RESEARCH NEWS

Parallel computing speeds-up cancer diagnosis

17 July 2007



Biomedical engineers at the University of Virginia School of Engineering and Applied Science have developed a new imaging tool that will dramatically improve medical ultrasound scans, potentially leading to quicker and more accurate diagnosis of breast cancer and other life threatening conditions.

Using Star-P software from Interactive Supercomputing, the university’s biomedical engineering research team, led by Associate Professor William F. Walker, created an advanced beam-forming algorithm – called the time-domain optimised near-field estimator (TONE) – which significantly improves the contrast and resolution of ultrasound images.

‘The potential applications for this algorithm are almost infinite,’ said James H. Aylor, dean of University of Virginia’s School of Engineering and Applied Science. ‘Not only can it be used in the medical community to benefit patients nationwide, it will have applications in the fields of radio astronomy, seismology and more.’

While conventional beam-forming algorithms have been used in ultrasound scanners for nearly a half century, they typically result in degraded images that are blurry or cluttered. This is caused by sound wave reflections coming from undesired locations within the organ or tissue.

The TONE algorithm reduces these signals, resulting in much higher definition images. However, it comes at the heavy price of a much greater computational load, which would overwhelm a typical computer’s processing ability. The team solved this problem by automatically parallelising their algorithms with Star-P to run on a powerful, memory-rich IBM Linux cluster.

Research associate Francesco Viola said: ‘It takes a huge amount of memory and computational resources to execute the algorithm. Typical resolution for ultrasound imaging systems is in the 200-300μm range. With Star-P, we were able to tap into the University’s supercomputing clusters to generate ultra high resolution images of 67μm, without having to become parallel programming experts.’