Lightweight Deep Learning Model for Assessment of Substitution Voicing and Speech after Laryngeal Carcinoma Surgery

Maskeliūnas, Rytis; Kulikajevas, Audrius; Damaševičius, Robertas; Pribuišis, Kipras; Ulozaitė-Stanienė, Nora; Uloza, Virgilijus

doi:10.3390/cancers14102366

Cited by 13 publications

(8 citation statements)

References 77 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…AVQI was used to assess general vocal quality, and ABI was used to assess breathiness, both of which were found to be relevant. Other researchers wanted to use cutting-edge deep learning research to objectively categorize, extract, and assess substitute voicing following laryngeal oncosurgery from audio signals [35]. Their technique had the highest true-positive rate of all of the cutting-edge approaches examined, reaching an acceptable overall accuracy and demonstrating the practical usage of voice quality devices.…”

Section: State Of the Art Reviewmentioning

confidence: 99%

Pareto-Optimized AVQI Assessment of Dysphonia: A Clinical Trial Using Various Smartphones

et al. 2023

Self Cite

View full text Add to dashboard Cite

Multiparametric indices offer a more comprehensive approach to voice quality assessment by taking into account multiple acoustic parameters. Artificial intelligence technology can be utilized in healthcare to evaluate data and optimize decision-making processes. Mobile devices provide new opportunities for remote speech monitoring, allowing the use of basic mobile devices as screening tools for the early identification and treatment of voice disorders. However, it is necessary to demonstrate equivalence between mobile device signals and gold standard microphone preamplifiers. Despite the increased use and availability of technology, there is still a lack of understanding of the impact of physiological, speech/language, and cultural factors on voice assessment. Challenges to research include accounting for organic speech-related covariables, such as differences in conversing voice sound pressure level (SPL) and fundamental frequency (f0), recognizing the link between sensory and experimental acoustic outcomes, and obtaining a large dataset to understand regular variation between and within voice-disordered individuals. Our study investigated the use of cellphones to estimate the Acoustic Voice Quality Index (AVQI) in a typical clinical setting using a Pareto-optimized approach in the signal processing path. We found that there was a strong correlation between AVQI results obtained from different smartphones and a studio microphone, with no significant differences in mean AVQI scores between different smartphones. The diagnostic accuracy of different smartphones was comparable to that of a professional microphone, with optimal AVQI cut-off values that can effectively distinguish between normal and pathological voice for each smartphone used in the study. All devices met the proposed 0.8 AUC threshold and demonstrated an acceptable Youden index value.

show abstract

Section: State Of the Art Reviewmentioning

confidence: 99%

Pareto-Optimized AVQI Assessment of Dysphonia: A Clinical Trial Using Various Smartphones

et al. 2023

Self Cite

View full text Add to dashboard Cite

show abstract

“…Objective evaluation is more challenging in patients who have undergone extended cordectomy, or partial or total laryngectomy as they utilize substitution voicing, where voice is generated by a single oscillating vocal fold against remaining laryngeal and pharyngeal structures, alaryngeal (esophageal or tracheoesophageal) processes, vocal prostheses, or electrolarynx. Maskeliunas et al [56 ▪▪ ,57] used DL to objectively categorize, extract, and assess substitution voicing from speech audio signals. Mel-frequency spectrograms (visual representation of the frequency spectrum of an acoustic signal) were used as input to a deep neural network architecture with an overall accuracy of 89.5%.…”

Section: Machine Learning Applications To Assess Voice and Swallowing...mentioning

confidence: 99%

Machine learning in the evaluation of voice and swallowing in the head and neck cancer patient

Srinivasan,

Liu,

Rameau

2023

Current Opinion in Otolaryngology &Amp; Head &Amp; Neck Surgery

View full text Add to dashboard Cite

Purpose of review The purpose of this review is to present recent advances and limitations in machine learning applied to the evaluation of speech, voice, and swallowing in head and neck cancer. Recent findings Novel machine learning models incorporating diverse data modalities with improved discriminatory capabilities have been developed for predicting toxicities following head and neck cancer therapy, including dysphagia, dysphonia, xerostomia, and weight loss as well as guiding treatment planning. Machine learning has been applied to the care of posttreatment voice and swallowing dysfunction by offering objective and standardized assessments and aiding innovative technologies for functional restoration. Voice and speech are also being utilized in machine learning algorithms to screen laryngeal cancer. Summary Machine learning has the potential to help optimize, assess, predict, and rehabilitate voice and swallowing function in head and neck cancer patients as well as aid in cancer screening. However, existing studies are limited by the lack of sufficient external validation and generalizability, insufficient transparency and reproducibility, and no clear superior predictive modeling strategies. Algorithms and applications will need to be trained on large multiinstitutional data sets, incorporate sociodemographic data to reduce bias, and achieve validation through clinical trials for optimal performance and utility.

