapy. Treatment outcomes were evaluated by measuring the size of the VPG relative to the length of the vocal folds before and after the final injection procedure. RESULTS: Results showed a significant reduction in VPG size from baseline of 27.74 ± 15.06 to 5.48 ± 8.95 (p < .001). 15 out of 23 patients were responsive (size reduction ≥ 75%) to intralesional steroid injection. Alcohol consumption and longer symptom duration were associated with a poor response (size reduction <75%), whereas prior intubation was associated with better response. CONCLUSIONS: For refractory VPG not responding to conservative treatment, intralesional steroid injection appears to be a promising alternative option without significant adverse effects.
RESUMO
This retrospective observational cohort study aims to assess the outcomes and associated factors in head and neck cancer (HNC) survivors with dysphagia, and to investigate the relationship between outcomes and speech and swallowing rehabilitation (SSR). We enrolled patients who were diagnosed with HNC between October 2016 and July 2018; we included 393 patients who developed dysphagia after definite treatment and were referred to speech-language pathologists (SLPs). We then classified patients into groups according to whether they received SSR. We used the clinical variables-including age, sex, site of malignancy, cancer stage, treatment modality, SSR, initial ECOG score, initial KPS, initial body weight (BW), and initial BMI-to evaluate the association between the percentage of BW change and overall survival (OS). There were 152 (39%) and 241 (61%) patients who received and did not receive SSR, respectively. In multivariate linear regression, SSR was significantly associated with percentage change in BW at 3 months post-treatment. Having SSR was positively associated with the percentage change in BW and decreased the BW loss [ß coefficient (95% CIs) = 2.53 (0.92 to 4.14)] compared to having no SSR. In the multivariate Cox regression, SSR was an independent factor for OS. Compared to no SSR, the hazard ratio (95% CIs) for patients who received SSR was 0.48 (0.31 to 0.74). SSR helps to avoid BW loss and increases overall survival. HNC patients who develop dysphagia after treatment should be encouraged to participate in SSR.
Assuntos
Transtornos de Deglutição , Neoplasias de Cabeça e Pescoço , Humanos , Deglutição , Transtornos de Deglutição/terapia , Fala , Estudos Retrospectivos , Sobreviventes , Redução de PesoRESUMO
BACKGROUND/PURPOSE: Voice therapy is frequently recommended as the first-line treatment for benign voice disorders. This study investigated the clinical effectiveness of voice therapy and the prognostic factors of treatment outcomes. METHODS: We recruited 103 consecutive patients with voice disorders, namely vocal nodules, polyps, and muscle tension dysphonia (MTD), from September 2014 to July 2016. All the patients received voice therapy as the primary treatment. Treatment outcomes were evaluated using auditory perceptual evaluation, acoustic analysis, maximum phonation time, and 10-item voice handicap index (VHI-10). Clinical effectiveness of voice therapy was defined by either 1) a posttreatment VHI-10 score ≤ 10 points or 2) decline of VHI-10 ≥ 4 points. RESULTS: After voice therapy, VHI-10 and perceptual rating of voice quality improved significantly (p < 0.05) in the three disease categories. In patients with nodules, all the outcome parameters improved significantly (p < 0.05). Patients with good adherence to voice therapy (attending more than four sessions) had a significantly higher effectiveness than those with poor adherence (87% vs. 64%, p < 0.05). Patients with high occupational vocal demand also demonstrated a better effectiveness than those with routine vocal demand (90% vs. 70%, p < 0.05). Subsequent multivariate analyses revealed that adherence and vocal demand were independently and significantly correlated with clinical effectiveness (p = 0.03). CONCLUSION: Voice therapy is effective for patients with vocal nodules, polyps, and MTD. Adherence to voice therapy and occupational vocal demand are significant prognostic factors for treatment outcomes.
