Decision Support System for Temporal Lobe Epilepsy

颞叶癫痫决策支持系统

• 批准号: 8727294
• 负责人: HAMID SOLTANIAN-ZADEH
• 金额: $ 41.61万
• 依托单位: HENRY FORD HEALTH SYSTEM
• 依托单位国家: 美国
• 财政年份: 2012
• 资助国家: 美国
• 起止时间: 2012-09-01 至 2016-08-31
• 项目状态: 已结题
• 来源: https://reporter.nih.gov/project-details/8727294
• 关键词: Amygdaloid structure; Archives; Area; Biological; Brain; Brain imaging; Characteristics; Clinical; Clinical Data; Computers; Data; Data Analyses; Databases; Decision Support Systems; Diagnosis; Diagnostic; Electrocorticogram; Electroencephalography; Epilepsy; Evaluation; Goals; Gold; Health Care Costs; Hippocampus (Brain); Human; Image; Image Analysis; Imagery; Individual; Magnetic Resonance Imaging; Manuals; Medial; Medical Imaging; Methods; Modality; Online Systems; Operative Surgical Procedures; Outcome; Patients; Pattern Recognition; Phase; Physicians; Postoperative Period; Preparation; Probability; Process; Prospective Studies; Recording of previous events; Research; Research Personnel; Retrospective Studies; Risk; Risk Factors; Role; Science; Sensitivity and Specificity; Shapes; Structure; Surface; System; Technology; Temporal Lobe; Temporal Lobe Epilepsy; Testing; Texture; Thalamic structure; Training; Validation; base; brain volume; clinical Diagnosis; clinical application; clinical decision-making; clinical practice; computerized; cost; data integration; data management; data mining; entorhinal cortex; evaluation/testing; imaging Segmentation; improved; knowledge base; multidisciplinary; novel; novel strategies; outcome forecast; predictive modeling; prospective; putamen; single photon emission computed tomography; treatment effect; treatment planning; user-friendly; web based interface

DESCRIPTION (provided by applicant): With the ever-increasing role of medical images in diagnosis, treatment planning, and evaluation of treatment effects, extraction of quantitative information from these images and efficient use of the results have become a necessity. In recent years, we have developed novel three-dimensional (3D) knowledge- based methods to segment brain structures from magnetic resonance images (MRI) automatically. These methods need to be optimized, fine-tuned, and compared to other methods for the segmentation of specific brain structures that may be involved in medial temporal lobe epilepsy (mTLE). Feature extraction methods also need to be developed and optimized to characterize (i.e., determine local and global multi- parametric intensity distribution, texture, shape, surface area, surface curvature, and volume of) the brain structures. Multi-modality analysis using multi-parametric MRI and SPECT needs to be developed for improved sensitivity and specificity. We have also developed our preliminary version of a content-based human brain image database system to hold the image analysis results with other clinical information (e.g., textual data) in a manner that can be searched, retrieved, and queried conveniently from any computer station. This system needs integrated methods for data preparation, missing value treatment, interactive rule-extraction, visualization, and user-inference to serve as a decision support system in clinical practice. A user-friendly, web-based interface will be critical for the ultimate use of the system by researchers and clinicians. Last but not least, the database needs to be populated with data from a large number of patients so that it can be confidently used for hypothesis testing and clinical applications. The goal of this project is to develop novel approaches for the above needs. Image analysis and feature extraction methods will segment and characterize hippocampus, amygdala, entorhinal cortex, thalamus, putamen, and other brain structures from MRI. The methods will be tested, evaluated, and validated using clinical data of epilepsy patients. Clinical diagnosis based on EEG studies and surgery outcome will be used as "gold standards" for evaluation and validation of the image analysis methods. The proposed decision support system will be populated with multi-modality data of 350 epilepsy patients to evaluate correlation between a variety of risk factors, imaging features, clinical diagnosis (lateralization), and post- operative outcomes, and to assist physicians with improved clinical diagnosis, reduced intracranial EEG studies (reduced risk and suffering of patients as well as their healthcare cost), optimal treatment options, and prediction of outcome in prospective studies. The proposed research will be a breakthrough in the application of computerized methods for medical image quantification and object characterization, and will advance image analysis science in the direction of integrating knowledge-based image segmentation and characterization methods with pattern recognition and data mining technology in decision support systems. The proposed approaches are applicable to the identification, segmentation, and characterization of other biological structures. They are also applicable to virtually any image analysis task for which object segmentation, quantification, and characterization are used.

描述（申请人提供）：随着医学图像在诊断、治疗计划和治疗效果评估中的作用日益增强，从这些图像中提取定量信息并有效利用结果已成为必然。近年来，我们开发了新颖的基于三维（3D）知识的方法，可以自动从磁共振图像（MRI）中分割大脑结构。这些方法需要优化、微调，并与其他方法进行比较，以分割可能与内侧颞叶癫痫 (mTLE) 相关的特定大脑结构。还需要开发和优化特征提取方法来表征（即确定局部和全局多参数强度分布、纹理、形状、表面积、表面曲率和体积）大脑结构。需要开发使用多参数 MRI 和 SPECT 的多模态分析，以提高敏感性和特异性。我们还开发了基于内容的人脑图像数据库系统的初步版本，以可以从任何计算机站方便地搜索、检索和查询的方式保存图像分析结果和其他临床信息（例如文本数据） 。该系统需要数据准备、缺失值处理、交互式规则提取、可视化和用户推理的集成方法，以作为临床实践中的决策支持系统。用户友好的、基于网络的界面对于研究人员和临床医生最终使用该系统至关重要。最后但并非最不重要的一点是，数据库需要填充大量患者的数据，以便可以放心地用于假设检验和临床应用。该项目的目标是开发满足上述需求的新方法。图像分析和特征提取方法将从 MRI 中分割和表征海马体、杏仁核、内嗅皮层、丘脑、壳核和其他大脑结构。这些方法将使用癫痫患者的临床数据进行测试、评估和验证。基于脑电图研究和手术结果的临床诊断将作为图像分析方法评估和验证的“金标准”。拟议的决策支持系统将填充 350 名癫痫患者的多模态数据，以评估各种危险因素、影像学特征、临床诊断（偏侧化）和术后结果之间的相关性，并协助医生改进临床诊断、减少颅内脑电图研究（降低患者的风险和痛苦以及医疗费用）、最佳治疗方案以及前瞻性研究结果的预测。所提出的研究将是计算机化方法在医学图像量化和对象表征应用方面的突破，并将推动图像分析科学朝着将基于知识的图像分割和表征方法与模式识别和数据挖掘技术在决策支持中相结合的方向发展。系统。所提出的方法适用于其他生物结构的识别、分割和表征。它们还适用于几乎任何使用对象分割、量化和表征的图像分析任务。