Convolutional Neural Network Classifies Pathological Voice Change in Laryngeal Cancer with High Accuracy
Open accessThe DOI matches its Crossref record; 42 of the 43 references it cites were found in a registry or an index.
Abstract
Voice changes may be the earliest signs in laryngeal cancer. We investigated whether automated voice signal analysis can be used to distinguish patients with laryngeal cancer from healthy subjects. We extracted features using the software package for speech analysis in phonetics (PRAAT) and calculated the Mel-frequency cepstral coefficients (MFCCs) from voice samples of a vowel sound of /a:/. The proposed method was tested with six algorithms: support vector machine (SVM), extreme gradient boosting (XGBoost), light gradient boosted machine (LGBM), artificial neural network (ANN), one-dimensional convolutional neural network (1D-CNN) and two-dimensional convolutional neural network (2D-CNN). Their performances were evaluated in terms of accuracy, sensitivity, and specificity. The result was compared with human performance. A total of four volunteers, two of whom were trained laryngologists, rated the same files. The 1D-CNN showed the highest accuracy of 85% and sensitivity and sensitivity and specificity levels of 78% and 93%. The two laryngologists achieved accuracy of 69.9% but sensitivity levels of 44%. Automated analysis of voice signals could differentiate subjects with laryngeal cancer from those of healthy subjects with higher diagnostic properties than those performed by the four volunteers.
Overview
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Integrity signals
Registry
Cites
Statistics
Figures







Verification
Claims the paper makes
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
The source backs part of the claim, not all of it — or backs it for a different population, condition or comparison.
In the citing paper
“most are associated with non-specific voice quality (VQ) modifications that remain difficult to attribute to a specific diagnosis through perceptual evaluation alone [10,12,13,14].”
From Diagnostic Accuracy of Artificial Intelligence in Laryngeal Disorders: An Integrative Review (2026) · cites this paper as [10]
Experimental — accuracy not yet measured
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
The self-check needs the paper’s full text.
Citations
Cites
Siegel, R. L., Miller, K. D., & Jemal, A. (2020) · CA: A Cancer Journal for Clinicians · doi:10.3322/caac.21590 · Cited by 20,444
Dermatologist-level classification of skin cancer with deep neural networks
Esteva, A., Kuprel, B., Novoa, R. A., Ko, J., Swetter, S. M., Blau, H. M., & Thrun, S. (2017) · Nature · doi:10.1038/nature21056 · Cited by 14,362 · cited 1×
“Significant new work has been reported recently using artificial intelligence techniques in the early detection of cancer, including skin and gastric cancer [36,37].”
BackgroundPraat, a system for doing phonetics by computer
Boersma, P. (2002) · UvA-DARE (University of Amsterdam) · Cited by 4,417
Applications of Support Vector Machine (SVM) Learning in Cancer Genomics
(2018) · Cancer Genomics & Proteomics · doi:10.21873/cgp.20063 · Cited by 1,528 · cited 2×
“In machine learning, the hyperparameter is a high-level configurator empirically chosen to control complexity, conservativeness, and overfitting of a model before training the networks [16].”
BackgroundMuda, L., Begam, M. H., & Elamvazuthi, I. (2010) · arXiv (Cornell University) · doi:10.48550/arxiv.1003.4083 · Cited by 803 · cited 2×
“The voice, a signal carrying infinite information, can be represented in a simpler form by introducing digital signal processing tools such as PRAAT or MFCCs, which improves optimization potential despite the small datasets [15].”
BackgroundMethodHirasawa, T., Aoyama, K., Tanimoto, T., Ishihara, S., Shichijo, S., Ozawa, T., Ohnishi, T., Fujishiro, M., Matsuo, K., Fujisaki, J., et al. (2018) · Gastric Cancer · doi:10.1007/s10120-018-0793-2 · Cited by 744 · cited 1×
“Significant new work has been reported recently using artificial intelligence techniques in the early detection of cancer, including skin and gastric cancer [36,37].”
BackgroundPhysiologic and acoustic differences between male and female voices
Titze, I. R. (1989) · The Journal of the Acoustical Society of America · doi:10.1121/1.397959 · Cited by 607 · cited 1×
“First, male and female voices are known to fall within different ranges of frequency [21].”