Assuntos
Distúrbios da Voz , Humanos , Prognóstico , Resultado do Tratamento , Distúrbios da Voz/terapiaRESUMO
Deep learning techniques such as convolutional neural networks (CNN) have been successfully applied to identify pathological voices. However, the major disadvantage of using these advanced models is the lack of interpretability in explaining the predicted outcomes. This drawback further introduces a bottleneck for promoting the classification or detection of voice-disorder systems, especially in this pandemic period. In this paper, we proposed using a series of learnable sinc functions to replace the very first layer of a commonly used CNN to develop an explainable SincNet system for classifying or detecting pathological voices. The applied sinc filters, a front-end signal processor in SincNet, are critical for constructing the meaningful layer and are directly used to extract the acoustic features for following networks to generate high-level voice information. We conducted our tests on three different Far Eastern Memorial Hospital voice datasets. From our evaluations, the proposed approach achieves the highest 7%-accuracy and 9%-sensitivity improvements from conventional methods and thus demonstrates superior performance in predicting input pathological waveforms of the SincNet system. More importantly, we intended to give possible explanations between the system output and the first-layer extracted speech features based on our evaluated results.
Assuntos
Distúrbios da Voz , Voz , Acústica , Humanos , Redes Neurais de Computação , Distúrbios da Voz/diagnósticoRESUMO
OBJECTIVES: In the clinical setting, a variety of inner ear test results are obtained from patients with unilateral Meniere's disease (MD). In this study, the authors use inner ear test results as parameters to illustrate the relationship between inner ear function and vertigo attack frequency. DESIGN: We retrospectively enrolled 50 unilateral MD patients. In addition to clinical symptoms, the results of pure-tone audiometry and caloric, acoustic cervical vestibular-evoked myogenic potential (cVEMP), galvanic cVEMP, vibratory ocular VEMP (oVEMP), and galvanic oVEMP tests were collected via chart review. The multiple linear regression method was used to examine which independent variables have a statistically significant influence on vertigo attacks. RESULTS: In affected ears, the abnormal rate of the caloric, acoustic cVEMP, galvanic cVEMP, vibratory oVEMP, and galvanic oVEMP tests was 74%, 76%, 48%, 34%, and 30%, respectively. According to the regression model, the abnormal galvanic cVEMP response and abnormal galvanic oVEMP response had significantly negative correlations with the frequency of vertigo attacks after logarithmic transformation. A predictive model for disease attack frequency using significant parameters and their regression coefficients was proposed: (Equation is included in full-text article.) CONCLUSIONS:: Using the proposed model with galvanic VEMP, clinicians could develop better strategies to manage vertigo attacks in patients with MD.
Assuntos
Orelha Interna , Doença de Meniere , Humanos , Doença de Meniere/diagnóstico , Estudos Retrospectivos , Vertigem/diagnóstico , Potenciais Evocados Miogênicos VestibularesRESUMO
BACKGROUND: Ultrasound-guided-fine-needle aspiration drainage (US-FNAD) and US-percutaneous ethanol injection (US-PEI) have been widely used in the management of benign neck cysts. However, the long-term results of US-FNAD and US-PEI are not well elucidated. METHODS: We retrospectively collated patients under neck US examinations from March 2007 to December 2017 and investigated the recurrence after US-FNAD and US-PEI. Univariate and multivariate Cox regression analyses were used to assess significant risk factors for recurrence after US-FNAD. RESULTS: A total of 1075 patients were recruited, and their age was 50 ± 15 (mean ± standard deviation) years. A total of 862 patients had thyroid cysts, 118 patients had thyroglossal duct cysts (TGDC), twenty patients had branchial cleft cysts, 64 patients had parotid sialocysts, and 11 patients had plunging ranulas. Majority of the patients (97%, 1037/1075) reported significant symptom improvement immediately. However, 38% of the patients had recurrence with a median 3-year follow-up period. In a multivariate Cox regression analysis with adjustment for age and gender, plunging ranula (hazard ratio [HR]: 2.44, 95% confidence interval [CI]: 1.19-4.99) and lateral dimension size ≥ 0.8 cm (HR: 1.32, 95% CI: 1.04-1.67) after US-FNAD were independent risk factors for recurrence. There were 15 male and 19 female patients who received US-PEI therapy after repeated US-FNAD, of whom 23 patients had thyroid cysts, 6 had plunging ranulas, 4 had TGDC, and one had a branchial cleft cyst. The overall success rate was 94% (32/34), with a median follow-up period of 1.6 years. Two recurrent symptomatic patients had plunging ranulas. Some patients stated mild pain (21%, 7/34) and swelling sensation (26%, 9/34) after the injection. No major complications, such as vocal fold paresis or airway compression, were found. CONCLUSION: US-FNAD is an effective tool in the management of benign neck cysts with a 38% recurrence rate. Plunging ranulas have the highest rate of recurrence after FNAD. US-PEI is effective for most recurrent neck cysts after repeated US-FNAD.
