使用卷积神经网络实现鼻窦计算机断层扫描放射学评分的自动化。
The use of a convolutional neural network to automate radiologic scoring of computed tomography of paranasal sinuses.
作者信息
Lee Daniel J, Hamghalam Mohammad, Wang Lily, Lin Hui-Ming, Colak Errol, Mamdani Muhammad, Simpson Amber L, Lee John M
机构信息
Department of Otolaryngology-Head and Neck Surgery, Unity Health TorontoSt. Michael's Hospital, University of Toronto, 30 Bond Street, 8 Cardinal Carter Wing, Toronto, ON, M5B 1W8, Canada.
Division of Otolaryngology-Head and Neck Surgery, Department of Surgery, North York General Hospital, University of Toronto, Toronto, Canada.
出版信息
Biomed Eng Online. 2025 Apr 27;24(1):49. doi: 10.1186/s12938-025-01376-7.
BACKGROUND
Chronic rhinosinusitis (CRS) is diagnosed with symptoms and objective endoscopy or computed tomography (CT). The Lund-Mackay score (LMS) is often used to determine the radiologic severity of CRS and make clinical decisions. This proof-of-concept study aimed to develop an automated algorithm combining a convolutional neural network (CNN) for sinus segmentation with post-processing to compute LMS directly from CT scans.
RESULTS
Radiology Information System was queried for outpatient paranasal sinus CTs at a tertiary institution. We identified 1,399 CT scans which were manually labelled with LMS of individual sinuses. Seventy-seven CT scans with 13,668 coronal images were segmented manually for individual sinuses. Our model for segmentation achieved a mean Dice score of 0.85 for all sinus regions, except for the osteomeatal complex. For individual Dice scores were 0.95, 0.71, 0.78, 0.93, 0.86 for the maxillary, anterior ethmoid, posterior ethmoid, sphenoid, and frontal sinuses, respectively. LMS was computed automatically by applying adaptive image thresholding and pixel counting to the CNN's segmented regions. A convolutional neural network (CNN) model was trained to segment each sinus region. Overall, the LMS model showed a high degree of accuracy with a score of 0.92, 0.99, 0.99, 0.97, 0.99, 0.86 for the maxillary, anterior ethmoid, posterior ethmoid, sphenoid, and frontal sinuses, respectively.
CONCLUSIONS
Reporting of paranasal sinus CT can be automated and potentially standardized with a CNN model to provide accurate Lund-Mackay score.
背景
慢性鼻-鼻窦炎（CRS）通过症状及客观的鼻内镜检查或计算机断层扫描（CT）进行诊断。Lund-Mackay评分（LMS）常用于确定CRS的放射学严重程度并做出临床决策。本概念验证研究旨在开发一种自动化算法，该算法结合用于鼻窦分割的卷积神经网络（CNN）及后处理，以直接从CT扫描计算LMS。
结果
在一家三级医疗机构的放射信息系统中查询门诊鼻窦CT。我们识别出1399例CT扫描，这些扫描对各个鼻窦的LMS进行了手动标注。对77例CT扫描的13668张冠状位图像进行了各个鼻窦的手动分割。除了骨窦复合体，我们的分割模型对所有鼻窦区域的平均Dice评分为0.85。上颌窦、前筛窦、后筛窦、蝶窦和额窦的个体Dice评分分别为0.95、0.71、0.78、0.93、0.86。通过对CNN分割区域应用自适应图像阈值处理和像素计数自动计算LMS。训练了一个卷积神经网络（CNN）模型来分割每个鼻窦区域。总体而言，LMS模型显示出高度准确性，上颌窦、前筛窦、后筛窦、蝶窦和额窦的评分分别为0.92、0.99、0.99、0.97、0.99、0.86。
结论
鼻窦CT报告可以通过CNN模型实现自动化并可能实现标准化，以提供准确的Lund-Mackay评分。
Zotero 插件