{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,9,22]],"date-time":"2026-09-22T20:14:19Z","timestamp":1790108059030,"version":"4.0.1"},"reference-count":47,"publisher":"Wiley","issue":"7","license":[{"start":{"date-parts":[[2025,2,5]],"date-time":"2025-02-05T00:00:00Z","timestamp":1738713600000},"content-version":"vor","delay-in-days":0,"URL":"http:\/\/onlinelibrary.wiley.com\/termsAndConditions#vor"}],"funder":[{"DOI":"10.13039\/100000049","name":"National Institute on Aging","doi-asserted-by":"publisher","award":["K76AG079040"],"award-info":[{"award-number":["K76AG079040"]}],"id":[{"id":"10.13039\/100000049","id-type":"DOI","asserted-by":"publisher"}]},{"DOI":"10.13039\/100015326","name":"Common Fund","doi-asserted-by":"publisher","award":["OT2 OD032720"],"award-info":[{"award-number":["OT2 OD032720"]}],"id":[{"id":"10.13039\/100015326","id-type":"DOI","asserted-by":"publisher"}]}],"content-domain":{"domain":["onlinelibrary.wiley.com"],"crossmark-restriction":true},"short-container-title":["The Laryngoscope"],"published-print":{"date-parts":[[2025,7]]},"abstract":"<jats:sec>\n                    <jats:title>Objective<\/jats:title>\n                    <jats:p>To develop and validate a deep\u2010learning classifier trained on voice data extracted from videolaryngostroboscopy recordings, differentiating between three different vocal fold (VF) states: healthy (HVF), unilateral paralysis (UVFP), and VF lesions, including benign and malignant pathologies.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Methods<\/jats:title>\n                    <jats:p>\n                      Patients with UVFP (\n                      <jats:italic>n<\/jats:italic>\n                      \u2009=\u2009105), VF lesions (\n                      <jats:italic>n<\/jats:italic>\n                      \u2009=\u200963), and HVF (\n                      <jats:italic>n<\/jats:italic>\n                      \u2009=\u200941) were retrospectively identified. Voice samples were extracted from stroboscopic videos (Pentax Laryngeal Strobe Model 9400), including sustained \/i\/ phonation, pitch glide, and \/i\/ sniff task. Extracted audio files were converted into Mel\u2010spectrograms. Voice samples were independently divided into training (80%), validation (10%), and test (10%) by patient. Pretrained ResNet18 models were trained to classify (1) HVF and pathological VF (lesions and UVFP), and (2) HVF, UVFP, and VF lesions. Both classifiers were further validated on an external dataset consisting of 12 UVFP, 13 VF lesions, and 15 HVF patients. Model performances were evaluated by accuracy and F1\u2010score.\n                    <\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Results<\/jats:title>\n                    <jats:p>When evaluated on a hold\u2010out test set, the binary classifier demonstrated stronger performance compared to the multi\u2010class classifier (accuracy 83% vs. 40%; F1\u2010score 0.90 vs. 0.36). When evaluated on an external dataset, the binary classifier achieved an accuracy of 63% and F1\u2010score of 0.48, compared to 35% and 0.25 for the multi\u2010class classifier.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Conclusions<\/jats:title>\n                    <jats:p>Deep\u2010learning classifiers differentiating HVF, UVFP, and VF lesions were developed using voice data from stroboscopic videos. Although healthy and pathological voice were differentiated with moderate accuracy, multi\u2010class classification lowered model performance. The model performed poorly on an external dataset. Voice captured in stroboscopic videos may have limited diagnostic value, though further studies are needed.