Some cancers, like the exceptionally deadly pancreatic, make the diagnosis simple: purge it as soon as possible! Pancreatic cancer is commonly undiagnosed until its late stages, at which point it's “all hands on deck.” Prostate cancer, on the other, poses problems on the other end of the spectrum: it is usually caught in such an early stage that “watchful waiting,” rather than any kind of surgery or radiation therapy, is a legitimate and often favored approach. In fact, more than 50 per cent of prostate cancer cases cause no symptoms and are never life threatening.

Urologists long for better clinical markers – diagnostic “road signs” – that will help them predict the behavior of prostate cancer tumors. Or as Colin Cooper, professor of cancer genetics at the University of East Anglia's Norwich Medical School,puts it, “...distinguishing the dangerous 'tigers' from the less threatening 'pussycats.'”

Cooper and his team are trying to improve urologists' diagnostic lot through the application of mathematics. They applied a method to cancer samples which they call “Latent Process Decomposition” (LPD). It evaluates the structure of a dataset in the absence of knowledge of clinical outcome. Gene expression levels in each cancer sample are then modeled. This allowed the researchers to identify a common process in the data they studied which was unique to clinically important cancer. In the past, such mathematical modeling has failed because prostate cancer samples are so diverse; the LPD process allows doctors to at last identify the common denominator.

The touchstone is "...categorized by low expression of a core set of 45 genes, many encoding proteins involved in the structure of cells, transport of ions and cell adhesion. This was common across the samples of cancers known to have a poor patient prognosis," according to Professor Vincent Moulton, from UEA's School of Computer Science.

The end result? Doctors now have another diagnostic tool to assist them in prescribing the right prostate cancer therapy at the right time, and reduce the risk of over-treatment.

The research has been published in the journal European Urology Focus.