show abstract

“…Furthermore, morphological characteristics supplied by CT are insufficient to evaluate the biological characteristics of the primary tumor. At present, radiomics overcomes the insufficiency of the above traditional imaging techniques, widely used in clinical diagnosis, treatment, and prognosis ( 15 , 16 ). The qualitative and quantitative assessment of lesion characteristics and intratumoral spatial heterogeneity can contribute to improving the non-invasive preoperative diagnosis accuracy of HNSCC.…”

Section: Introductionmentioning

confidence: 99%

Clinical application of CT-based radiomics model in differentiation between laryngeal squamous cell carcinoma and squamous cell hyperplasia

Cui,

Khodrog,

Liu

et al. 2024

Front. Med.

View full text Add to dashboard Cite

ObjectiveTo evaluate the clinical application of the CT-based radiomics prediction model for discriminating SCC and SCH.MethodsA total of 254 clinical samples were selected from 291 patients with larynx-occupying lesions who underwent primary surgery. All lesions were validated via histopathological examination at The Second Hospital of Jilin University between June 2004 and December 2019. All patients were randomly allocated to the training (n = 177) and validation (n = 77) cohorts. After the acquisition of CT images, manual 3D tumor segmentation was performed using the CT images of the arterial, venous, and non-contrast phases via ITK-SNAP software. Subsequently, radiomics features were extracted using A.K. software. Based on the above features, three different diagnostic models (CTN, CTA+CTV, and CTN+CTA+CTV) were constructed to classify squamous cell carcinoma (SCC) and squamous cell hyperplasia (SCH). Additionally, receiver operating characteristic (ROC) and decision curve analysis (DCA) curves were measured to evaluate the diagnostic characteristics and clinical safety of the proposed three prognostic models.ResultsIn the radiomic prediction Model 1 (CTN), the area under the curve (AUC), accuracy, sensitivity, specificity, positive predictive value (PPV) and negative predictive value (NPV) of the training cohorts in differentiating SCC and SCH were 0.883, 0.785, 0.645, 1.000, 1.000, and 0.648, while in the testing cohorts, these values were 0.852, 0.792, 0.66, 1.000, 1.000, and 0.652, respectively. In the radiomic prediction Model 2 (CTA+CTV), the AUC, accuracy, sensitivity, specificity, PPV, and NPV values of the training cohorts were 0.965, 0.91, 0.916, 0.9, 0.933, and 0.875, respectively, while in the testing cohorts, the corresponding values were 0.902, 0.805, 0.851, 0.733, 0.833, and 0.759, respectively. In the radiomic prediction Model 3(CTN+CTA+CTV), the AUC, accuracy, sensitivity, specificity, PPV, and NPV values of the training cohorts were 0.985, 0.944, 0.953, 0.929, 0.953, and 0.929, while in the testing cohorts, the corresponding values were 0.965, 0.857, 0.894, 0.8, 0.875, and 0.828, respectively.ConclusionThe radiomic prediction Model 3, based on the arterial-venous-plain combined scan phase of CT, achieved promising diagnostic performance, expected to be regarded as a preoperative imaging tool in classifying SCC and SCH to guide clinicians to develop individualized treatment programs.

show abstract

Lightweight Deep Learning Model for Assessment of Substitution Voicing and Speech after Laryngeal Carcinoma Surgery

Cited by 13 publications

References 77 publications

Pareto-Optimized AVQI Assessment of Dysphonia: A Clinical Trial Using Various Smartphones

Pareto-Optimized AVQI Assessment of Dysphonia: A Clinical Trial Using Various Smartphones

Machine learning in the evaluation of voice and swallowing in the head and neck cancer patient

Clinical application of CT-based radiomics model in differentiation between laryngeal squamous cell carcinoma and squamous cell hyperplasia

Contact Info

Product

Resources

About