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Level of Evidence<\/jats:title>\n                    <jats:p>\n                      4\n                      <jats:italic>Laryngoscope<\/jats:italic>\n                      , 135:2428\u20132436, 2025\n                    <\/jats:p>\n                  <\/jats:sec>","DOI":"10.1002\/lary.32036","type":"journal-article","created":{"date-parts":[[2025,2,5]],"date-time":"2025-02-05T07:14:06Z","timestamp":1738739646000},"page":"2428-2436","update-policy":"https:\/\/doi.org\/10.1002\/crossmark_policy","source":"Crossref","is-referenced-by-count":9,"title":["A Deep\u2010Learning Model for Multi\u2010class Audio Classification of Vocal Fold Pathologies in Office Stroboscopy"],"prefix":"10.1002","volume":"135","author":[{"ORCID":"https:\/\/orcid.org\/0009-0008-7201-7545","authenticated-orcid":false,"given":"Yeo E.","family":"Kim","sequence":"first","affiliation":[{"name":"Sean Parker Institute for the Voice, Department of Otolaryngology\u2010Head and Neck Surgery Weill Cornell Medicine  New York New York U.S.A."}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Maria","family":"Dobko","sequence":"additional","affiliation":[{"name":"Sean Parker Institute for the Voice, Department of Otolaryngology\u2010Head and Neck Surgery Weill Cornell Medicine  New York New York U.S.A."}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Haomiao","family":"Li","sequence":"additional","affiliation":[{"name":"Cornell Tech  New York New York U.S.A."}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Tianlan","family":"Shao","sequence":"additional","affiliation":[{"name":"Sean Parker Institute for the Voice, Department of Otolaryngology\u2010Head and Neck Surgery Weill Cornell Medicine  New York New York U.S.A."}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Preethi","family":"Periyakoil","sequence":"additional","affiliation":[{"name":"Sean Parker Institute for the Voice, Department of Otolaryngology\u2010Head and Neck Surgery Weill Cornell Medicine  New York New York U.S.A."}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Courtney","family":"Tipton","sequence":"additional","affiliation":[{"name":"Sean Parker Institute for the Voice, Department of Otolaryngology\u2010Head and Neck Surgery Weill Cornell Medicine  New York New York U.S.A."}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-1617-1133","authenticated-orcid":false,"given":"Christine","family":"Colasacco","sequence":"additional","affiliation":[{"name":"The Center for Voice and Swallowing, Department of Otolaryngology\u2010Head and Neck Surgery, New York\u2010Presbyterian Hospital Columbia University Irvine Medical Center  New York City New York U.S.A."}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Aisha","family":"Serpedin","sequence":"additional","affiliation":[{"name":"Weill Cornell Medicine\u2010Qatar  Doha Qatar"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Olivier","family":"Elemento","sequence":"additional","affiliation":[{"name":"Englander Institute for Precision Medicine Weill Cornell Medicine  New York New York U.S.A."}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Mert","family":"Sabuncu","sequence":"additional","affiliation":[{"name":"Department of Radiology Weill Cornell Medicine  New York New York U.S.A."}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0001-6637-4302","authenticated-orcid":false,"given":"Michael","family":"Pitman","sequence":"additional","affiliation":[{"name":"The Center for Voice and Swallowing, Department of Otolaryngology\u2010Head and Neck Surgery, New York\u2010Presbyterian Hospital Columbia University Irvine Medical Center  New York City New York U.S.A."}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0001-8989-2073","authenticated-orcid":false,"given":"Lucian","family":"Sulica","sequence":"additional","affiliation":[{"name":"Sean Parker Institute for the Voice, Department of Otolaryngology\u2010Head and Neck Surgery Weill Cornell Medicine  New York New York U.S.A."