Model of detection for prostate tumours

A two-year project could result in a system to help doctors detect prostate cancer and make earlier, life-saving diagnoses.

Visualisation specialist See3D, an Aberystwyth University spin-off, will develop computer-generated models that will be able to identify smaller cancerous regions which are currently difficult to detect.

'At the moment radiologists will just look at MR (magnetic resonance) images. We hope to provide computer-assisted evaluation of those images, so we will add a computer to determine where to look for certain images or for certain abnormalities within the data,' said Dr Reyer Zwiggelaar of Oncomorph Analysis, the medical imaging software company that commissioned the project.

'We've shown with other applications [breast and lung cancers] that radiologists will be able to pick up abnormalities that are smaller — and they will also be able to pick them up earlier.'

The researchers have carried out studies using the technology for other applications which they claim reveals the system to be able to give a fair second opinion on a scan reading because it is 'as good as a radiologist'.

'In other applications there was an improvement of between five and 25 per cent in the performance of radiologists' ability to detect things early,' said Zwiggelaar.

While the new technology will be based on similar existing software, one main difference is that it will process MR-based images rather than X-ray, as is the case in mammography.

The team hopes to develop a system that can either be provided direct to radiologists or incorporated into MR-based systems.