Personalized Management of Intracranial Aneurysms Using Computer-aided Analytics

摘要 主要目标是开发一种以人工智能为中心的定量和非侵入性软件 可集成到3D血管造影扫描仪（DSA、CTA或MRA）中以提供引导的平台 关于颅内动脉瘤（IA）的诊断和管理。出血性卒中继发于 颅内动脉瘤破裂导致显著的发病率和死亡率， 美国的无症状性颅内动脉瘤的诊断随着脑血管造影的使用而增加。 显像然而，关于应该治疗哪些动脉瘤的指导尚未取得进展。 利用计算科学和技术的最新进展，特别是人工智能， 基于两种使能技术构建的拟议软件平台可以（1）推动自动化的“患者特异性” 血流动力学评价纳入临床工作流程，（2）进行IA的“数据驱动”风险评估 在个人基础上破裂。具体的研究目标是：（1）开发一个面向临床的CFD平台 这使得能够对IA进行自动化的“患者特异性”血流动力学评价，（2）研究数据驱动的 分析预测IA的破裂风险，以及（3）在盲态下评估数据驱动的分析 study. 一旦确认，计划进行后续R01项目，以检查所提议软件的临床实用性 平台作为计算机辅助脑功能评价的单一网关 动脉瘤

Abstract The primary objective is to develop an artificial intelligence-centric, quantitative and noninvasive software platform that can be integrated into 3D angiographic scanners (DSA, CTA or MRA) to provide guidance regarding the diagnosis and management of intracranial aneurysms (IA). Hemorrhagic stroke secondary to ruptured IAs leads to significant morbidity and mortality and affects over 35,000 patients on a yearly basis in the United States. The diagnosis of asymptomatic IAs is on the rise with the increasing use of cerebral imaging. However, guidance regarding which aneurysms should be treated has not advanced. Leveraging recent advances in computational science and technology, particularly artificial intelligence, the proposed software platform built on two enabling technologies can (1) propel automated “patient-specific” hemodynamic evaluations into the clinical workflow and (2) conduct “data-driven” risk assessments of IA rupture on an individual basis. Specific research aims are to (1) develop a clinically-oriented CFD platform that enables automated “patient-specific” hemodynamic evaluations of IAs, (2) investigate data-driven analytics toward prediction of rupture risk for IAs and (3) evaluate the data-driven analytics in a blind study. Once validated, a follow-up R01 project is planned to examine the clinical utility of the proposed software platform in a prospective clinical study as a single gateway for computer-aided evaluation of cerebral aneurysms.