}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0003-1543-2634","authenticated-orcid":false,"given":"Ana\u00efs","family":"Rameau","sequence":"additional","affiliation":[{"name":"Sean Parker Institute for the Voice, Department of Otolaryngology\u2010Head and Neck Surgery Weill Cornell Medicine  New York New York U.S.A."}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"311","published-online":{"date-parts":[[2025,2,5]]},"reference":[{"key":"e_1_2_6_2_1","volume-title":"Stroboscopy","author":"Woo P","year":"2009"},{"key":"e_1_2_6_3_1","doi-asserted-by":"publisher","DOI":"10.1002\/lary.28539"},{"key":"e_1_2_6_4_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.amjoto.2021.103327"},{"key":"e_1_2_6_5_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.jvoice.2021.05.002"},{"key":"e_1_2_6_6_1","doi-asserted-by":"publisher","DOI":"10.1097\/MOO.0000000000000697"},{"key":"e_1_2_6_7_1","doi-asserted-by":"publisher","DOI":"10.1002\/lary.29886"},{"key":"e_1_2_6_8_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.jvoice.2020.08.003"},{"key":"e_1_2_6_9_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.amjoto.2023.103800"},{"key":"e_1_2_6_10_1","doi-asserted-by":"publisher","DOI":"10.1002\/lio2.754"},{"key":"e_1_2_6_11_1","doi-asserted-by":"publisher","DOI":"10.1001\/jamaoto.2023.4807"},{"key":"e_1_2_6_12_1","doi-asserted-by":"crossref","unstructured":"HanY LiS CaoY MaQ YoshikawaM.Voice\u2010indistinguishability: protecting voiceprint in privacy\u2010preserving speech data release. 2020 IEEE International Conference on Multimedia and Expo (ICME) 2020:1\u20136.","DOI":"10.1109\/ICME46284.2020.9102875"},{"key":"e_1_2_6_13_1","doi-asserted-by":"publisher","DOI":"10.1111\/1460-6984.12783"},{"key":"e_1_2_6_14_1","doi-asserted-by":"publisher","DOI":"10.1109\/EMBC.2018.8513222"},{"key":"e_1_2_6_15_1","doi-asserted-by":"publisher","DOI":"10.3390\/s22176634"},{"key":"e_1_2_6_16_1","doi-asserted-by":"publisher","DOI":"10.1002\/lio2.259"},{"key":"e_1_2_6_17_1","doi-asserted-by":"publisher","DOI":"10.2196\/25247"},{"key":"e_1_2_6_18_1","doi-asserted-by":"publisher","DOI":"10.1016\/S0892-1997(01)00037-6"},{"key":"e_1_2_6_19_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.otc.2007.05.011"},{"key":"e_1_2_6_20_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.jvoice.2023.12.007"},{"key":"e_1_2_6_21_1","doi-asserted-by":"crossref","unstructured":"McFeeB RaffelC LiangD et al.Librosa: audio and music signal analysis in Python.2015.","DOI":"10.25080\/Majora-7b98e3ed-003"},{"key":"e_1_2_6_22_1","unstructured":"audEERING GmbH.Opensmile (version 2.3).2017."},{"key":"e_1_2_6_23_1","doi-asserted-by":"crossref","unstructured":"LowDM RaoV RandolphG SongPC GhoshSS.Identifying bias in models that detect vocal fold paralysis from audio recordings using explainable machine learning and clinician ratings. medRxiv.2024.","DOI":"10.1371\/journal.pdig.0000516"},{"key":"e_1_2_6_24_1","doi-asserted-by":"publisher","DOI":"10.1109\/TAFFC.2015.2457417"},{"key":"e_1_2_6_25_1","first-page":"1","article-title":"A deep learning approach for automated diagnosis and multi\u2010class classification of Alzheimer's disease stages using resting\u2010state fMRI and residual neural networks","volume":"44","author":"Ramzan F","year":"2019","journal-title":"J Med Syst"},{"key":"e_1_2_6_26_1","doi-asserted-by":"publisher","DOI":"10.3390\/bioengineering11030295"},{"key":"e_1_2_6_27_1","doi-asserted-by":"crossref","unstructured":"HeK ZhangX RenS SunJ.Deep residual learning for image recognition.2016:770\u2013778.","DOI":"10.1109\/CVPR.2016.90"},{"key":"e_1_2_6_28_1","unstructured":"LoshchilovI HutterF.Decoupled weight decay regularization. arXiv preprint arXiv: 171105101 2017."},{"key":"e_1_2_6_29_1","doi-asserted-by":"crossref","unstructured":"HuangJ XianhuiW QiH ChengJ TaoranZ.Diabetic retinopathy detection with weighted cross\u2010entropy loss.2023:8814\u20138819.","DOI":"10.23919\/CCC58697.2023.10240083"},{"key":"e_1_2_6_30_1","article-title":"Pytorch: an imperative style, high\u2010performance deep learning library","volume":"32","author":"Paszke A","year":"2019","journal-title":"Adv Neural Inform Process Syst"},{"key":"e_1_2_6_31_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.jrras.2024.101080"},{"key":"e_1_2_6_32_1","unstructured":"LundbergS.A unified approach to interpreting model predictions. arXiv preprint arXiv:1705078742017."