BackgroundSex Disparities in Cancer Mortality and Survival
Cook, M. B., McGlynn, K. A., Devesa, S. S., Freedman, N. D., & Anderson, W. F. (2011) · Cancer Epidemiology, Biomarkers & Prevention · doi:10.1158/1055-9965.epi-11-0246 · Cited by 515 · cited 1×
“Laryngeal cancer is known to show a skewed gender distribution [25,26] with an approximately four- to six-fold higher risk in males [27] and poor prognosis compared with females [28].”
BackgroundOn rectified linear units for speech processing
Zeiler, M. D., Ranzato, M., Monga, R., Mao, M., Yang, K., Le, Q. V., Nguyen, P., Senior, A., Vanhoucke, V., Dean, J., et al. (2013) · 2013 IEEE International Conference on Acoustics, Speech and Signal Processing · doi:10.1109/icassp.2013.6638312 · Cited by 420 · cited 2×
“Batch normalization and ReLU activation are used for fa ter and stable training [19].”
MethodNonparametric regression using deep neural networks with ReLU activation function
Schmidt-Hieber, J. (2020) · The Annals of Statistics · doi:10.1214/19-aos1875 · Cited by 292
Detection of Pathological Voice Using Cepstrum Vectors: A Deep Learning Approach
Fang, S.-H., Tsao, Y., Hsiao, M.-J., Chen, J.-Y., Lai, Y.-H., Lin, F.-C., & Wang, C.-T. (2019) · Journal of Voice · doi:10.1016/j.jvoice.2018.02.003 · Cited by 278 · cited 4×
“However, in past studies, the number of cancer patients was either too small [10,11] or often assessed as a single group together with other voice disorders.”
BackgroundMethodResultInfluentialGodino-Llorente, J. I., & Gomez-Vilda, P. (2004) · IEEE Transactions on Biomedical Engineering · doi:10.1109/tbme.2003.820386 · Cited by 264 · cited 1×
“However, in past studies, the number of cancer patients was either too small [10,11] or often assessed as a single group together with other voice disorders.”
BackgroundVoice Pathology Detection and Classification Using Convolutional Neural Network Model
Mohammed, M. A., Abdulkareem, K. H., Mostafa, S. A., Khanapi Abd Ghani, M., Maashi, M. S., Garcia-Zapirain, B., Oleagordia, I., Alhakami, H., & AL-Dhief, F. T. (2020) · Applied Sciences · doi:10.3390/app10113723 · Cited by 208 · cited 3×
“contributive bands in both detection and classification ranged between 1000 and 8000 Hz [7].”
BackgroundMethodInfluentialHao, X., Bao, Y., Guo, Y., Yu, M., Zhang, D., Risacher, S. L., Saykin, A. J., Yao, X., & Shen, L. (2020) · Medical Image Analysis · doi:10.1016/j.media.2019.101625 · Cited by 202
Deep learning in head & neck cancer outcome prediction
Diamant, A., Chatterjee, A., Vallières, M., Shenouda, G., & Seuntjens, J. (2019) · Scientific Reports · doi:10.1038/s41598-019-39206-1 · Cited by 200 · cited 1×
“Recent studies have shown that the CNN can be used to predict the outcome automatically [40] or detect laryngeal cancer based on laryngoscope images with accuracy levels of”
BackgroundFeature extraction and classification of heart sound using 1D convolutional neural networks
Li, F., Liu, M., Zhao, Y., Kong, L., Dong, L., Liu, X., & Hui, M. (2019) · EURASIP Journal on Advances in Signal Processing · doi:10.1186/s13634-019-0651-3 · Cited by 167 · cited 1×
“Although our results are consistent with recent studies that showed the superior performance of the 1D-CNN compared with the 2D-CNN in a heart sound classification study [24], one has to be conscious of the fact that performance of these algorithms may change depending on the nature of the data.”
ResultComputer-aided diagnosis of laryngeal cancer via deep learning based on laryngoscopic images
Xiong, H., Lin, P., Yu, J.-G., Ye, J., Xiao, L., Tao, Y., Jiang, Z., Lin, W., Liu, M., Xu, J., et al. (2019) · EBioMedicine · doi:10.1016/j.ebiom.2019.08.075 · Cited by 162 · cited 1×
“86% [41], which are similar to our accuracy levels of 85.”
