Skip to main content
Premium Trial:

Request an Annual Quote

Noninvasive Prostate Cancer MDx Test Enters Validation Phase After Initial Results Published in JCO


NEW YORK (GenomeWeb) – Publishing a study last month in the Journal of Clinical Oncology's Precision Oncology edition, members of an international team have laid the groundwork for a new epigenetic assay to help clinicians assess the risk of prostate cancer using a urine sample — an advancement that could help patients avoid unnecessary biopsies and other follow-up procedures.

The study, which appeared online Jan. 14, described the team's Epicapture (Epigenetic Cancer of the Prostate Test in Urine) assay, which detects aberrant methylation in six prostate cancer-associated genes (GSTP1, SFRP2, IGFBP3, IGFBP7, APC, and PTGS2) via PCR.

Investigators reported that the test achieved area under the receiver operating characteristic curves of 0.64, 0.86, and 0.83 for detecting all prostate cancer, only high-grade cancers, and high-risk cancers, respectively. Although the sensitivity for overall cancer doesn't improve much on currently used prostate specific antigen tests, the performance for these different clinical categorizations of high-risk tumors was superior to PSA.

Antoinette Perry, assistant professor in genetics and cell biology at University College Dublin, and a co-author of the study in JCO, said in an interview that she and her colleagues are now working on clinically validating the test to prepare for potential clinical use, and they are looking at ways to make the assay more easy to commercialize and more applicable to wide-scale prostate cancer screening.

Currently, the analysis requires urine samples taken after a digital rectal examination, which triggers the release of prostate material into the urine. A less involved sampling process might make the test more easily translatable to the clinic, especially in the context of screening. The group also now has a second phase of the project funded by Enterprise Ireland.

"We're looking to clinically validate these results in another large, independent cohort of patients, to show that we can reproduce these findings and maybe look at added clinical data from the patients … maybe building in other features such as the age of the patients … but basically looking now to prove what we've just shown that we can we can reproduce it again," she said.

According to Perry, the creation and validation of the Epicapture signature was catalyzed by funding from the Movember Foundation, which connected Perry with collaborators at other European and US Institutions to pool resources and samples.

"They had a number of GAP studies, which are global action plans, and the first was focused on prostate cancer biomarkers that could distinguish aggressive disease from clinically indolent disease," Perry said. In addition to urine, Movember GAP projects were also funded to look at things like blood and exosomes, and results from those efforts are also being finalized and published, she added.

In their report, Perry and her coauthors cited as an impetus for their efforts the observation that widespread use of prostate specific antigen (PSA) testing has significantly increased the detection and thus the overtreatment of low-risk cancers with little likelihood of clinical manifestation.

Although PSA accuracy has been improved by the development of algorithms incorporating different protein isoforms, dissatisfaction with the biomarker continues to translate into demand for prognostic alternatives, something that the international team behind Epicapture is not alone in trying to fill.

Some companies already offer commercial tests with a similar clinical function as Epicapture would provide, and others have said they are developing urine- or blood-based risk assays, and reported performance from various other commercial and academic efforts will set a benchmark for the Epicapture team as they move forward. MDx Health, for example, has offered SelectMDx, a urine test for prostate cancer risk stratification, since 2016.

The firm has not yet penetrated a majority of the projected market, but it presented a variety of data last year intended to help persuade the community of the utility of the test, which it reported has an area under the receiver operating curve characteristic of 0.89.

MDx Health President and CEO Jan Groen said this week that the company recently reported SelectMDx testing grew 58 percent in 2018 compared to 2017. "Since the launch of the [test] slightly over two years ago, we have tested over 30,000 patients," he wrote in an email.

Exosome Diagnostics also sells a urine-based assay, the ExoDx Prostate IntelliScore (EPI) prostate cancer test, which tests for three exosomal RNA biomarkers, and has been engaged in its own efforts to demonstrate clinical utility and drive health plan coverage.

Academic teams have published on a variety of prognostics that improve on PSA, like the University of Michigan's Mi Prostate Score, which incorporates detection of the TMPRSS2-ERG fusion and PCA3 transcripts in urine with serum PSA and delivers AUCs up to 0.78 for detecting high-grade cancers.

