文献速递咽喉与嗓音

基于诊室喉频闪视频的多类音频深度学习模型对声带病理的分类

A Deep-Learning Model for Multi-class Audio Classification of Vocal Fold Pathologies in Office Stroboscopy.
PubMed 文献 · Laryngoscope · 2025-02-05

本研究回顾性纳入声带单侧麻痹105例、声带病变63例和健康声带41例,从喉频闪视频中提取声音数据生成梅尔频谱图,训练ResNet18模型进行二分类(健康与病理)和多分类(健康、麻痹、病变)。二分类在留出测试集准确率83%、F1分数0.90,多分类准确率40%、F1分数0.36;外部数据集上二分类准确率63%、F1分数0.48,多分类准确率35%、F1分数0.25。结论认为健康与病理声音可中等准确区分,但多分类降低性能,外部数据集表现差,喉频闪视频声音的诊断价值可能有限,需进一步研究。

要点

  • 二分类模型在留出测试集准确率83%、F1分数0.90,优于多分类的40%和0.36。
  • 外部数据集上二分类准确率63%、F1分数0.48,多分类准确率35%、F1分数0.25。
  • 作者认为喉频闪视频声音诊断价值可能有限,需进一步研究。

核心结论

  • 深度学习分类器能中等准确区分健康与病理声音,但多分类降低模型性能,外部数据集表现差,喉频闪视频声音的诊断价值可能有限,需进一步研究。

研究设计与样本

  • 回顾性识别声带单侧麻痹105例、声带病变63例和健康声带41例,从喉频闪视频提取声音样本。
  • 声音样本按患者分为训练集80%、验证集10%和测试集10%,训练预训练ResNet18模型进行二分类和多分类。
  • 外部验证数据集包括声带单侧麻痹12例、声带病变13例和健康声带15例。

主要发现

  • 留出测试集上,二分类准确率83%、F1分数0.90,多分类准确率40%、F1分数0.36。
  • 外部数据集上,二分类准确率63%、F1分数0.48,多分类准确率35%、F1分数0.25。

临床怎么用

  • 原文未据此给出推荐。

局限性

  • 原文未报告局限性。

以上解读依据论文摘要(全文未开放获取或许可不允许翻译转载)。

原文

OBJECTIVE: To develop and validate a deep-learning 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. METHODS: Patients with UVFP (n = 105), VF lesions (n = 63), and HVF (n = 41) 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-spectrograms. 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-score. RESULTS: When evaluated on a hold-out test set, the binary classifier demonstrated stronger performance compared to the multi-class classifier (accuracy 83% vs. 40%; F1-score 0.90 vs. 0.36). When evaluated on an external dataset, the binary classifier achieved an accuracy of 63% and F1-score of 0.48, compared to 35% and 0.25 for the multi-class classifier. CONCLUSIONS: Deep-learning 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-class 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. LEVEL OF EVIDENCE: 4 Laryngoscope, 135:2428-2436, 2025.

来源:Laryngoscope,查看原文。版权归原作者所有。

引用

  1. Kim YE, Dobko M, Li H, et al. A Deep-Learning Model for Multi-class Audio Classification of Vocal Fold Pathologies in Office Stroboscopy. Laryngoscope. 2025. PMID: 39907244. doi:10.1002/lary.32036

供执业医师工作参考,不替代临床判断;数值与分级以指南原文为准。