RESUMO
BACKGROUND: Studies have used questionnaires of dysphonic symptoms to screen voice disorders. This study investigated whether the differential presentation of demographic and symptomatic features can be applied to computerized classification. METHODS: We recruited 100 patients with glottic neoplasm, 508 with phonotraumatic lesions, and 153 with unilateral vocal palsy. Statistical analyses revealed significantly different distributions of demographic and symptomatic variables. Machine learning algorithms, including decision tree, linear discriminant analysis, K-nearest neighbors, support vector machine, and artificial neural network, were applied to classify voice disorders. RESULTS: The results showed that demographic features were more effective for detecting neoplastic and phonotraumatic lesions, whereas symptoms were useful for detecting vocal palsy. When combining demographic and symptomatic variables, the artificial neural network achieved the highest accuracy of 83 ± 1.58%, whereas the accuracy achieved by other algorithms ranged from 74 to 82.6%. Decision tree analyses revealed that sex, age, smoking status, sudden onset of dysphonia, and 10-item voice handicap index scores were significant characteristics for classification. CONCLUSION: This study demonstrated a significant difference in demographic and symptomatic features between glottic neoplasm, phonotraumatic lesions, and vocal palsy. These features may facilitate automatic classification of voice disorders through machine learning algorithms.
Assuntos
Redes Neurais de Computação , Aprendizado de Máquina Supervisionado , Distúrbios da Voz/classificação , Adulto , Fatores Etários , Idoso , Consumo de Bebidas Alcoólicas/epidemiologia , Algoritmos , Demografia , Feminino , Glote/lesões , Glote/fisiopatologia , Humanos , Neoplasias Laríngeas/complicações , Neoplasias Laríngeas/diagnóstico , Neoplasias Laríngeas/fisiopatologia , Masculino , Pessoa de Meia-Idade , Estudos Retrospectivos , Índice de Gravidade de Doença , Fatores Sexuais , Fumar/epidemiologia , Avaliação de Sintomas , Paralisia das Pregas Vocais/complicações , Paralisia das Pregas Vocais/diagnóstico , Paralisia das Pregas Vocais/fisiopatologia , Distúrbios da Voz/epidemiologia , Qualidade da Voz , Ferimentos e Lesões/diagnósticoRESUMO
Ultrasound investigations and correct identification of malignant thyroid nodules depend on the experience and qualifications of the investigators; thus, a model that provides better evaluation before needle aspiration is desired. Data from 687 patients with 726 thyroid nodules comprising 65 malignant nodules (61 papillary and 4 follicular carcinoma) and 661 benign nodules were used to construct a predictive model. Presence of micro-calcification, taller-than-wide shape, predominant solid echostructure, and irregular margins were shown to be good independent predictive parameters. A thyroid nodule was predicted as malignant with a score ≥3.3. Internal validation of this predictive tool by the bootstrapping method showed excellent overall model performance.
