文献速递耳科与耳神经

神经影像洞察:梅尼埃病的结构改变与分类

Neuroimaging Insights: Structural Changes and Classification in Ménière's Disease.
PubMed 文献 · Ear Hear · 2024-05-24

本研究纳入55例单侧梅尼埃病(MD)患者与70例健康对照(HC),采用基于体素的形态测量和基于表面的形态测量分析神经影像结构差异,并构建机器学习分类模型。MD患者额叶和扣带回灰质体积显著减少,皮质厚度改变模式与情绪加工和感觉整合脑区相关。分类模型区分MD与HC的准确率达84%,曲线下面积(AUC)为0.92。

要点

  • MD患者额叶和扣带回灰质体积显著减少。
  • 皮质厚度改变见于情绪加工和感觉整合相关脑区。
  • 机器学习模型区分MD与HC准确率84%,AUC 0.92。

核心结论

  • MD患者存在灰质体积减少和特定皮质厚度改变模式,机器学习分类模型准确率84%,AUC 0.92,显示诊断潜力。

研究设计与样本

  • 纳入55例单侧MD患者和70例HC,采用基于体素的形态测量和基于表面的形态测量分析神经影像数据,并构建机器学习分类模型。

主要发现

  • MD患者灰质体积显著减少,尤其位于额叶和扣带回。
  • 皮质厚度改变模式见于情绪加工和感觉整合相关脑区。
  • 机器学习分类模型区分MD与HC的准确率为84%,AUC为0.92,精确率和召回率表现稳健,F1分数均衡。

关键数值与推荐

  • 机器学习分类模型区分MD与HC的准确率为84%。
  • 受试者工作特征曲线分析显示AUC为0.92。

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

OBJECTIVES: This study aimed to comprehensively investigate the neuroanatomical alterations associated with idiopathic Ménière's disease (MD) using voxel-based morphometry and surface-based morphometry techniques. The primary objective was to explore nuanced changes in gray matter volume, cortical thickness, fractal dimension, gyrification index, and sulcal depth in MD patients compared with healthy controls (HC). Additionally, we sought to develop a machine learning classification model utilizing these neuroimaging features to effectively discriminate between MD patients and HC. DESIGN: A total of 55 patients diagnosed with unilateral MD and 70 HC were enrolled in this study. Voxel-based morphometry and surface-based morphometry were employed to analyze neuroimaging data and identify structural differences between the two groups. The selected neuroimaging features were used to build a machine learning classification model for distinguishing MD patients from HC. RESULTS: Our analysis revealed significant reductions in gray matter volume in MD patients, particularly in frontal and cingulate gyri. Distinctive patterns of alterations in cortical thickness were observed in brain regions associated with emotional processing and sensory integration. Notably, the machine learning classification model achieved an impressive accuracy of 84% in distinguishing MD patients from HC. The model's precision and recall for MD and HC demonstrated robust performance, resulting in balanced F1-scores. Receiver operating characteristic curve analysis further confirmed the discriminative power of the model, supported by an area under the curve value of 0.92. CONCLUSIONS: This comprehensive investigation sheds light on the intricate neuroanatomical alterations in MD. The observed gray matter volume reductions and distinct cortical thickness patterns emphasize the disease's impact on neural structure. The high accuracy of our machine learning classification model underscores its diagnostic potential, providing a promising avenue for identifying MD patients. These findings contribute to our understanding of MD's neural underpinnings and offer insights for further research exploring the functional implications of structural changes.

来源:Ear Hear,查看原文。版权归原作者所有。

引用

  1. Li J, Cheng Q, Leng Y, et al. Neuroimaging Insights: Structural Changes and Classification in Ménière's Disease. Ear Hear. 2024. PMID: 38783421. doi:10.1097/AUD.0000000000001519

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。