And other commercial firms, like MDNA and Chronix Biomedical, have made moves over the last few years to enter the space with strategies based on analyzing alterations in mitochondrial DNA or sequencing circulating tumor mutations, though neither firm currently lists a non-invasive prostate cancer test on its website.

Groen said that MDx Health believes that the 98 percent negative predictive value it has reported for SelectMDx remains industry-leading.

In their study, Perry and Colleagues began by analyzing their six-gene methylation signature in archived tissue samples using Illumina microarrays, in order to confirm that the methylation of the genes in question was indeed associated with prostate cancer and that this epigenetic alteration increased with aggressiveness.

Based on positive results from this effort, the group then translated the signature to a quantitative methylation-specific PCR assay, which they applied to urine samples from a multi-center cohort of about 450 men scheduled for ultrasound-guided biopsy prompted by elevated and/or increasing PSA or a suspicious rectal examination.

The group created logistic regression models using a training set of 319 of the total cohort members (70 percent) and then evaluated these models for their ability to correctly predict either overall cancer, high-grade cancer, or high-risk cancer in the remaining 30 percent.

"Analyzing all six genes together, [the test] consistently performed better than individual genes or combinations thereof," Perry and her coauthors wrote. One important aspect of the study design, she added, was the team's decision to assess the assay performance not just in detecting prostate cancer overall, but also in predicting high-risk and high-grade tumors as defined by clinical definition tools that are used widely by urologists.

"Gleason score is a good prognostic measure … but there are other risk stratification tools that are used as well clinically … the ones that we used in the study being D'Amico or CAPRA," Perry said. "What those tools do is they encompass the Gleason score, but they also bring in other features as well. So, the D'Amico uses Gleason, it uses PSA, and then it uses the clinical stage of the disease. And the CAPRA does the same thing, but it also gives a weight depending on the age of the patient and what percentage of the biopsy cores were positive.

"They're just a little bit more informative or holistic than just looking at Gleason score, and … we thought well let's see how well [our test] does not only just using Gleason but using these risk stratification tools that are used clinically," she added.

For prostate cancer in general, the Epicapture test's AUC was only 0.64, which isn't a real improvement over PSA. But for detecting high-grade cancers as defined by Gleason score, the D'Amico tool, and CAPRA stratification, it rose to 0.86, 0.83, and 0.80, respectively.

That translated to sensitivities of 0.85, 0.70, and 0.67 and NPV greater than 0.94 for all three prognostic end points, the authors wrote.

Finally, the team evaluated the training set portion of the cohort to select an Epicapture score cutoff that would best inform a physician's decision on whether to perform a prostate biopsy.

"Applying [the resulting threshold] delivered a 79 percent improvement over the tumor specificity of PSA," the authors wrote.

Overall, the group concluded, Epicapture represents a "potential improvement on existing methods for patient stratification and for determining the need to perform biopsy, detecting 85 percent of high-grade tumors, with a [negative predictive value] of 98 percent."

Additionally, because the cohort studied was young and biopsy was the only endpoint used to assess the test, it's also possible that the false-positive rate could be pushed lower, considering the possibility that some Epicapture-positive individuals who had negative or low-grade biopsy results actually did have aggressive cancers missed by the biopsy itself, the authors wrote.

This was illustrated by one case in the current study, Patient SJH149, in which biopsy yielded a relatively low-risk result: "a minute focus of Gleason score 6 adenocarcinoma," despite an Epicapture score that was "strongly positive."  A repeat biopsy then revealed high-grade cancer in 40 percent of cores on the right side.

"This finding could suggest that the source of the tumor DNA analyzed in urine could have been the high-grade tumor, which was not sampled on the original biopsy," the authors wrote.

For the false-negatives in the study, Perry and her colleagues wrote that the analysis suggests that the assay failed to detect aberrant methylation because these cases had insufficient or undetectable prostate cells present in their urine, whether due to biological differences or to variation in how samples were collected. According to the authors, future work may allow the incorporation of a prostate-specific marker into the assay, which would flag samples with insufficient prostate cell content for analysis.