APPLICATIONS NEWS

HPC helps doctors detect breast cancer in early stages

9 October 2007



High-performance computing is helping doctors to detect breast cancer in the very early stages of development, leading to a greater chance of successful treatment and recovery. The computers perform computationally intensive algorithms to analyse images from a new technique that works with x-ray mammography to provide a better diagnosis.

The technique, currently being investigated by Dr Susan Hagness at the University of Wisconsin-Madison, fires ultra-short pulses of low-power electromagnetic waves at the patient, which provides three-dimensional images of the breasts. This process is completely harmless, but gives more information than an x-ray image alone.

However, to retrieve this useful information, highly complex image analysis is necessary, including inverse scattering algorithms which take up a lot of processing power. These processes would take too long on conventional computers, so Hagness chose an HPC solution from Acceleware to increase the speed of the computations.

'Acceleware's hardware solution enables us to complete our computations and generate images in hours instead of days,' said Hagness. 'Minimising the time required to generate complex images is fundamental to helping save lives and reduce suffering.'

Related internet links

Acceleware
University of Wisconsin-Madison