Development and validation of a collaborative framework for assessment of peripheral facial paralysis using facial image regions of interest.
Background: While accurate evaluation of PFP is crucial for determining optimal treatment strategies, current clinical assessments rely heavily on subjective evaluations, leading to considerable variability between inter- and intra-observer ratings. Objective: This study aimed to develop and validat...
| Publicado en: | Acta Oto-Laryngologica Vol. 145; no. 8; pp. 759 - 770 |
|---|---|
| Autores principales: | , , , , , , , |
| Formato: | pictorial research tables/charts Journal Article |
| Publicado: |
Taylor & Francis Ltd
Aug2025
|
| Acceso en línea: | Ver este registro en EBSCOhost |
| Sumario: | Background: While accurate evaluation of PFP is crucial for determining optimal treatment strategies, current clinical assessments rely heavily on subjective evaluations, leading to considerable variability between inter- and intra-observer ratings. Objective: This study aimed to develop and validate a collaborative framework for evaluating PFP based on regions of interest in facial images. Methods: We developed and tested two approaches: (1) a collaborative framework integrating image interpretation techniques (representation learning via CNN) with predefined handcrafted features based on regions of interest in facial images, and (2) a convolutional neural network (CNN) model trained exclusively on full-face patient images. The diagnostic accuracy of both systems was evaluated using a test set and compared with otologists' assessments. Results: The collaborative framework achieved a mean Area Under the Curve (AUC) of 0.92 for PFP prediction in the test set, surpassing the 0.76 AUC achieved by the CNN trained on full-face images. The framework's performance matched that of experienced otologists (accuracy: 80.0% vs. 77.2%; sensitivity: 85.3% vs. 77.7%). Moreover, system assistance improved primary clinicians' mean accuracy by 17.7 percentage points. Conclusions: These findings demonstrate that our collaborative framework-based automated diagnosis system can effectively assist clinicians in PFP diagnosis. 虽然准确评估PFP对于确定最佳治疗策略至关重要, 但目前的临床评估高度依赖主观评价, 导致观察者评分之间和观察者自身评分之间存在相当大的差异。 本研究旨在开发并验证一个基于面部图像感兴趣区域(ROI)评估PFP的协作性框架。 我们开发并测试了两种方法: (1) 一个将图像解读技术(通过CNN进行表征学习)与基于面部图像感兴趣区域预先定义的手工特征相结合的协作性框架；以及 (2) 一个专门针对患者全脸图像进行训练的卷积神经网络(CNN)模型。使用测试集评估这两个系统的诊断准确性, 并与耳科医生的评估进行比较。 该协作性框架在测试集中对PFP预测的平均曲线下面积(AUC)达到0.92, 超过了在全脸图像上训练的CNN模型的0.76 AUC。该框架的性能与经验丰富的耳科医生相当(准确率: 80.0% 相较于 77.2%；灵敏度: 85.3% 相较于 77.7%)。此外, 系统辅助将初级临床医生的平均准确率提高了17.7个百分点。 这些结果表明, 我们的基于协作性框架的自动化诊断系统可以有效地协助临床医生进行PFP诊断。 |
|---|