Assuntos
Adenocarcinoma Folicular/diagnóstico por imagem , Carcinoma Papilar/diagnóstico por imagem , Simulação por Computador , Software , Neoplasias da Glândula Tireoide/diagnóstico por imagem , Nódulo da Glândula Tireoide/diagnóstico por imagem , Ultrassonografia Doppler em Cores , Adenocarcinoma Folicular/patologia , Adulto , Idoso , Idoso de 80 Anos ou mais , Biópsia por Agulha Fina , Carcinoma Papilar/patologia , Feminino , Humanos , Metástase Linfática/patologia , Masculino , Pessoa de Meia-Idade , Valor Preditivo dos Testes , Sistemas de Informação em Radiologia , Projetos de Pesquisa , Estudos Retrospectivos , Glândula Tireoide/diagnóstico por imagem , Glândula Tireoide/patologia , Neoplasias da Glândula Tireoide/patologia , Nódulo da Glândula Tireoide/patologia , Ultrassonografia de Intervenção , Adulto JovemRESUMO
OBJECTIVE: This study investigated whether artificial intelligence (AI) models combining voice signals, demographics, and structured medical records can detect glottic neoplasm from benign voice disorders. METHODS: We used a primary dataset containing 2-3 s of vowel "ah", demographics, and 26 items of structured medical records (e.g., symptoms, comorbidity, smoking and alcohol consumption, vocal demand) from 60 patients with pathology-proved glottic neoplasm (i.e., squamous cell carcinoma, carcinoma in situ, and dysplasia) and 1940 patients with benign voice disorders. The validation dataset comprised data from 23 patients with glottic neoplasm and 1331 patients with benign disorders. The AI model combined convolutional neural networks, gated recurrent units, and attention layers. We used 10-fold cross-validation (training-validation-testing: 8-1-1) and preserved the percentage between neoplasm and benign disorders in each fold. RESULTS: Results from the AI model using voice signals reached an area under the ROC curve (AUC) value of 0.631, and additional demographics increased this to 0.807. The highest AUC of 0.878 was achieved when combining voice, demographics, and medical records (sensitivity: 0.783, specificity: 0.816, accuracy: 0.815). External validation yielded an AUC value of 0.785 (voice plus demographics; sensitivity: 0.739, specificity: 0.745, accuracy: 0.745). Subanalysis showed that AI had higher sensitivity but lower specificity than human assessment (p < 0.01). The accuracy of AI detection with additional medical records was comparable with human assessment (82% vs. 83%, p = 0.78). CONCLUSIONS: Voice signal alone was insufficient for AI differentiation between glottic neoplasm and benign voice disorders, but additional demographics and medical records notably improved AI performance and approximated the prediction accuracy of humans. LEVEL OF EVIDENCE: NA Laryngoscope, 2024.
RESUMO
OBJECTIVES: To determine the efficacy of real-time elastography (RTE), compared with our previously proposed prediction model, in the detection of malignancy in cervical lymph nodes (LNs). METHODS: One hundred and thirty-one patients underwent ultrasound-guided fine needle aspiration biopsy (ultrasound FNAB) after ultrasound and RTE evaluation. The formula of the RTE scoring system was a four-point visual scale, based on a previously determined model. The formula of the prediction model was: [Formula: see text]. An extended model was constructed with four previous predictors and elasticity scores, using a logistic regression model. RESULTS: Final histology revealed 77 benign and 54 malignant LNs. In the elasticity score system, sensitivity was 66.7 %, specificity was 57.1 %, the positive predictive value (PPV) was 52.2 % and the negative predictive value (NPV) was 71.0 %. In the prediction model system, sensitivity was 79.6 %, specificity was 92.2 %, the PPV was 87.8 % and the NPV was 86.6 %. When the extended and the original model were compared, the areas under the receiver operating characteristic curve (c-statistic) was 0.94 and 0.95, respectively (P > 0.05). CONCLUSIONS: Qualitative RTE offers no additional value over conventional ultrasound in predicting malignancy in cervical LNs. KEY POINTS: ⢠An ultrasound system can help in the assessment of cervical lymph nodes. ⢠Grey-scale and power Doppler ultrasound remain fundamental for neck nodal evaluation. ⢠Qualitative real-time elastography provided no additional value compared with current prediction models.
