Risk stratification of uncomplicated type B aortic dissection using clinical and engineering analysis

项目摘要 B型主动脉夹层（TBAD）是一种致命的疾病， 主动脉的内层（内膜层），导致主动脉壁层分离（剥离） 产生"真"和"假"流明。存在任一器官的复杂性TBAD 灌注不良或主动脉破裂有很高的住院死亡率，需要紧急手术 或血管内治疗。简单的TBAD传统上是用最佳的 药物治疗（OMT），包括积极的抗高血压治疗和监测 显像OMT的住院死亡率低，但长期生存率低，为48- 66%，继发于主动脉瘤的总体无介入生存率低于50% 形成和破裂。这些不佳的长期结果支持了 治疗不复杂的TBAD。因此，存在以下迫切且未满足的临床需求： 及时识别那些在急性期OMT可能失败的无并发症的TBAD患者， 并因此受益于早期干预，例如胸血管内主动脉修复术（TEVAR）。 因此，本项目的目标是开发一个风险分层模型，用于预测 OMT失败和无并发症TBAD患者的最佳干预时机。到 为实现这一目标，将对约500例无并发症的TBAD进行回顾性分析 埃默里主动脉数据库的病人临床和解剖数据将从 电子病历和图像研究，以确定OMT失败的预测因素。接下来，使用 相同的患者数据库，将进行一系列机械实验，以获得 TBAD组织的超弹性和失效特性，从中破裂/撕裂风险指标将 发展。流体-结构相互作用（FSI）分析将得到验证和应用， 血流动力学和壁应力场的"热图"。因此，风险指数将 提取。对于具有纵向影像学数据的患者，将使用以下指标预测TBAD进展： 一个综合的生长和重塑（G & R）和解剖传播模型。关键 生物力学参数将被确定为OMT失败的潜在预测因子。最后， 机器学习（ML）技术将用于联合收割机临床和生物力学预测 开发一个多因素的个性化TBAD风险分层模型。评价 为了评估所提出的方法的性能，我们将招募并进行纵向随访研究 35例急性无并发症TBAD患者，通过比较ML模型， 预测结果与实际临床结果。

Project Summary Type B Aortic Dissection (TBAD) is a lethal disease which occurs when a tear develops in the inner lining (intimal layer) of the aorta, causing the layers of the aortic wall to separate (dissect) creating “true” and “false” lumens. Complicated TBADs with presence of either organ malperfusion or aortic rupture have a high in-hospital mortality rate and require emergent surgical or endovascular therapy. Uncomplicated TBADs have been traditionally managed with optimal medical therapy (OMT) consisting of aggressive anti-hypertensive therapy and surveillance imaging. OMT results in low in-hospital mortality rates, but dismal long-term survival rates of 48- 66%, and overall intervention-free survival rates of less than 50% secondary to aortic aneurysm formation and rupture. These poor long-term outcomes support a paradigm change in the treatment of the uncomplicated TBADs. Thus, there is an urgent and unmet clinical need for promptly identifying those uncomplicated TBAD patients that will likely fail OMT in the acute phase, and thus benefit from early intervention such as Thoracic Endovascular Aortic Repair (TEVAR). Therefore, the objective of this project is to develop a risk stratification model for predicting both failure of OMT and the optimal timing of intervention in uncomplicated TBAD patients. To achieve this goal, a retrospective analysis will be conducted for about 500 uncomplicated TBAD patients from the Emory Aortic Databank. Clinical and anatomic data will be harvested from the electronic medical record and image studies to identify predictors of OMT failure. Next, using the same patient database, a series of mechanical experiments will be performed to obtain hyperelastic and failure properties of the TBAD tissues, from which rupture/tear risk metrics will be developed. Fluid-structure interaction (FSI) analyses will be validated and applied to obtain “heat maps” of hemodynamic and wall stress fields. The risk indices will be consequently extracted. For patients with longitudinal imaging data, TBAD progression will be predicted using an integrated growth and remodeling (G&R) and dissection propagation model. Critical biomechanical parameters will be identified as potential predictors of OMT failure. Finally, machine learning (ML) techniques will be used to combine clinical and biomechanical predictors to develop a multi-factorial, personalized TBAD risk stratification model. To evaluate the performance of the proposed approach, we will recruit and perform a longitudinal follow-up study of 35 acute uncomplicated TBAD patients to validate our approach by comparing the ML-model- prediction results with actual clinical outcomes.