文献速递睡眠呼吸障碍咽喉与嗓音

利用深度学习预测舌下神经刺激的治疗反应

Predicting Therapeutic Response to Hypoglossal Nerve Stimulation Using Deep Learning.
PubMed 文献 · Laryngoscope · 2024-06-27

该研究在三级转诊中心纳入接受药物诱导睡眠内镜(DISE)及后续舌下神经刺激器(HGNS)植入的患者,利用DISE图像训练6种深度学习模型和5种机器学习算法,以Sher标准定义反应者。结果显示基于腭咽(VP)图像的模型总体准确率高于舌根(BOT)或两者联合,VCG-16深度学习模型在VP图像集上表现最佳(训练准确率0.833)。作者认为深度神经网络有潜力预测HGNS疗效,需多机构数据开发通用模型。

要点

  • 基于腭咽图像的模型总体准确率高于舌根或联合图像集
  • VCG-16深度学习模型在腭咽图像集上训练准确率0.833、F1分数0.78、召回率0.883
  • 逻辑回归模型在机器学习模型中准确率最高(0.685)

核心结论

  • 深度神经网络有潜力利用DISE图像预测HGNS治疗疗效,有助于更好地选择植入患者。
  • 开发多机构数据和图像集将有助于建立可推广的预测模型。

研究设计与样本

  • 在三级转诊中心纳入接受DISE及后续HGNS植入的患者。
  • 共纳入127例患者的25,040张图像,其中16,515张(69.3%)来自反应者,8,262张(30.7%)来自非反应者。
  • 训练6种深度学习模型和5种机器学习算法,使用舌根和腭咽图像。
  • 反应者按Sher标准定义:呼吸暂停低通气指数(AHI)降低50%且AHI<15次/小时。
  • 评估指标包括精确率、召回率、F1分数和总体准确率。

主要发现

  • 基于腭咽数据集的模型总体准确率高于单独舌根及腭咽与舌根联合图像集。
  • VCG-16深度学习模型在腭咽图像集上总体表现最佳,训练准确率0.833、F1分数0.78、召回率0.883。
  • 机器学习模型中逻辑回归模型准确率最高(0.685),F1分数0.813。

临床怎么用

  • 作者认为深度神经网络有潜力利用DISE图像预测HGNS治疗疗效,有助于更好地选择植入患者。
  • 作者指出开发多机构数据和图像集将有助于建立可推广的预测模型。

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

OBJECTIVES: To develop and validate machine learning (ML) and deep learning (DL) models using drug-induced sleep endoscopy (DISE) images to predict the therapeutic efficacy of hypoglossal nerve stimulator (HGNS) implantation. METHODS: Patients who underwent DISE and subsequent HGNS implantation at a tertiary care referral center were included. Six DL models and five ML algorithms were trained on images from the base of tongue (BOT) and velopharynx (VP) from patients classified as responders or non-responders as defined by Sher's criteria (50% reduction in apnea-hypopnea index (AHI) and AHI < 15 events/h). Precision, recall, F1 score, and overall accuracy were evaluated as measures of performance. RESULTS: In total, 25,040 images from 127 patients were included, of which 16,515 (69.3%) were from responders and 8,262 (30.7%) from non-responders. Models trained on the VP dataset had greater overall accuracy when compared to BOT alone and combined VP and BOT image sets, suggesting that VP images contain discriminative features for identifying therapeutic efficacy. The VCG-16 DL model had the best overall performance on the VP image set with high training accuracy (0.833), F1 score (0.78), and recall (0.883). Among ML models, the logistic regression model had the greatest accuracy (0.685) and F1 score (0.813). CONCLUSION: Deep neural networks have potential to predict HGNS therapeutic efficacy using images from DISE, facilitating better patient selection for implantation. Development of multi-institutional data and image sets will allow for development of generalizable predictive models. LEVEL OF EVIDENCE: NA Laryngoscope, 134:5210-5216, 2024.

来源:Laryngoscope,查看原文。版权归原作者所有。

引用

  1. Alapati R, Renslo B, Jackson L, et al. Predicting Therapeutic Response to Hypoglossal Nerve Stimulation Using Deep Learning. Laryngoscope. 2024. PMID: 38934474. doi:10.1002/lary.31609

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。