This article is more than 5 years old. Information may no longer be current.
Urine biomarker shows potential in detection of aggressive prostate cancer
Investigators at UCLA Jonsson Comprehensive Cancer Center and University of Toronto have identified a novel biomarker in urine that could aid in the detection of aggressive prostate cancer.
“We currently do not have accurate biomarkers to help determine the aggressiveness of prostate cancer that is clinically localized,” Paul C. Boutros, PhD, director of cancer data science at UCLA Jonsson Comprehensive Cancer Center, said in a press release. “We developed a three-stage experimental strategy to maximize statistical and data science considerations and give us the best chance of finding a biomarker that predicts prostate cancer aggressiveness.”
Boutros and colleagues sought to measure the longitudinal stability of 673 microRNAs in urine samples from 10 men with localized prostate cancer. They then measured temporally stable urine microRNAs among 99 patients in an independent training cohort and developed a biomarker predictive of Gleason grade. The biomarker was validated in an independent cohort of 40 patients.
Results showed the urine biomarker identified men at high risk for aggressive prostate cancer with accuracy similar to that of invasive tissue-based tests.
HemOnc Today spoke with Boutros about what prompted this research, what the data showed and plans for additional research.
Question: What prompted this research?
Answer: For men newly diagnosed with localized prostate cancer, the problem in general is not detection of the disease, but detection of aggressive disease. The existing PSA test and imaging with multiparametric MRI are effective at identifying prostate lesions. However, large numbers of men are diagnosed with lower-grade disease that is localized, relatively small in volume and unlikely to be lethal. For these patients, standard of care is active surveillance. This would include monitoring with different frequencies, but 6-month annual imaging or biopsy supplemented by PSA testing is typical. However, many men do not feel confident in this treatment decision; thus, a significant number elect to forgo active surveillance. The need here is to identify aggressive tumors so those men will move immediately to definitive local therapy and potentially to systemic therapy with androgen deprivation.
Q: How did you conduct the study?
A: We first considered ways to approach developing a test for these men. There are all sorts of options with tissue-based biomarkers, but those are not amenable to repeat monitoring. Also, biopsies are uncomfortable, morbid and have many side effects, as well as significant costs. We tried to identify something that would be minimally or noninvasive and that could be done regularly from a longitudinal perspective to maximize the chance of repeat use. This led us quickly to think about urine. Of course, the prostate is ideal for urine-based biomarkers because of its proximity — it is possible to find many prostate-secreted proteins and other molecules in urine. Given this decision, our first question was: Is it useful to come up with a biomarker if we find out the characteristics of people’s urine inherently vary over time? Based upon what someone eats, how they work out or any similar characteristic, the urine may contain different molecules.
Working with our colleagues in Toronto, we identified a set of men on active surveillance protocols who provided annual urine samples. We were then able to longitudinally look at how the urine of an individual man changed across several years. This allowed us to use a profiling technique that looked at a specific class of genes known as miRNAs. We could see which of those changed over time and which were quite stable within an individual patient. This enabled us to omit a large fraction of genes. We then focused on key molecules associated with disease.
We ran a very standard cohort design study that included a large number of men, and processed their urine to identify those associated with high-grade aggressive disease. Data from those men were used to identify a risk signature that could predict whether a patient had aggressive disease from the urine alone. We took the final risk signature and validated it in yet another cohort of men to develop an assessment of accuracy.
Q: What did you find?
A: It was surprising to see just how much individual variability there was in human urine over time. We almost have to obtain it up front to make these types of studies work, and this may yield insights into other liquid biopsy-based biomarkers. The second thing that we found was that if we examine the tumor, the urine is a good molecular surrogate. There are all sorts of things we cannot measure because of the longitudinal instability, but for those we could measure, there is a strong correlation between the levels in the urine and those in the primary tumor. This suggests that at least in prostate cancer and for miRNAs that are not physiologically variable, we can estimate what the tumor looks like just by looking at what is in the urine. We put all of this together into a risk signature and were able to validate the same sensitivity as a PSA test, but with more specificity. This test could reduce the numbers of false positives and men who might unnecessarily go to definitive therapy.
Q: Are there plans for subsequent research on this?
A: Absolutely. Our next step is to conduct a much larger validation study, for which we are working to miniaturize the assay and make it easier to execute. We then intend to analyze about 1,000 men with a little more clinical variability in their initial characteristics to quantify the generalizability of this biomarker. This will give a better estimate of its accuracy and put it in the context of multiparametric MRI assays to see if the test is identifying tumors that would have been missed by radiology.
Q: What is the anticipated timeline for results to be published from this subsequent research?
A: We are actively conducting the research now and results hopefully will become available during the next year or 2.
Q: Is there anything else that you would like to mention?
A: This work requires high-quality biobanking and deep interactions with clinician scientists. In this case, my partner in crime was Stanley Liu, MD, radiation oncologist at Sunnybrook Health Sciences Center in Toronto, who is deeply engaged in biobanking. This type of clinician-scientist partnership is key in studies such as this. – by Jennifer Southall
Reference:
Jeon J, et al. J Natl Cancer Inst. 2019;doi: 10.1093/jnci/djz112.
For more information:
Paul C. Boutros , PhD, can be reached at UCLA, Department of Human Genetics, Box 957088, 57200A South Tower CHS, Los Angeles, CA 90095; email: pboutros@mednet.ucla.edu.
Disclosure: Boutros reports no relevant financial disclosures.