Urine test predicts bladder cancer risk 12 years prior to symptom onset

A simple urine test looking at genetic mutations in DNA predicts urothelial cancer – the most common type of bladder cancer – 12 years before patients develop symptoms in a recent study.

The urinary comprehensive genomic profiling (uCGP) test has the potential to be used as a screening tool for people at increased risk for bladder cancer by reason of genetics, smoking, or environmental exposures to known carcinogens, said the researchers. More importantly, it could reduce the frequency of unwanted cystoscopies.

The uCGP test used an NGS assay and was able to identify mutations in 60 genes associated with bladder cancer, details of which were reported at EAU23. [Abstract A0268]

Detection made easy

In training and validation cohorts, the test accurately predicted future bladder cancer in 66 percent of urine samples, including those collected more than 10 years prior to the testing.

“These findings provide the first evidence from a population-based cohort study of preclinical urothelial cancer detection with uCGP,” said Dr Florence Le Calvez-Kelm from the International Agency for Research on Cancer (IARC) in Lyon, France. “Results were similar for patients with known risk factors for bladder cancer and those without.”

Commenting on the study, Dr Joost Boormans, a urologist at the Erasmus University Medical Center in Rotterdam, Netherlands said the study was very “encouraging as it shows that our ability to identify molecular alterations in liquid biopsies, urine in this case which can indicate cancer,  is constantly improving.”

“This urine test for genetic mutations could show its value is in reducing cystoscopies and scans in bladder cancer patients who are being monitored for recurrence, including those referred for the presence of blood in the urine,” said Boormans.

“A simple urine test would be far easier for patients to undergo than invasive procedures or scans and could be less costly for health services."

Le Calvez-Kelm and her team first used the 60-gene assay to create a training set using urine samples from 46 patients with de novo urothelial cancer, 40 with recurrent cancer, and 140 healthy controls.

Next, they tested the model in two validation cohorts – the first were samples taken from 22 patients with de novo cancer, 48 with recurrent urothelial cancer, and 96 controls from a case-control study conducted at Massachusetts General Hospital (MGH) in Boston, Massachusetts and Ohio State University (OSU) in Columbus, Ohio, both in the US; the second included 29 patients from the prospective Golestan Cohort Study who subsequently developed urothelial cancer and 98 controls.

Ten genes were identified for inclusion in the screening model, which was targeted to an overall sensitivity of 88 percent and a 97 percent sensitivity for high-grade tumours, with a specificity of 94 percent.

In the MGH/OSU cohort, the sensitivity of the models was 71 percent, and the specificity was 94 percent.  In the Golestan cohort, the sensitivity was 66 percent, and the specificity was 94 percent.

Higher sensitivity closer to diagnosis

Interestingly, when the analysis was broken down according to the lag time between urine collection and diagnosis, sensitivity increased as the time to diagnosis decreased. “The closer we got to the diagnosis, the higher was the sensitivity,” Le Calvez-Kelm reported.

When the analysis was limited to urothelial cancers diagnosed within 7 years of sample collection, the sensitivity for detecting preclinical cancer improved to 86 percent.

In the Golestan cohort, uCGP-predicted positive results were associated with a more than eightfold higher risk for worse cancer-free survival compared with uCGP-predicted negatives (hazard ratio 8.5, p<0.0001).

More studies are warranted to determine the clinical utility of the findings in bigger and longer studies.