Why do some people develop malfunction of their mitral valve (MV) after a heart attack? The answers are complex, and treatments are not yet perfected. But biomedical engineers are hot on the trail of better solutions thanks to advancements in computer modeling and simulation.

Fig. 1 - Submodel of the mitral valve includes the annulus (entire red area), the leaflets (the anterior and posterior portions of the annulus are the flaps that open and close the valve), and the chordae tendons (yellow) that attach to the inside wall of the heart ventricle and control valve function.

Functional mitral regurgitation (FMR) is fairly common following a myocardial infarction (MI).1 Characterized by incomplete closure of the critical valve between the left atrium and ventricle, FMR allows oxygenated blood to seep back into the atrium instead of being pushed out the ventricle to the rest of the body. The patient becomes out-of-breath and fatigued.

Surgeons can restore MV function by implanting a flexible annuloplasty ring that constricts the opening of the valve (the annulus) so its two leaflets can meet (coapt) properly again.2 However, regurgitation recurs in more than 60 percent of patients within three to five years of this kind of repair.3

Project engineer Yunjie Wang, PhD, along with colleagues at Thornton Tomasetti Applied Science, decided to look more deeply into this clinical challenge. “For some patients, the repair fails after several years because the design of the ring wasn't ideal for their particular anatomy in the first place,” she says. “In other cases, remodeling of heart tissue after an MI gradually changes the shape of the annulus so it no longer matches the ring that was implanted. We need deeper insight into these kinds of patient-specific differences, so we can customize annuloplasty rings to fit individual hearts.”

Although she earned her bachelor's and master's degrees in mechanical engineering and solid mechanics, Wang turned to biomechanics for her doctorate work. “I switched because I realized that, while we understand the mechanics of things like rockets and cars very well, we still don't understand all we need to about the human body,” she says. “Even after five years studying the structure and behavior of blood vessels, I still have questions. That's what makes our work so compelling.”

Refining a Mitral Valve Simulation Inside The Living Heart Model

Wang's team has a virtual laboratory in which to study mitral valve disease: The Living Heart Human Model (LHHM). The project to develop this massive computer simulation, founded on finite element analysis (FEA), has been ongoing for more than six years, thanks to contributions from software engineers, biomedical researchers, and clinicians around the world. Both academic and commercial licenses, from project founder Dassault Systèmes SIMULIA, provide access to the high-performance computer model of a beating heart, complete with mechanical, electrical, and fluid-flow characteristics and a variety of submodels of different areas within the organ. While Thornton Tomasetti has a commercial license to The Living Heart for proprietary IP development and private client projects, its Applied Science group remains an active contributor to the overall LHHM as well, notes Wang. “The heart is a very big bioengineering challenge so we all need to collaborate on The Living Heart Project to reach the final objective,” says Wang. “Our own team achieved an improved model of a closed mitral valve shape that we contributed to this year's software release.”

That improved model — one that is closer to medical imaging data of a normal human valve — was only a starting point for Wang's team. They went on to create a submodel of the MV region (see Figure 1) that focused on parameters of the aforementioned annulus and leaflets — and adjusted the chordae, which are the string-like tendons that run between the leaflets and the papillary muscles of the heart to control the opening and closing of the valve. For accuracy (since the valve functions inside a beating heart) this new model included boundary conditions and pressure differences as driven variables from the full LHHM.

Simulating an MI

Fig. 2 - Simulating the effects of a heart attack.

Now it was time to create the effects of a heart attack on the model. As seen in Figure 2, the yellow region of the global model of the heart (left) depicts tissue that has died after a heart attack stops blood flow to that part of the organ. In a cross section of the heart (middle), the yellow damaged region is visible inside the left ventricle. At right is the mitral valve (beige) with some chordae not affected by the MI (blue) and others in dark yellow that are now attached to the dead portion of the heart muscle.

“Simulating an MI is a challenging task, as both passive and active material properties of the infarcted region change immediately after the heart attack and over the course of healing — and there are likely changes to the electrical behavior of the heart as well,” says Wang. While the LHHM has the potential to take into account all these effects, in this study the team was able to simulate an FMR-disease state by modifying just the active material properties of their submodel. “That was the most relevant parameter for our purposes,” she says. “The MI causes asymmetric behavior of the mitral valve because some chordae remain attached to healthy heart muscle while others do not. As a result, the leaflets no longer coapt completely, and blood regurgitates back into the atrium.”

With their post-MI, disease-state valve model in hand, the group then shifted focus to treatment of the mitral valve malfunction with an annuloplasty ring implant. This is where 3D modeling really shines in life-science research: a virtual model of a ring could be inserted into the virtual model of the diseased mitral valve to see how well it worked.

Evaluating a Medical Device for Optimum Treatment Results

Fig. 3 - Model of annuloplasty ring (left) with subvalvular arm. Model of ring and arm installed in a mitral valve (right). Note how the arm contacts some of the chordae below the valve.

The engineers chose a model of a previously designed annuloplasty ring (see Figure 3) with a subvalvular component (an arm that extends below the valve and touches some of the chordae), adjusting it slightly to accommodate the dimensions of their valve model. They then “sutured” the ring model to the valve model, ran the simulation, and compared the closed dimensions of the valve before and after the ring was implanted. A quick visual inspection showed that the installation of the annuloplasty ring and arm resulted in improved coaption. In a separate workflow, simulations of blood flow through the annulus (using smooth particle hydrodynamics, SPH) were used to verify that leakage was significantly reduced in the final design.

Aware of the long-term failure history of annuloplasty rings in those 60 percent of patients, Wang's group next wanted to explore what geometrical changes could be made to the subvalvular arm of the ring to improve its overall effectiveness. Three parameters were chosen to be optimized: arm height, arm length, and the radius of curvature of the arm. Powered by Isight software, a design of experiments (DOE) was performed to understand the effects of different design variables on the response of the valve. By optimizing the interactions of the three parameters, the engineers sought to find out which geometric configuration would minimize the amount of blood regurgitating back through the annulus (see Figure 4).

Fig. 4 - Left: Model of mitral-valve closure shape in a diseased heart with MI — note the small opening where the leaflets have not properly coapted, allowing regurgitation of blood — and after an annuloplasty ring was installed (the ring is not shown in this image). Right: SPH simulation of blood flow in malfunctioning valve revealing leakage and the final results when the optimum ring design was achieved.

The data showed that leakage increased as the values for the height and radius of the arm increased, but decreased as the length increased. The design variable with the strongest effect overall turned out to be changing the length of the arm, something Wang's team thought was likely due to different chordae being engaged at different arm lengths. “This indicated the importance of identifying the chordae that need to either be included or excluded in the engagement with the ring in order to minimize FMR,” says Wang. Optimization with Isight enabled the engineers to identify the annuloplasty ring implant that best fit the design space and reduced blood leakage the most.

While this project is a just snapshot in time of Wang's work, she feels her team's progress has demonstrated what a powerful platform the LHHM can be for medical device design and evaluation. “We have developed the capability to adequately model the effects of an MI on the mitral valve, FMR in particular, and then optimize device design to mitigate those effects,” she says. “The automated methodology we've developed provides a firm foundation for future work in simulating patient-specific MV cases. It also has the potential to expand in scope to include other parts of the human heart.”

This article was written by Lynn Manning, a science and technology writer based in Providence, RI. For more information, click here .