Assuntos
Técnicas de Imagem por Elasticidade/métodos , Linfonodos/diagnóstico por imagem , Metástase Linfática/diagnóstico , Pescoço/diagnóstico por imagem , Neoplasias/diagnóstico por imagem , Ultrassonografia Doppler/métodos , Adulto , Idoso , Idoso de 80 Anos ou mais , Biópsia por Agulha Fina , Elasticidade , Feminino , Humanos , Masculino , Pessoa de Meia-Idade , Neoplasias/patologia , Variações Dependentes do Observador , Valor Preditivo dos Testes , Distribuição Aleatória , Análise de Regressão , Reprodutibilidade dos Testes , Sensibilidade e Especificidade , Adulto JovemRESUMO
OBJECTIVE: Fat injection laryngoplasty (FIL) is a common procedure used to correct glottic insufficiency. Nevertheless, few studies have discussed potential treatments for cases with poor voice recovery after FIL. METHODS: Eighteen patients with unfavorable vocal outcomes from FIL were analyzed. Each of these patients presented persistent dysphonia for more than 2 months following FIL, together with bulging vocal folds and poor mucosal wave. We applied microsurgery as the standard treatment to remove excessive fat. Vocal fold steroid injection (VFSI) was administered to patients that were hesitant or declined to undergo microsurgery. Voice outcomes were evaluated using the Voice Handicap Index-10 (VHI-10), grade-roughness-breathiness (GRB) scores, and smoothed cepstral peak prominence (CPPs). RESULTS: Six patients underwent microsurgery directly, 6 patients received only VFSIs as a salvage treatment, and the remaining 6 patients received 1 to 3 courses of VFSIs before the decision to undergo microsurgery. Pathology reports were available for 10 patients, and contained 5 instances of adipose tissues, 3 of fat necrosis, 1 of chronic inflammation, and 1 of fibrosis. Seventeen patients reported satisfactory or improved outcomes. We found remarkable improvements in VHI-10, GRB, and CPPs (all P < .05) after salvage treatments for FIL. Subgroup analyses showed comparable voice outcomes for patients undergoing direct microsurgery, VFSI alone, and VFSI followed by microsurgery (P > .05). CONCLUSIONS: This study demonstrated that fat overinjection and/or fibrotic change in the injected vocal folds may cause poor voice outcomes after FIL. Both microsurgery and VFSI could be applied as salvage treatments with good voice recovery profiles. LEVEL OF EVIDENCE: Level 4.
Assuntos
Disfonia , Laringoplastia , Voz , Humanos , Terapia de Salvação , Laringoplastia/métodos , Resultado do Tratamento , Prega Vocal/cirurgia , Prega Vocal/patologia , Disfonia/etiologia , Disfonia/cirurgia , Estudos RetrospectivosRESUMO
Objective: To analyze the risk factors for postoperative vocal fold fibrosis (PVF) in patients undergoing microlaryngeal surgery (MLS) for benign vocal fold lesions. Study Design: Retrospective study. Methods: We retrospectively included patients who had undergone MLS for vocal polyps, nodules, mucus retention cysts, fibrous mass, or Reinke's edema. Data on the patients' clinicodemographic characteristics and intraoperative findings were obtained by reviewing their clinical records. PVF was defined by the presence of an adynamic segment of membranous vocal folds or a marked reduction in mucosal wave amplitude on post-MLS (6 weeks) videolaryngostroboscopy. The risk factors for PVF were analyzed through univariate and multivariate logistic regressions. Results: This study included 89 patients, of whom 16 (18%) were given a diagnosis of PVF. A significantly increased incidence of PVF was noted in patients with fibrous mass (p < .01). The univariate analysis indicated that lesion attachment to the vocal ligament, prolonged surgical duration (>60 min), and symptom duration (>12 months) were significantly correlated with PVF (p < .05). The multivariate analysis confirmed that diagnosis of fibrous masses, lesion attachment to the vocal ligament and symptom duration are significant risk factors for PVF. Conclusion: PVF is more common in patients with fibrous masses. Lesions attachment to the vocal ligament and prolonged symptom duration appear to be other significant risk factors for PVF. Level of Evidence: 4.
RESUMO
Purpose: The aim of this pilot study was to explore the efficacy of an intensive 3 week water resistance phonation (WRP) therapy program for people with presbyphonia.Method: Participants included 13 people with presbyphonia who received intensive WRP therapy. All participants completed eight sessions of therapy over 3 weeks. Auditory perceptual ratings, and acoustic and aerodynamic assessments were performed before and after treatment. Participants also completed the voice-related quality of life questionnaire before and after the treatment.Result: After 3 weeks of intensive voice therapy, significant improvements were demonstrated in acoustic, aerodynamic, and auditory perceptual parameters, as well as patient perceptions of voice-related quality of life. Acoustically, it was found there were significant decreases in shimmer (p = 0.019), noise-to-harmonic ratio (p = 0.016), and smoothed cepstral peak prominence (p = 0.001). Perceptually, the clients with presbyphonia showed significant reductions in the ratings of the overall grade, roughness, asthenia, and strain. Moreover, there was a significant increase in the total score of the Mandarin version of the Voice-Related Quality of Life measure post-therapy.Conclusion: The investigation provides preliminary evidence that people with presbyphonia can improve their vocal function and voice-related quality of life through intensive WRP therapy within a short period of time.
