Cardiovascular disease (CVD) is one of the leading causes of death worldwide. It is estimated that over 130 million adults in the US population (45.1%) are projected to suffer from CVD by 2035. When treating highly calcified coronary artery with stents, interventional cardiologists make stressful treatment decisions that can lead to inadequate stent deployment or calamitous events. Current therapeutic options such as cutting/scoring balloon, rotational/orbital atherectomy, or laser angioplasty might modify a severely calcified lesion before stent deployment; however, there are no clear guidelines to indicate the necessity for calcium modification. To address these challenges, I will create live-time stent intervention planning software, which will allow quantitative and comprehensive evaluations of coronary arteries for predicting acute stent deployment outcomes in intravascular optical coherence tomography (IVOCT) images. Specific aims are: 1) Develop IVOCT image analysis methods for automated plaque characterization and stent assessment using deep learning model. 2) Develop powerful machine/deep learning regression solutions to predict acute stent deployment outcomes in the presence of calcifications and evaluate using extensive in-vivo and ex-vivo data sets. I will use both the manually labeled in-vivo data as well as ex-vivo data set with multiple lesions including heavily calcified tissues. With success, the project will lead to an informed, evidence-based, future decision-support software for live-time treatment planning from IVOCT imaging. The project consortium will build on expertise in interventions, quantitative image analysis of IVOCT, and finite element modeling.
|Program type||Postdoctoral Fellowship|
|Effective start/end date||01/01/2020 → 12/31/2021|