Addressing algorithmic and data challenges to deep learning based segmentation of spine anatomy

解决基于深度学习的脊柱解剖分割的算法和数据挑战

• 批准号: 10367207
• 负责人: Bilwaj Gaonkar
• 金额: $ 7.79万
• 依托单位: UNIVERSITY OF CALIFORNIA LOS ANGELES
• 依托单位国家: 美国
• 财政年份: 2022
• 资助国家: 美国
• 起止时间: 2022-08-01 至 2022-12-31
• 项目状态: 已结题
• 来源: https://reporter.nih.gov/project-details/10367207
• 关键词: Address; Affect; Age; Algorithms; Anatomy; Architecture; Area; Automobile Driving; Cervical; Clinical; Clinical Research; Clinical/Radiologic; Collection; Consensus; Consult; Data; Data Collection; Data Set; Databases; Decision Making; Dependence; Development; Diagnosis; Diagnostic; Discipline; Failure; Female; Goals; Health Care Costs; Heterogeneity; Human; Image; Intervertebral disc structure; Label; Laws; Link; Magnetic Resonance Imaging; Manuals; Mathematics; Measurement; Measures; Medical Imaging; Medicine; Methods; Modeling; Operative Surgical Procedures; Outcome; Pathology; Patient imaging; Patients; Performance; Physicians; Radiology Specialty; Recommendation; S-nitro-N-acetylpenicillamine; Sample Size; Scanning; Slice; Specificity; Spinal; Spinal Canal; Spinal Diseases; Spondylolisthesis; Standardization; Stenosis; System; Techniques; Technology; Testing; Training; Unnecessary Surgery; Validation; Variant; Vertebral column; Work; automated segmentation; base; biomedical imaging; computerized; cost; deep learning; deep learning algorithm; deep learning model; design; experimental study; imaging Segmentation; imaging biomarker; improved; interest; intervertebral disk degeneration; learning network; male; network architecture; novel; public database; quantitative imaging; relating to nervous system; segmentation algorithm; surgery outcome; treatment planning

Project Summary/Abstract In spine medicine, subjective interpretation of biomedical images often leads to wrong diagnoses, prolonged non-surgical treatment for surgically treatable patients, and surgical treatment when none is necessary. Objective computerized analysis of the aforementioned images using deep learning has the potential to improve surgical outcomes while driving down the cost of surgery by eliminating unnecessary surgery and expediting necessary ones. Yet several barriers stymie the development and deployment of deep learning technology to operationalize imaging biomarker-based treatment recommendation in surgical practice. First, a publicly available database is absent to help train and validate algorithms for spinal pathologies. Second deep learning techniques remain difficult to train and operationalize in the clinical setting, due to various challenges. These include – 1. The lack of a framework to link generalization error to training data in deep learning-based segmentation, due to which performance estimates of algorithms are untenable prior to deployment 2. the lack of a disciplined approach to improve deep network performance on medical image segmentation and 3. the lack of frameworks that enable deep networks to identify and flag a difficult case and failed cases where a human expert should be consulted. First, we propose to develop a publicly accessible spine imaging database to promote the development of deep learning algorithms. Second, we aim to address the aforementioned technical challenges by 1. Developing a power-law scaling based framework to link training sample size and generalization error analytically 2. Proposing and validating a mathematical framework to create deep learning ensembles from deep learning models to guarantee improvement in segmentation performance 3. Developing and validating a Von-Neumann information- based score to endow deep learning ensembles with the ability to identify difficult cases and predict failure.

项目总结/摘要 在脊柱医学中，对生物医学图像的主观解释常常导致错误的诊断、长时间的不确定性和不确定性。 非手术治疗可手术治疗的患者，和手术治疗时，没有必要。目的 使用深度学习对上述图像进行计算机化分析有可能改善外科手术。 结果，同时通过消除不必要的手术和加快必要的手术， 一个然而，一些障碍阻碍了深度学习技术的开发和部署， 在外科实践中基于成像生物标志物的治疗建议。首先，公开可用的数据库是 没有帮助训练和验证脊柱病理学的算法。第二，深度学习技术 由于各种挑战，难以在临床环境中进行培训和操作。其中包括-1。缺乏 在基于深度学习的分割中，将泛化错误与训练数据联系起来的框架， 算法的性能估计在部署2之前是站不住脚的。缺乏一个纪律性的方法， 提高深度网络在医学图像分割上的性能; 3.缺乏框架， 深入的网络，以确定和标记一个困难的情况下，失败的情况下，应咨询人类专家。 首先，我们建议开发一个可公开访问的脊柱成像数据库，以促进深度 学习算法其次，我们的目标是通过1.开发一 基于幂律标度的框架，将训练样本大小和泛化误差解析地联系起来2.提出 并验证数学框架，从深度学习模型创建深度学习集合， 保证分割性能的提高3.开发和验证冯诺依曼信息- 基于分数，赋予深度学习组合识别困难案例和预测失败的能力。