RESUMO
OBJECTIVES: Current treatments for benign vocal lesions consist mainly voice therapy and phonomicrosurgery. However, these options are not always suitable for professional voice users because of their tight performance schedule and limited time for voice rest. This study investigated vocal fold steroid injection (VFSI) as an alternative treatment. STUDY DESIGN: Matched case series. METHODS: We retrospectively enrolled 28 professional voice users (i.e., singers, actors and news anchors) who received VFSI for vocal nodules and polyps in an office setting of a tertiary teaching hospital. Outcomes were evaluated using videolaryngostroboscopy (VLS), the 10-item Voice Handicap Index (VHI-10), maximum phonation time, and acoustic and perceptual analyses before and 1 month after VFSI. Study results were compared with 56 patients of nonprofessional voice users (i.e., routine or high occupational vocal demands), matched in a 1:2 ratio by age, sex, and treatment date. RESULTS: After VFSI, VLS revealed substantial lesion resolution in 82% professional voice users. One professional voice user developed a self-limited vocal fold hematoma after VFSI. VHI-10 scores in the professional group decreased from 21 to 14 points, compared with 23 to 16 points in the non-professional group, demonstrating a significant within-group effect (P < 0.01, GEE) and a nonsignificant between-group effect (P = 0.86). Other outcomes also improved significantly after VFSI (P < 0.05), without significant differences between the two groups. CONCLUSION: VFSI can be an effective and safe alternative treatment for professional voice users with benign vocal lesions.
Assuntos
Doenças da Laringe , Distúrbios da Voz , Humanos , Prega Vocal/cirurgia , Estudos Retrospectivos , Resultado do Tratamento , Doenças da Laringe/diagnóstico , Doenças da Laringe/tratamento farmacológico , Doenças da Laringe/patologia , Esteroides/efeitos adversos , Distúrbios da Voz/diagnóstico , Distúrbios da Voz/tratamento farmacológico , Distúrbios da Voz/patologiaRESUMO
Objectives: This study developed the Mandarin Chinese version of the Aging Voice Index (AVI), with preliminary validation of the scale for potential clinical applications. Study Design: Scale development. Methods: The experimental procedure involved: (1) cross-cultural adaptation of the original AVI into the Mandarin Chinese version (CAVI); (2) evaluation by expert panel; (3) back translation; (4) pilot testing; (5) development of the final CAVI; (6) scale validation with 68 older adults of 60-89 years old (29 females and 39 males), 34 with voice disorders and 34 age-matched with normal voice. Internal consistency reliability, test-retest reliability, content validity, criterion-related validity, and discriminatory ability (diagnostic accuracy) of the CAVI were evaluated. Results: There were high internal consistency (Cronbach's alpha = 0.9733), high test-retest reliability (intraclass correlation coefficient = 0.9578, p < 0.01), high content validity (content validity index = 0.9710), high criterion-related validity (Pearson's r = 0.9439, p < 0.01 between CAVI and Voice Handicap Index-10; r = 0.8070, p < 0.01 between CAVI and voice-related quality of life [V-RQOL]), and significant difference in CAVI scores between the two groups with huge effect size (t(34.69) = -11.59, Cohen's d = 2.81, p < 0.001). Receiver operating characteristic analysis revealed a high diagnostic accuracy of the CAVI, with an area under the curve of 0.9974 (p < 0.001) and a cut-off score of 12.0 with 100% sensitivity and 97.1% specificity. Conclusion: Our findings suggested that the CAVI could be a reliable and valid standardized self-assessment questionnaire tool for clinical evaluation of the impact of voice problems specifically for Mandarin-speaking older adults. Further studies should explore a full-scale validation of the CAVI for being a standard clinical tool, including for older adults in Mainland China. Level of evidence: 3b (case-control study).