BackgroundSkuk, V. G., & Schweinberger, S. R. (2013) · Journal of Speech Language and Hearing Research · doi:10.1044/1092-4388(2013/12-0314 · Cited by 149
Trends in head and neck cancers in England from 1995 to 2011 and projections up to 2025
Louie, K. S., Mehanna, H. M., & Sasieni, P. D. (2015) · Oral Oncology · doi:10.1016/j.oraloncology.2015.01.002 · Cited by 143 · cited 1×
“Laryngeal cancer is known to show a skewed gender distribution [25,26] with an approximately four- to six-fold higher risk in males [27] and poor prognosis compared with females [28].”
BackgroundMueller, P. (1997) · Seminars in Speech and Language · doi:10.1055/s-2008-1064070 · Cited by 139 · cited 1×
“Likewise, voice changes in elderly males are characterized by an increase of jitter values [42].”
BackgroundAl-nasheri, A., Muhammad, G., Alsulaiman, M., & Ali, Z. (2016) · Journal of Voice · doi:10.1016/j.jvoice.2016.01.014 · Cited by 126
Ali, Z., Hossain, M. S., Muhammad, G., & Sangaiah, A. K. (2018) · Future Generation Computer Systems · doi:10.1016/j.future.2018.02.021 · Cited by 102
Pathological voice quality assessment using artificial neural networks
Ritchings, R. T., McGillion, M., & Moore, C. J. (2002) · Medical Engineering & Physics · doi:10.1016/s1350-4533(02)00064-4 · Cited by 90
Raitiola, J. P. H. (1999) · Acta Oto-Laryngologica · doi:10.1080/00016489950180531 · Cited by 87 · cited 1×
“In addition, the inclusion of these supraglottic cancer patients who usually remain asymptomatic in the early stages and are difficult to diagnose [35] may be clinically relevant.”
BackgroundByeon, H., & Cha, S. (2020) · Scientific Reports · doi:10.1038/s41598-020-61565-3 · Cited by 87 · cited 2×
“Instead, this feature is a more prominent marker in voice change observed in smokers [34].”
BackgroundMethodDirect speech feature estimation using an iterative EM algorithm for vocal fold pathology detection
Gavidia-Ceballos, L., & Hansen, J. H. L. (1996) · IEEE Transactions on Biomedical Engineering · doi:10.1109/10.486257 · Cited by 80 · cited 2×
“Gavidia-Ceballos and Hansen [22] demonstrated accuracy levels of 88.7% in patients with vocal fold cancer, but their sample was limited to 20 glottic cancer and 10 healthy subjects.”
BackgroundInfluentialLaryngeal cancer mortality trends in E uropean countries
Chatenoud, L., Garavello, W., Pagan, E., Bertuccio, P., Gallus, S., Vecchia, C. L., Negri, E., & Bosetti, C. (2015) · International Journal of Cancer · doi:10.1002/ijc.29833 · Cited by 73 · cited 1×
“Laryngeal cancer is known to show a skewed gender distribution [25,26] with an approximately four- to six-fold higher risk in males [27] and poor prognosis compared with females [28].”
BackgroundPolesel, J., Furlan, C., Birri, S., Giacomarra, V., Vaccher, E., Grando, G., Gobitti, C., Navarria, F., Schioppa, O., Minatel, E., et al. (2017) · Oral Oncology · doi:10.1016/j.oraloncology.2017.02.009 · Cited by 71 · cited 1×
“Despite the rising incidence, early diagnosis remains challenging, resulting in delayed treatment [2,3].”
BackgroundFraile, R., Sáenz-Lechón, N., Godino-Llorente, J. I., Osma-Ruiz, V., & Fredouille, C. (2009) · Folia Phoniatrica et Logopaedica · doi:10.1159/000219950 · Cited by 62 · cited 1×
“Despite our results favoring the 1D-CNN, due to the uneven gender distribution in laryngeal cancer, the gender composition of the data should be considered with caution when developing future deep learning algorithms since the female voice has broader distributions in cepstral domain analysis [29].”
ResultTao, J., Qin, C., Li, W., & Liu, C. (2019) · Sensors · doi:10.3390/s19153280 · Cited by 48
Changes in head and neck oncologic practice during the COVID ‐19 pandemic
Brody, R. M., Albergotti, W. G., Shimunov, D., Nicolli, E. A., Patel, U. A., Harris, B. N., & Bur, A. M. (2020) · Head & Neck · doi:10.1002/hed.26233 · Cited by 46
Causes for delay before specialist consultation in head and neck cancer
Nieminen, M., Aro, K., Jouhi, L., Bäck, L., Mäkitie, A., & Atula, T. (2018) · Acta Oncologica · doi:10.1080/0284186x.2018.1497297 · Cited by 42 · cited 1×
“Despite the rising incidence, early diagnosis remains challenging, resulting in delayed treatment [2,3].”
