Skip to content Skip to navigation

Research Overview

Current projects in the Marsden Lab

This is a list of some of our ongoing projects.   Our research spans numerical methods development to clinical application of cardiovacsular modeling methods in pediatric and adult cardiovascular disease.  

Pediatric and Congenital Heart Disease

Collaborators: Dr. Jeffrey Feinstein, Tain Yen Hsia, Richard Figliola, Marlene Rabinovitch, Doff McElhinney, John Eaton
Lab members:   Ingrid Lan, Nicole Schiavone, Weiguang Yang, Chris  Elkins, Melody Dong
 
We apply computational modeling to pediatric and congenital heart disease for surgial planning, medical device design, and understanding biomechanics of disease progression.  Current applications include single-ventricle physiology, Tetralogy of Fallot, Williams and Alagille Syndromes, and pulmonary hypertension.  In single ventricle physiology, we have developed novel surgical methods, including the Fontan Y-graft and the Assisted Bidirectional Glenn.   In pulmonary hypertension, we are using multiscale modeling to understand mechanical forces leading to disease progression and remodeling of the vessel wall in the proximal and distal pulmonary arterial bed.   In Tetralogy of Fallot, we are combining experimental and computational methods to understand the wide variation of valve performance and longevity in children recieving surgical replacement valves.  In Williams and Alagille syndromes, we are comparing stenting vs. surgical interventions.  

Coronary Artery Hemodynamics

Collaborators: Andrew Kahn, MD, PhD, Jane Burns, MD, Jack Boyd, MD, Kristy Red-Horse, PhD

Lab members:  Owais Khan, Suhaas Anbazhakan, Jongmin Seo

 

Our lab is performing patient specific modeling of coronary artery hemodynamics in children and adults.  In adults, we are simulating coronary bypass graft surgery to elucidate causes of vein graft failure.   Vein grafts fail at alarmingly high rates of nearly 50% within 5-10 years of surgery.   We use multiscale modeling tools to characterize and compare the in vivo environment of arterial and vein grafts to uncover mechanisms for vein graft failure.  We also use virtual surgery to evaluate optimal surgical methods for individual patients.  In children, we are using patient specific modeling to evaluate the risk of thrombosis in patients with coronary artery aneurysms caused by Kawasaki disease.  We are also exploring the functional significance of collateral arteries in the coronaries in a collaboration with developmental biologists. 

Optimization and Uncertainty Quantification

Collaborators: Daniele Schiavazzi, John Dennis, Gianluca Geraci
Lab members: Casey Fleeter, Jongmin Seo
 
We are developing tools for optimization of cardiovascular geometries using derivative-free pattern search methods with surrogates.  We are investigating a range of clinically relevant cost functions and constraints including energy efficiency, wall shear stress, flow distribution and surgical feasibility.   We extended our surrogate optimization code to run in parallel in a high performance computing environment.   We are also developing a framework to perform comprehensive uncertainty analysis for cardiovascular simulations that provide confidence intervals on simulation outputs, taking into account sources of noise in simulation parameters such as image data artifacts, material properties, and clinical data.  We have developed a framework for multi-level multi-fidelity UQ that interfaces SimVascular with Dakota in collaboration with Sandia National Lab.  
 

Computational methods for Patient-Specific Modeling

Collaborators: Vijay Vedula, Irene Vignon-Clementel
Lab members:  Ju Liu, Ingrid Lan, Jongmin Seo
 
We are developing novel computational methods for patient-specific modeling and blood flow simulation.   We are working to improve image segmentation methods using machine learning and convolutional neural networks.  We are also developing fluid strucutre interaction methods to account for ventricular motion.   Multi-domain modeling allows us to couple 3D fluid dynamics simulations to reduced order models of the circulatory physiology, and we have developednovel coupling methods and linear solver methods to enhance efficiency.  We have recently proposed a novel unified formulation for fluid structure interaction using finite elements.