RESUMO
OBJECTIVE: Voice disorders significantly compromise individuals' ability to speak in their daily lives. Without early diagnosis and treatment, these disorders may deteriorate drastically. Thus, automatic classification systems at home are desirable for people who are inaccessible to clinical disease assessments. However, the performance of such systems may be weakened due to the constrained resources and domain mismatch between the clinical data and noisy real-world data. METHODS: This study develops a compact and domain-robust voice disorder classification system to identify the utterances of health, neoplasm, and benign structural diseases. Our proposed system utilizes a feature extractor model composed of factorized convolutional neural networks and subsequently deploys domain adversarial training to reconcile the domain mismatch by extracting domain-invariant features. RESULTS: The results show that the unweighted average recall in the noisy real-world domain improved by 13% and remained at 80% in the clinic domain with only slight degradation. The domain mismatch was effectively eliminated. Moreover, the proposed system reduced the usage of both memory and computation by over 73.9%. CONCLUSION: By deploying factorized convolutional neural networks and domain adversarial training, domain-invariant features can be derived for voice disorder classification with limited resources. The promising results confirm that the proposed system can significantly reduce resource consumption and improve classification accuracy by considering the domain mismatch. SIGNIFICANCE: To the best of our knowledge, this is the first study that jointly considers real-world model compression and noise-robustness issues in voice disorder classification. The proposed system is intended for application to embedded systems with limited resources.
Assuntos
Compressão de Dados , Distúrbios da Voz , Humanos , Distúrbios da Voz/diagnóstico , Redes Neurais de ComputaçãoRESUMO
OBJECTIVE: Doctors, nowadays, primarily use auditory-perceptual evaluation, such as the grade, roughness, breathiness, asthenia, and strain scale, to evaluate voice quality and determine the treatment. However, the results predicted by individual physicians often differ, because of subjective perceptions, and diagnosis time interval, if the patient's symptoms are hard to judge. Therefore, an accurate computerized pathological voice quality assessment system will improve the quality of assessment. METHOD: This study proposes a self_attention-based system, with a deep learning technology, named self_attention-based bidirectional long-short term memory (SA BiLSTM). Different pitches [low, normal, high], and vowels [/a/, /i/, /u/], were added into the proposed model, to make it learn how professional doctors evaluate the grade, roughness, breathiness, asthenia, and strain scale, in a high dimension view. RESULTS: The experimental results showed that the proposed system provided higher performance than the baseline system. More specifically, the macro average of the F1 score, presented as decimal, was used to compare the accuracy of classification. The (G, R, and B) of the proposed system were (0.768±0.011, 0.820±0.009, and 0.815±0.009), which is higher than the baseline systems: deep neural network (0.395±0.010, 0.312±0.019, 0.321±0.014) and convolution neural network (0.421±0.052, 0.306±0.043, 0.3250±0.032) respectively. CONCLUSIONS: The proposed system, with SA BiLSTM, pitches, and vowels, provides a more accurate way to evaluate the voice. This will be helpful for clinical voice evaluations and will improve patients' benefits from voice therapy.
RESUMO
BACKGROUND: How to properly manage clinically negative neck of head and neck cancer patients is a controversial topic. Research is now directed toward finding a method sensitive enough to bring the risk of occult metastases below 20%. The aim of this review was to compare the diagnostic accuracy of different imaging modalities, including CT, MRI, PET and US, in clinically N0 head and neck cancer patients. METHODS: For this systematic review and meta-analysis, PubMed and the Cochrane Database were searched for relevant original articles published up to May 2011. Inclusion criteria were as follows: articles were reported in English; CT, MRI, PET or US were performed to identify cervical metastases in clinically N0 head and neck squamous cell carcinoma; and data were sufficient for the calculation of true-positive or false-negative values. A bivariate random effect model was used to obtain pooled sensitivity and specificity. The positive and negative test probability of neck metastasis was generated based on Bayesian theory and collected data for different pre-test possibilities. RESULTS: Of the 168 identified relevant articles, 7 studies fulfilled all inclusion criteria for CT, 6 studies for MRI, 11 studies for PET and 8 studies for US. There was no difference in sensitivity and specificity among these imaging modalities, except CT was superior to US in specificity. The pooled estimates for sensitivity were 52% (95% confidence interval [CI], 39% ~ 65%), 65% (34 ~ 87%) 66% (47 ~ 80%), and 66% (45 ~ 77%), on a per-neck basis for CT, MRI, PET and US, respectively. The pooled estimates for specificity were 93% (87% ~ 97%), 81% (64 ~ 91%), 87% (77 ~ 93%), and 78% (71 ~ 83%) for CT, MRI, PET and US, respectively. With pre-examination nodal metastasis probabilities set at 10%, 20% and 30%, the post-exam probabilities of positive nodal metastasis rates were 47%, 66% and 77% for CT; 27%, 46% and 59% for MRI; 36%, 56% and 69% for PET; and 25%, 42% and 56% for US, respectively. Negative nodal metastasis probabilities were 95%, 89% and 82% for CT; 95%, 90% and 84% for MRI; 96%, 91% and 86% for PET; and 95%, 90% and 84% for US, respectively. CONCLUSIONS: Modern imaging modalities offer similar diagnostic accuracy to define and diagnose clinically N0 neck. Minimizing morbidity and avoiding elective neck dissection is acceptable in some select cases.