BackgroundBalaguer, M., Pommée, T., Farinas, J., Pinquier, J., Woisard, V., & Speyer, R. (2019) · Head & Neck · doi:10.1002/hed.25949 · Cited by 35 · cited 2×
“Subjective perception of early voice changes can be difficult [5].”
BackgroundVoice outcomes for early laryngeal cancer
Kinshuck, A. J., Shenoy, A., & Jones, T. M. (2017) · Current Opinion in Otolaryngology & Head & Neck Surgery · doi:10.1097/moo.0000000000000363 · Cited by 24 · cited 1×
“Controversies surrounding the role of fundamental frequencies exist, with some suggesting these values decrease in laryngeal cancer and smokers [31], compared to healthy groups [33].”
BackgroundIndividualized prediction of late‐onset dysphagia in head and neck cancer survivors
Aylward, A., Park, J., Abdelaziz, S., Hunt, J. P., Buchmann, L. O., Cannon, R. B., Rowe, K. G., Snyder, J., Deshmukh, V. G., Newman, M. G., et al. (2020) · Head & Neck · doi:10.1002/hed.26039 · Cited by 19 · cited 1×
“With a delay of diagnosis, laryngeal cancer may lead to the most severe debilitating disabilities in phonation, swallowing [4] and overall quality of life.”
Backgroundde Bruijn, M., ten Bosch, L., Kuik, D. J., Langendijk, J. A., Leemans, C. R., & Leeuw, I. M. V. (2011) · Logopedics Phoniatrics Vocology · doi:10.3109/14015439.2011.606227 · Cited by 18
DNN-based Approach to Detect and Classify Pathological Voice
Chuang, Z.-Y., Yu, X., Chen, J., Hsu, Y.-T., Xu, Z., Wang, C., Lin, F., & Fang, S. (2018) · IEEE International Conference on Big Data (BigData) · doi:10.1109/bigdata.2018.8622317 · Cited by 16 · cited 3×
“[9] have used normalized MFCCs features and have shown that a deep neural network (DNN) can detect abnormal voice changes in voice disorders.”
MethodResultDiverging incidence trends for larynx and tonsil cancer in low socioeconomic regions of the US
Tang, J. A., & Lango, M. N. (2019) · Oral Oncology · doi:10.1016/j.oraloncology.2019.02.024 · Cited by 15 · cited 1×
“Laryngeal cancer is known to show a skewed gender distribution [25,26] with an approximately four- to six-fold higher risk in males [27] and poor prognosis compared with females [28].”
BackgroundCancer larynx detection using glottal flow parameters and statistical tools
Aicha, A. B., & Ezzine, K. (2016) · International Symposium on Signal, Image, Video and Communications (ISIVC) · doi:10.1109/isivc.2016.7893963 · Cited by 14 · cited 1×
“A few studies have used features associated with hypernasalance in oropharyngeal cancer [38] and glottal flow in glottic cancer [39] and employed ANNs with mixed results.”
ResultFactors affecting voice quality in early glottic cancer before and after radiotherapy
Mekiš, J., Strojan, P., & Boltežar, I. H. (2019) · Radiology and Oncology · doi:10.2478/raon-2019-0050 · Cited by 9 · cited 2×
“Controversies surrounding the role of fundamental frequencies exist, with some suggesting these values decrease in laryngeal cancer and smokers [31], compared to healthy groups [33].”
BackgroundVoice Pathology Recognition and Classification using Noise Related Features
Hamdi, R., Salah, H., & Chérif, A. (2018) · International Journal of Advanced Computer Science and Applications · doi:10.14569/ijacsa.2018.091112 · Cited by 8 · cited 1×
“Abnormal HNR values reflect an asthenic voice and dysphonia [32], which may be expected with the mass effects.”
BackgroundJeon, J., Han, Y. J., Park, G., Sohn, D. G., Lee, S., & Im, S. (2019) · Clinical Neurophysiology · doi:10.1016/j.clinph.2019.06.005 · Cited by 2 · cited 1×
“Image analysis based on the use of computational algorithms in radiology is now expanding to signal processing in other fields such as electrodiagnosis [6].”
