报告人：王梁助理教授

东南大学生物科学与医学工程学院

题目：预测斑块增长的多风险因子策略：基于心血管超声随访数据的三维流固耦合模型研究

时间：2018年7月28日上午09:00

地点：海韵数理楼661

摘要：Objective: Plaque progression and vulnerability are influenced by many risk factors. Our goal is to find simple method to combine multiple risk factors for better plaque development prediction.

Approach: A sample size of 254 intravascular ultrasound (IVUS) slices with noticeable change of plaque burden was obtained from 9 patients who underwent baseline and follow-up IVUS, and fluid-structure interaction (FSI) modeling for plaque mechanical conditions. Data of six key morphological and biomechanical factors were extracted from each slice to predict plaque development measured by plaque burden (PB) increase defined as: PBI = (PB at follow-up) – (PB at baseline).

A multi-factor decision-making strategy was proposed to assign a binary predictive outcome YW (W represents any combination of these six factors) based on the simple “threshold-value” idea to predict the ground truth YPBI: YPBI=1 if PBI>0; YPBI=0 if PBI<0 . the threshold value was determined to achieve best agreement rate between yw and ypbi for each combination.

Results: The results showed that plaque wall stress (PWS) was the best single-factor predictor for PBI with agreement rate 63.4%. The slice group with YPWS=1 had 1.89 times greater odd to have PBI>0 than the slice group with YPWS=0. Among all 63 combinations, combining lipid percent (LP), PWS and wall shear stress (WSS) was the best predictor for PBI among all 63 combinations, achieving an agreement rate 68.5%.  

Conclusion: The method is simple, efficient and robust, and could be used to combine predictors from different sources to improve prediction accuracy and help decision-making in clinical practice.

报告人简介：王梁博士于2017年5月获得美国伍斯特理工大学博士学位。毕业后在东南大学生物科学与医学工程学院任讲师。王梁博士的研究方向为基于血管超声随访数据的三维流固耦合模型研究，并利用三维流固耦合计算模型对血管粥样硬化斑块的增长进行多风险因子预测。

联系人:黄雪莹副教授

欢迎广大师生参加！