What is it about?
We developed the Prostate Urine Risk (PUR) test using machine learning to look at gene expression in urine from samples collected from 537 men attending a clinic with suspected prostate cancer. By examining the cell-free expression of 167 genes in urine samples, we found a mathematical combination of 35 different genes that could be used to produce the PUR risk signatures. Previous urine biomarker tests have been designed specifically for single purposes such as the detection of prostate cancer on re-biopsy (PCA3 test). But this new test uses four PUR signatures to provide a simultaneous assessment of non-cancerous tissue and risk groups (low, intermediate and high-risk) to show how aggressive the cancer is. This research shows that our urine test could be used to not only diagnose prostate cancer without the need for an invasive needle biopsy but to identify a patient’s level of risk. This means that we could predict whether or not prostate cancer patients already on active surveillance would require treatment. The really exciting thing is that the test predicted disease progression up to five years before it was detected by standard clinical methods. Furthermore, the test was able to identify men that were up to eight times less likely to need treatment within five years of diagnosis. If this test was to be used in the clinic, large numbers of men could avoid an unnecessary initial biopsy and the repeated, invasive follow-up of men with low-risk disease could be drastically reduced.
Photo by Erik-Jan Leusink on Unsplash
Why is it important?
Prostate cancer is more commonly a disease men die with rather than from. Unfortunately, we currently lack the ability to tell which men diagnosed with prostate cancer will need radical treatment and which men will not. Current practice assesses a patient’s disease using a PSA blood test, prostate biopsy and MRI. But up to 75 per cent of men with a raised PSA level are negative for prostate cancer on biopsy. Meanwhile 15 percent of patients who do not have a raised PSA are found to have prostate cancer – with a further 15 percent of these cancers being aggressive. A policy of ‘active surveillance’ has been developed as a way to combat this uncertainty, but it requires invasive follow-ups and constant reminders that a patient has a cancer with an uncertain natural history. This results in up to 50 per cent of men on active surveillance self-electing for treatment - whether they need it or not. It’s clear that there is a considerable need for additional, more accurate, tests for determining whether a patient has prostate cancer, how aggressive it is, and when treatment becomes necessary. If this test was to be used in the clinic, large numbers of men could avoid an unnecessary initial biopsy and the repeated, invasive follow-up of men with low-risk disease could be drastically reduced. It would also give additional reassurance that a patient does not need the disease treated.
Read the Original
This page is a summary of: A four-group urine risk classifier for predicting outcomes in patients with prostate cancer, BJU International, June 2019, Wiley, DOI: 10.1111/bju.14811.
You can read the full text:
The following have contributed to this page