MethodSelf-citation
Cited by
Clinical Applications of Artificial Intelligence—An Updated Overview
Busnatu, Ș., Niculescu, A.-G., Bolocan, A., Petrescu, G. E. D., Păduraru, D. N., Năstasă, I., Lupușoru, M., Geantă, M., Andronic, O., Grumezescu, A. M., et al. (2022) · Journal of Clinical Medicine · doi:10.3390/jcm11082265 · Cited by 172
An Improvised Deep-Learning-Based Mask R-CNN Model for Laryngeal Cancer Detection Using CT Images
Sahoo, P. K., Mishra, S., Panigrahi, R., Bhoi, A. K., & Barsocchi, P. (2022) · Sensors · doi:10.3390/s22228834 · Cited by 113
“Furthermore, effective utilization of CE–NBI pictures in early prognosis majorly depends upon the otolaryngologist’s practice that demands long years of training in a significant manner [8,9].”
BackgroundRespiratory Diseases Diagnosis Using Audio Analysis and Artificial Intelligence: A Systematic Review
Kapetanidis, P., Kalioras, F., Tsakonas, C. G., Tzamalis, P., Kontogiannis, G., Karamanidou, T., Stavropoulos, T. G., & Nikoletseas, S. (2024) · Sensors · doi:10.3390/s24041173 · Cited by 113
“For example, voice pitch and volume changes could indicate an early sign of conditions such as laryngeal carcinoma and vocal cord dysfunction [3].”
BackgroundUr Rehman, M., Shafique, A., Azhar, Q.-U.-A., Jamal, S. S., Gheraibia, Y., & Usman, A. B. (2024) · Engineering Applications of Artificial Intelligence · doi:10.1016/j.engappai.2024.108047 · Cited by 73
Voice disorder classification using speech enhancement and deep learning models
Chaiani, M., Selouani, S., Boudraa, M., & Yakoub, M. S. (2022) · Journal of Applied Biomedicine · doi:10.1016/j.bbe.2022.03.002 · Cited by 71
Metadata
- Authors
- Kim, H., Jeon, J., Han, Y. J., Joo, Y., Lee, J., Lee, S., & Im, S.
- Affiliations
- Department of Otolaryngology-Head and Neck Surgery, Bucheon St. Mary’s Hospital, College of Medicine, The Catholic University of Korea, Seoul 06591, Korea
- Pohang University of Science and Technology (KR)
- The Catholic University of Korea Bucheon St. Mary's Hospital (KR)
- Department of Mechanical Engineering, Pohang University of Science and Technology (POSTECH), Pohang 37673, Korea
- Corresponding author
- Im, Sun
- Venue
- Journal of Clinical Medicine
- Publisher
- MDPI AG
- Published
- 2020-10-25
- Volume
- 9
- Issue
- 11
- Pages
- 3415
- ISSN
- 2077-0383 (ISSN-L)
- PMC
- PMC7692693
- OpenAlex
- W3094604100
- Cited by
- 97
- References
- 43
- Citations by year
- 2026: 6 · 2025: 23 · 2024: 28 · 2023: 16 · 2022: 16 · 2021: 7 · 2020: 1
- FWCI
- 6.02
- Type
- journal article
- Language
- en
- Open access
- gold
- doi.org
- Licence
- Locations
- Journal of Clinical Medicine (publishedVersion, cc-by)
- PubMed Central (publishedVersion, cc-by)
- PubMed (publishedVersion)
- Open Access System for Information Sharing (Pohang University of Science and Technology) (submittedVersion)
- Funders
- Ministry of Science and ICT, South Korea (2020R1A2C1009744.)
- National Research Foundation of Korea (2020R1F1A1065814)
- Postech-Catholic Biomedical Engineering Institute (No. 5-2019-B0001-00119)
- Topics
- Health Sciences › Health Professions › Speech and Hearing › Voice and Speech Disorders
- Physical Sciences › Computer Science › Artificial Intelligence › AI in cancer detection
- Physical Sciences › Computer Science › Artificial Intelligence › Speech Recognition and Synthesis
- Keywords
- laryngeal cancer, Mel-frequency cepstral coefficients, convolutional neural networks, 1D convolutional neural network, 2D convolutional neural network, Praat, vowel /a/, early cancer detection
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.