When it comes to prostate-specific antigen (PSA) testing, researchers at the University of California, San Francisco (UCSF) discovered that factors in an individual's genetic makeup can skew the results, contributing to the somewhat unreliable predictions for prostate cancer observed in recent years with this controversial biomarker.
By correcting for these genetic factors, the UCSF team found that about 20 percent of biopsies for non-prostate cancer cases in two previous clinical trials would have been avoided in men referred for this invasive procedure based on standard PSA findings.
"Our findings are exciting because we were able to show that we can use genetic discoveries that are coming out of genome-wide association studies to actually improve the detection of prostate cancer and hopefully try to make PSA a more useful and accurate screening biomarker," said Linda Kachuri, MPH, PhD, a postdoctoral fellow in the Department of Epidemiology & Biostatistics at UCSF.
Presented during the 2022 American Association for Cancer Research (AACR) Annual Meeting, held April 8-13 in New Orleans, the research results stem from what is believed to be the largest ever genome-wide association study (GWAS) of PSA that included 95,000 men without diagnosed prostate cancer, using data from five cohorts from the United States, United Kingdom, and Sweden.
"I think that tools that give us more precision typically will turn out to have value," said Louis M. Weiner, MD, Director of the Georgetown Lombardi Comprehensive Cancer Center, who moderated an AACR press conference on this study.
"And if this [genetically adjusted PSA] is validated, as it seems to be getting validated right now, it could create a new paradigm for giving clinicians useful, actionable information to inform their patients," he added.
Prostate cancer is the second most common cancer worldwide, with about 1.25 million cases diagnosed globally in 2018. Until about 2008, doctors and professional organizations encouraged yearly PSA screening for men beginning at age 50, with age 45 for those considered at higher risk. However, subsequent studies revealed low sensitivity and specificity for PSA, leading to overdiagnosis in patients with otherwise benign prostate conditions and underdiagnosis of patients with more aggressive disease.
"Although [PSA] is a biomarker that has tremendous potential, right now it's not really being used systematically for screening," Kachuri said.
Given the controversy over its use, researchers around the world have been searching for an inexpensive, minimally invasive, and more effective tool to replace what was once considered the gold standard biomarker for the detection and management of prostate cancer. Several of these are in various stages of testing. Others have been trying to understand why PSA tests have proven somewhat unreliable, with a goal of improving its accuracy.
Among other things, studies have shown that PSA levels in the general population may reflect factors other than prostate cancer, including both benign and malignant conditions due to increasing age, ethnicity and, more importantly, genetic factors.
"One way you can think about this is that the observed PSA value is a bit noisy," said Kachuri. "It can reflect variations in PSA due to cancer, but it can also capture other factors that influence PSA. One of these factors is genetics."
In essence, the UCSF team is trying to reduce the "noisy" background inherent in the standard PSA test by taking into consideration, or adjusting for, an individual's genetic makeup.
As a first step in this study, the UCSF researchers conducted a genome-wide scan to identify genetic variants that were predictive of PSA levels in a large cohort of 95,000 men not previously diagnosed with prostate cancer. The goal was to find those genetic predictors of PSA elevation in men who do not have prostate cancer.
The analysis identified 129 PSA-associated different variants, of which 82 were novel. Using this data, the team created a polygenic score for PSA levels that provided a measure of each individual's genetic risk for high PSA levels. Essentially, the polygenic score is a cumulative measure of someone's genetic predisposition for high PSA, calculated by adding up the variants for each individual.
For the next step, the researchers sought to validate the predictive value of these genetic variants. "Every person has a particular value for this genetic score which captures their unique genetics," Kachuri said. "What we'd like to do is account for that information in our PSA values."
Using data from two clinical prostate cancer trials, the team showed that their genetic score explained 7.3 percent and 8.8 percent of the variation in baseline PSA levels in the two cohorts, PCPT and SELECT. Importantly, the score was not associated with prostate cancer, confirming that it reflects benign PSA variation.
Finally, the team wanted to see how their adjusted PSA scores may have altered the diagnosis and treatment in men from the previous clinical trials for prostate cancer.
"All of these individuals were biopsied, and so we were then able to see, had we gone back in time and used their genetically corrected PSA instead of their observed PSA, how many people would still have been referred to biopsy," Kachuri said.
Using their genetically adjusted PSA would have avoided 19.6 percent biopsies for men who were referred for biopsies which turned out to be negative based on their PSA level. It also would have resulted in 15.7 percent fewer biopsies in cases with low-grade disease, which accounted for 71 percent of all patients who would have avoided a biopsy. Further, use of the genetically adjusted PSA score in both the PCPT and SELECT cohorts was more robustly associated with aggressive prostate cancer than observed PSA values.
"We achieved the best predictive performance for aggressive disease by combining the genetic risk score for prostate cancer that hasn't previously been developed and reported with the genetically adjusted PSA measures that we developed here," Kachuri added.
While the study was very large, almost 90 percent of the participants were of predominantly European ancestry, thus representing a key limitation.
"In our subsequent effort, we are really trying to focus on having larger and much more diverse studies so we can really comprehensively examine PSA genetics in individuals of all ancestries where they represent our target patient population," Kachuri said.
Warren Froelich is a contributing writer.