使用贝叶斯网络的可解释放射学方法来预测口腔癌中人类乳头瘤病毒的状态,使用贝叶斯网络
1Institute of Biomedical Engineering, Bogazici University, Istanbul, Turkey; Biomedical Engineering, Namik Kemal University, Tekirdağ, Turkey.
概括
这项研究开发了一个可解释的贝叶斯网络 (BN) 模型,使用CT扫描中的放射学来分类口腔癌患者的HPV状态. 该BN模型为HPV检测提供了传统实验室方法的非侵入性,可解释的替代方案.
科学领域:
- 在瘤学瘤学.
- 放射学 放射学是一门学科.
- 人工智能的人工智能
背景情况:
- 人类乳头瘤病毒 (HPV) 是口腔癌发展的关键因素.
- 目前的HPV检测方法通常依赖于侵入性实验室技术.
- 放射学提供了使用成像数据的非侵入性方法,但往往缺乏可解释性.
研究的目的:
- 开发一种简单,可解释的贝叶斯网络 (BN) 模型,用于对口腔癌中的HPV状况进行分类.
- 评估使用基于放射学的BN用于非侵入性HPV检测的可行性.
- 提供一个工具,帮助临床医生在治疗决策.
主要方法:
- 利用对比度增强的计算机断层扫描 (CT) 图像从246名口腔癌患者 (216名HPV阳性) 中获取.
- 提取了851个放射学特征,并采用了Mens eX Machina (MXM) 方法来确定关键预测因素:球性和max2DDiameterRow.
- 开发了一个BN模型,并将其性能 (曲线下的面积 - AUC) 与支持向量机 (SVM) 模型进行了比较.
主要成果:
- 在MXM方法中,确定了球状性和max2DDiameterRow作为最相关的预测指标.
- 在BN模型中,训练数据的AUC为0.78,测试数据的AUC为0.72.
- 使用 25 个特征进行比较的 SVM 模型在测试数据上获得了更高的 0.83 的 AUC.
结论:
- 开发的BN模型提供了一种可解释的,非侵入性的方法,用于从CT图像中检测HPV状态.
- 该模型的可解释性可以帮助临床医生做出明智的治疗决策.
- 虽然提供可解释性,但BN模型的准确性可能低于SVM等更复杂,更难以解释的方法.
更多相关视频
06:57Chromogenic In Situ Hybridization as a Tool for HPV-Related Head and Neck Cancer Diagnosis
Published on: June 14, 2019
10.4K
07:43Four-color Fluorescence Immunohistochemistry of T-cell Subpopulations in Archival Formalin-fixed, Paraffin-embedded Human Oropharyngeal Squamous Cell Carcinoma Samples
Published on: July 29, 2017
9.7K
相关概念视频
01:21Cancer Survival Analysis
381
Cancer survival analysis focuses on quantifying and interpreting the time from a key starting point, such as diagnosis or the initiation of treatment, to a specific endpoint, such as remission or death. This analysis provides critical insights into treatment effectiveness and factors that influence patient outcomes, helping to shape clinical decisions and guide prognostic evaluations. A cornerstone of oncology research, survival analysis tackles the challenges of skewed, non-normally...
381
01:15Receiver Operating Characteristic Plot
280
A ROC (Receiver Operating Characteristic) plot is a graphical tool used to assess the performance of a binary classification model by illustrating the trade-off between sensitivity (true positive rate) and specificity (false positive rate). By plotting sensitivity against 1 - specificity across various threshold settings, the ROC curve shows how well the model distinguishes between classes, with a curve closer to the top-left corner indicating a more accurate model. The area under the ROC curve...
280