In her laboratory at the University of Washington, mathematician Kristin Rae Swanson peers into the future of brain cancer patients--on her computer screen. She has created a software program that uses data from magnetic resonance imaging scans to simulate how fast a patient's brain tumor is likely to spread. She can pinpoint with uncanny precision where a tumor will grow months ahead of time and predict how long a patient is likely to live under various treatment scenarios.
She first proposed the idea ten years ago and "was laughed out of the room" by skeptical doctors who figured brain tumor growth was too erratic to predict, Swanson says. But she has developed an equation that takes into account how fast tumors divide and disperse through brain tissue and can predict their path. It takes three hours on a PC to run a patient's data through this simulation.