谷歌浏览器插件
订阅小程序
在清言上使用

New Compartmental Reading Method for MRI Enables Accurate Localization of Cholesteatomas with High Sensitivity and Specificity.

Otology & neurotology(2021)

引用 0|浏览5
暂无评分
摘要
Objectives: Cholesteatoma is an inflammatory disease, frequently observed in childrens and young adults, with a risk of relapse or recurrence. The few studies which analyzed cholesteatoma localization on magnetic resonance imaging (MRI) usually merged CT-MR images or relied on their authors’ anatomical knowledge. We propose a compartmental reading method of the compartments of the middle ear cavity for an accurate localization of cholesteatomas on MR images alone. Material and Methods: Our method uses easily recognizable anatomical landmarks, seen on both computed tomography (CT) and MRI, to delimit the middle ear compartments (epitympanum, mesotympanum, hypotympanum, retrotympanum, protympanum, antrum-mastoid cavity). We first tested it on 50 patients on non-enhanced temporal bone CT. Then, we evaluated its performances for the localization of cholesteatomas on MRI, compared with surgery on 31 patients (validation cohort). Results: The selected anatomical landmarks that delimited the middle ear compartments were applicable in 98 to 100% of the cases. In the validation cohort, we were able to accurately localize the cholesteatoma on MRI in 83% of the cases (n = 26) with high sensitivity (95.7%) and specificity (98.6%). Conclusion: With our compartmental reading method, based on the recognition of well-known anatomical landmarks to differentiate the compartments of the middle ear cavity on MRI, we were able to accurately localize the cholesteatoma with high (>90%) sensitivity and specificity. Such landmarks are widely applicable and only require limited learning time based on key images. Accurate localization of the cholesteatoma is useful for the choice of surgical approach.
更多
查看译文
关键词
Cholesteatoma,Epitympanum,Hypotympanum,Localization,Mesotympanum,Middle ear cavity,Magnetic resonance imaging,Retrotympanum
AI 理解论文
溯源树
样例
生成溯源树,研究论文发展脉络
Chat Paper
正在生成论文摘要