Assuntos
Diagnóstico por Imagem/métodos , Neoplasias de Cabeça e Pescoço/diagnóstico , Linfonodos/patologia , Humanos , Metástase Linfática , Reprodutibilidade dos Testes , Sensibilidade e EspecificidadeRESUMO
Background: Unilateral vocal fold paralysis (UVFP) after thyroid surgery often leads to significant morbidity including dysphonia, dysphagia, and aspiration. Injection laryngoplasty (IL) is an effective treatment of UVFP with numerous readily available materials. However, few studies focus on IL for UVFP following thyroidectomy. Objectives: This review aims to critically review current literature to determine the timing, materials, methods and outcomes of IL for UVFP after thyroid surgery. Type of Review: Literature review. Methods: A literature review was performed using the Pubmed, Medline and EMBASE databases. All relevant articles published in English addressing the effect of IL in post thyroid surgery related UVFP were analyzed. Studies using IL for UVFP of multiple etiologies were excluded. Meta-analysis was conducted using fixed and random effect model. Results: Five original studies were identified, including 214 patients received IL for UVFP following thyroid surgery. Two studies injected autologous fat via direct suspension laryngoscope under general anesthesia, while the other 3 studies injected polyacrylamide, hyaluronic acid, and polymethyl methacrylate from cricothyroid membrane under local anesthesia. All 5 studies reported improved voice outcomes of IL for post-thyroidectomy UVFP. Meta-analysis showed MPT increased for 3.18â s (95% CI: 2.40-3.96, fix effect model) after IL. Another common acoustic parameter, jitter (%) also improved for 1.46 (95% CI: 0.73-2.19, random effects model) after IL for post-thyroidectomy UVFP. Conclusions: This review supported that IL can improve the voice outcome for post-thyroidectomy UVFP. Autologous fat remains a good augmentation material with a potential longer lasting effect. More research and long-term surveys are needed to document the safety and longevity of other synthetic materials.
RESUMO
OBJECTIVES: The present study aims to evaluate the reliability and validity of the Mandarin version of the Singing Voice Handicap Index-10 (MSVHI-10). METHODS: One hundred and fifteen singers, aged from 21 to 65, participated in this study. All the subjects completed the Mandarin version of the SVHI-10 and the Voice Handicap Index-10 (VHI-10). Other gathered information include sex, age, type of singing performance, and the accumulated years of singing. We also asked the participants to self-report the presence of singing voice complaint. Statistical analyses were performed to evaluate the validity, reliability, and the Receiver Operating Characteristic (ROC) curve of MSVHI-10. RESULTS: The results showed high internal consistency (Cronbach's α =0.96), with the scores of each question ranged from 0.57 to 0.90. Forty-three singers re-filled the MSVHI-10 with an interval of 12.7 ± 9.3 days (mean ± standard deviation), and the results showed a high test-retest reliability (r = 0.98, P < 0.01). The SVHI-10 scores were significantly correlated with the VHI-10 (r = 0.54-0.72, P < 0.01). ROC analysis revealed a high differentiating accuracy for distinguishing between the subjects with or without singing voice complaints, with an area under the curve (AUC) of 0.97. At the cut-off score of 21.5, the sensitivity and specificity were 85.7% and 95.9%, respectively. CONCLUSIONS: The Mandarin version of the SVHI-10 has been shown to be a reliable and valid instrument for evaluating the self-perception of singing voice complaints.