From voice biomarkers to telemedicine screening: developing and evaluating a voice-based AI model for laryngeal lesion detection using the Bridge2AI-Voice dataset - Takeaways - MDSpire

From voice biomarkers to telemedicine screening: developing and evaluating a voice-based AI model for laryngeal lesion detection using the Bridge2AI-Voice dataset

  • By

  • Phillip D. Jenkins

  • Steven Bedrick

  • Lisa Karstens

  • William Hersh

  • the Bridge2AI-Voice Consortium

  • David A. Dorr

  • July 1, 2026

  • 0 min

Share

  • 1

    The Bridge2AI-Voice initiative provides a large-scale, ethically sourced dataset for laryngeal lesion detection using voice analysis.

  • 2

    The study analyzed data from 205 participants to evaluate a voice-based AI model for high-sensitivity screening of laryngeal lesions.

  • 3

    The OpenSMILE-based model achieved an AUC of 0.812, with sensitivity of 0.870 and specificity of 0.566, surpassing an age-only baseline.

  • 4

    Subgroup analysis indicated consistent sensitivity across benign and precancerous lesion subgroups, with no incremental benefit from alternative features.

  • 5

    The findings suggest that the derived acoustic features support effective discrimination of vocal fold lesions, warranting further investigation.

Original Source(s)

From voice biomarkers to telemedicine screening: developing and evaluating a voice-based AI model for laryngeal lesion detection using the Bridge2AI-Voice dataset

  • by Phillip D. Jenkins, Steven Bedrick, Lisa Karstens, William Hersh, the Bridge2AI-Voice Consortium, David A. Dorr

  • July 1, 2026

Related Content

Occult clival chordoma initially unrecognized on MRI presenting with cerebrospinal fluid rhinorrhea: a case report

Predicting overall survival in synchronous metastatic nasopharyngeal carcinoma using a stacking ensemble machine learning model: a multicenter retrospective study

Combined biofeedback and vestibular rehabilitation therapy for vestibular migraine: clinical efficacy and neurobiochemical correlates