The mathematical model allows the researchers to change the parameters of the program, such as the stent configurations, the materials, the shape of the blood-vessel wall, and the drug-flow properties, so that they can test different experimental conditions. "We created an automatic algorithm so that we have the flexibility to visualize different elements without having to start from scratch," says Vijaya Kolachalama, a postdoctoral associate at HST working on the program. (There are currently only four drug-eluting stent designs that are approved by the U.S. Food and Drug Administration.)

A simulation yielding optimal drug-releasing properties could let researchers know what drug to discharge and in what fashion, says Park. "It's a model for building next-generation stents."

The model has been "eye opening," says Edelman. "It is surprising how often the drugs don't penetrate or deposit where expected." He compares the blood flow across a stent to white-water rapids flowing over a rock: some of the water strikes the base, flies up in the air, and comes back down, instead of flowing over the rock. So the water continuously recirculates in the same area, making the design of a stent "critically important."

Parts of the computational findings have been validated by animal and in vitro models. The work was recently published in the Journal of Controlled Release and is funded by the National Institute of Health.

As stents become more sophisticated, Edelman says, there is a huge gap between how researchers think the devices work and how they actually behave in animal models and humans. Trials in live subjects can be costly and time consuming, and in some cases they even result in death.

The MIT researchers are pushing to make computer modeling part of the FDA regulatory process and have made their algorithms available to others so that the software can evolve.