},{"key":"e_1_2_6_33_1","article-title":"Diagnosis of early glottic cancer using laryngeal image and voice based on ensemble learning of convolutional neural network classifiers","author":"Kwon I","year":"2022","journal-title":"J Voice"},{"key":"e_1_2_6_34_1","doi-asserted-by":"publisher","DOI":"10.1017\/ATSIP.2019.7"},{"key":"e_1_2_6_35_1","doi-asserted-by":"publisher","DOI":"10.1038\/s41746-022-00712-8"},{"key":"e_1_2_6_36_1","doi-asserted-by":"publisher","DOI":"10.1002\/lary.31303"},{"key":"e_1_2_6_37_1","doi-asserted-by":"publisher","DOI":"10.1109\/TASL.2010.2104141"},{"key":"e_1_2_6_38_1","first-page":"147","volume-title":"Vocal Fold Physiology: Voice Quality Control","author":"Stevens KN","year":"1995"},{"key":"e_1_2_6_39_1","first-page":"128","article-title":"Sound spectrographic analysis of the voice of patients with recurrent Larngeal nerve paralysis: the Korean Society of Laryngology","author":"Kim K\u2010M","year":"1995","journal-title":"Phoniatr Logoped"},{"key":"e_1_2_6_40_1","doi-asserted-by":"publisher","DOI":"10.1016\/S0892-1997(96)80019-1"},{"key":"e_1_2_6_41_1","doi-asserted-by":"crossref","unstructured":"SeedatN AharonsonV HamzanyY.Automated and interpretable m\u2010health discrimination of vocal cord pathology enabled by machine learning 2020 IEEE Asia\u2010Pacific conference on computer science and data engineering (CSDE).2020:1\u20136.","DOI":"10.1109\/CSDE50874.2020.9411529"},{"key":"e_1_2_6_42_1","doi-asserted-by":"publisher","DOI":"10.1109\/ACCESS.2018.2816338"},{"key":"e_1_2_6_43_1","doi-asserted-by":"publisher","DOI":"10.3390\/electronics11223795"},{"key":"e_1_2_6_44_1","doi-asserted-by":"publisher","DOI":"10.1002\/ohn.809"},{"key":"e_1_2_6_45_1","doi-asserted-by":"publisher","DOI":"10.1002\/lary.31260"},{"key":"e_1_2_6_46_1","article-title":"Visualizing data using t\u2010SNE","volume":"9","author":"Van der Maaten L","year":"2008","journal-title":"J Mach Learn Res"},{"key":"e_1_2_6_47_1","doi-asserted-by":"crossref","unstructured":"SelvarajuRR CogswellM DasA VedantamR ParikhD BatraD.Grad\u2010CAM: visual explanations from deep networks via gradient\u2010based localization 2017 IEEE international conference on computer vision (ICCV).2017:618\u2013626.","DOI":"10.1109\/ICCV.2017.74"},{"key":"e_1_2_6_48_1","doi-asserted-by":"publisher","DOI":"10.1038\/s41598-023-34461-9"}],"container-title":["The Laryngoscope"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/onlinelibrary.wiley.com\/doi\/pdf\/10.1002\/lary.32036","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2025,11,14]],"date-time":"2025-11-14T05:32:29Z","timestamp":1763098349000},"score":1,"resource":{"primary":{"URL":"https:\/\/onlinelibrary.wiley.com\/doi\/10.1002\/lary.32036"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2025,2,5]]},"references-count":47,"journal-issue":{"issue":"7","published-print":{"date-parts":[[2025,7]]}},"alternative-id":["10.1002\/lary.32036"],"URL":"https:\/\/doi.org\/10.1002\/lary.32036","archive":["Portico"],"relation":{},"ISSN":["0023-852X","1531-4995"],"issn-type":[{"value":"0023-852X","type":"print"},{"value":"1531-4995","type":"electronic"}],"subject":[],"published":{"date-parts":[[2025,2,5]]},"assertion":[{"value":"2024-05-02","order":0,"name":"received","label":"Received","group":{"name":"publication_history","label":"Publication History"}},{"value":"2025-01-16","order":2,"name":"accepted","label":"Accepted","group":{"name":"publication_history","label":"Publication History"}},{"value":"2025-02-05","order":3,"name":"published","label":"Published","group":{"name":"publication_history","label":"Publication History"}}]}}