Inducing Vascular Grammars for Anomaly Classification in Brain Angiograms
Inducing Vascular Grammars for Anomaly Classification in Brain Angiograms
复制标题
诱导血管语法用于脑血管造影中的异常分类
DOI：
10.1115/1.4053424
复制
发表时间：
2022
期刊：
影响因子：
--
通讯作者：
Cagan, Jonathan
中科院分区：
文献类型：
--
作者：
Whiting, Mark E.;Mettenburg, Joseph;Novelli, Enrico M.;Santini, Tales;Martins, Tiago;Ibrahim, Tamer S.;LeDuc, Philip R.;Cagan, Jonathan
随着机器学习在医疗诊断领域的应用，很少有方法能提供人类医生可以直接学习的诊断学。这项工作介绍了一种方法，利用人类可观察到的结构，如宏观血管形成，用于产生评估的医疗条件与相对较少的培训情况下，并发现模式，是潜在的诊断辅助工具。该方法借鉴了形状语法，一种基于规则的技术，在设计和架构中处于领先地位，并通过递归子图挖掘算法加速。规则实例在数据中的分布被用作中间表示，使得能够采用常见的分类和异常检测方法来识别具有相对较小的数据集的指示性规则。该方法适用于7特斯拉的飞行时间血管造影MRI（n = 54）的人脑血管。对数据进行分段和归纳，生成具有代表性的语法规则。将规则集合隔离以可靠地暗示血管状况。该应用程序演示了自动结构化中间表示的力量，用于评估细微差别的生物形式关系，以及形状语法的强度，特别是用于识别复杂血管网络中的指示性模式。
As machine learning is used to make strides in medical diagnostics, few methods provide heuristics from which human doctors can learn directly. This work introduces a method for leveraging human observable structures, such as macroscale vascular formations, for producing assessments of medical conditions with relatively few training cases, and uncovering patterns that are potential diagnostic aids. The approach draws on shape grammars, a rule-based technique, pioneered in design and architecture, and accelerated through a recursive subgraph mining algorithm. The distribution of rule instances in the data from which they are induced is then used as an intermediary representation enabling common classification and anomaly detection approaches to identify indicative rules with relatively small data sets. The method is applied to seven-tesla time-of-flight angiography MRI (n = 54) of human brain vasculature. The data were segmented and induced to generate representative grammar rules. Ensembles of rules were isolated to implicate vascular conditions reliably. This application demonstrates the power of automated structured intermediary representations for assessing nuanced biological form relationships, and the strength of shape grammars, in particular for identifying indicative patterns in complex vascular networks.