This Gene Test Might Guide PSA Screening: Two Views

Nick Mulcahy

January 10, 2018

A one-time gene-based test can be administered "long before the onset of prostate cancer" and "substantially inform the decision" of whether or not a patient should undergo prostate cancer screening with a prostate-specific antigen (PSA) test, say an international team of investigators reporting a new study.

The study was published online January 10 in the BMJ.

The test calculates a "polygenic hazard score" that represents a man's fixed genetic risk for prostate cancer, including aggressive disease, say the authors.

"This is personalized prostate cancer screening ― personalized by the patient's own genetics," lead author Tyler Seibert, MD, PhD, of the Center for Multimodal Imaging and Genetics, University of California, San Diego, told Medscape Medical News.

"We are suggesting that this score can be used to help decide which patients get PSA testing," he added.

Andrew Vickers, PhD, from the Department of Epidemiology and Biostatistics at the Memorial Sloan Kettering Cancer Center in New York City, who was asked for comment, said that the genetically based score is "intriguing and promising."

But the score is currently not better than a single midlife PSA test score in being prognostic of later metastatic and lethal prostate cancer, said Dr Vickers, who has led the research that promotes the idea of such a single midlife PSA test to determine subsequent prostate cancer risk.

"In terms of clinical practice, at the moment, this has got nothing on PSA in terms of the ability to predict the long-term risk of aggressive prostate cancer," Dr Vickers told Medscape Medical News.

Overall, PSA testing is not very good for detecting prostate cancer, he said. "But PSA is a very good predictor of metastatic or lethal prostate cancer."

He was referring to two studies, both conducted by his team, that show that a single PSA measurement obtained when a man is in his 40s strongly predicts his risk for either advanced-stage or lethal prostate cancer later in life.

Dr Seibert emphasized the potential of the polygenic hazard score in the clinic.

"The score has clinical utility for primary care physicians as long as its prediction is accurate. We have already shown remarkably high accuracy in an independent dataset [from the ProtecT study], so I think we are most of the way there," he offered.

I think we are most of the way there. Dr Tyler Seibert

"That said, it is always important to replicate results, and we intend to do that," Dr Seibert added.

Currently, there are no similar genotype tests in routine clinical use for predicting risk for any cancer. But Dr Seibert expects that to change "in coming years," because research is being conducted in multiple settings: "There is a lot of excitement about this area of research in, for example, breast cancer, lung cancer, and colorectal cancer."

Genotyping is also becoming more common outside of medicine, with "millions of people already having been genotyped at their own expense to learn about their ancestry or risks of other diseases," he added.

Effect Size Issues

The new study has two data sets. In the development set, the researchers created a preliminary gene-based hazard score model out of genotype- and age-related data from 21 studies in the PRACTICAL consortium.

They analyzed more than 200,000 gene variants (ie, single nucleotide polymorphisms) from 31,747 white men with and without prostate cancer and identified 54 gene variants that were associated with increased risk for prostate cancer.

The gene variant data were folded into a survival analysis to estimate their effects on age at diagnosis of aggressive prostate cancer in the form of a hazard score.

A second and final model was then applied to data from the above-mentioned ProtecT clinical trial, which was conducted in 6411 men with prostate cancer who were randomly allocated to receive treatment or to undergo surveillance. This data set was used to validate the predictions of prostate cancer–free survival.

In this independent validation, "the hazard score was a highly significant predictor of age at diagnosis of aggressive cancer," write the study authors.

That conclusion is based on the investigators' comparing men with "high" hazard scores (>98th percentile) with men with "average" scores (30th to 70th percentile). In that comparison, they found that the hazard ratio for aggressive cancer was 2.9 among the high-score men (95% confidence interval, 2.4 - 3.4).

In other words, men whose scores were in the top 2% had an almost threefold greater relative risk for aggressive prostate cancer compared with men in middle percentiles, which accounted for 40% of the scores.

That is not typical methodology in epidemiology, observed Dr Vickers.

The "big effect" — a hazard ratio of nearly 3 — that the researchers report comes about by "focusing on this very small group — those at highest risk," he said.

"If they had done something more conventional, like comparing the top and bottom quintile, you wouldn't have seen quite that large effect," Dr Vickers said.

But Dr Seibert said that comparing the top 20% to the lowest 20% should give a greater effect, not lesser, because "we would be comparing the extremes."

Queried by Medscape Medical News about the issue, he ran such a calculation and found that the hazard ratio for the prediction of aggressive disease for top quintile vs bottom quintile of the score was 3.5, up from 2.9, as reported in the article.

The study investigators also included family history, which is considered a risk factor for prostate cancer, in a combined model, but this did not improve prediction of onset of aggressive prostate cancer (P = .59). In their various analyses, they found that "polygenic hazard score performance remained high when family history was accounted for."

The team also assessed the positive predictive value (PPV) of PSA testing within the validation set.

"Positive predictive value is directly dependent on prevalence, so if the polygenic hazard score predicts age of onset of prostate cancer, the positive predictive value of PSA should vary with the score," they write.

This held true in the validation set, they report.

In other words, the PPV of PSA screening for aggressive prostate cancer was increased with increasing polygenic hazard scores, they say.

More on the Comparison With PSA at Midlife

In one study by Dr Vickers and colleagues, men whose PSA value was higher than 1.5 ng/mL between the ages of 45 and 49 years accounted for nearly half of the prostate cancer deaths in a Swedish cohort over the next 30 years or so. Only 10% of the men in the study had such high PSA values at this relatively young age, as reported by Medscape Medical News.

Dr Vickers said that the effect size reported with the polygenic hazard score is "much smaller" than that found in midlife PSA studies.

Dr Seibert and the study authors acknowledge, in their new article's discussion section, that the midlife PSA score has the "best supportive evidence" of any current risk-stratification metric. They also state that their new score might be eventually be used in combination with the midlife PSA concentration.

Other studies have used genotyping to predict risk for prostate cancer, the current study authors also point out. But those studies are case-control studies. Epidemiologic studies, such as the current study, are preferable, Dr Seibert and colleagues say. That's because, in part, they can demonstrate that prostate cancer development is "highly" dependent on increasing age and is not a simple dichotomy, as in case-control studies.

Study limitations include the possibility that other, unmeasured factors may have influenced their results.

Also, the current study is limited to white men. "Validation of the score in other ethnic groups — and, if necessary, custom models for each — is needed. We plan to investigate this important question," the authors state.

This study was funded in part by grants from the US Department of Defense, the Prostate Cancer Foundation, the Research Council of Norway, KG Jebsen Stiftelsen, and the South East Norway Health Authority. Dr Seifert has disclosed no relevant financial relationships. Other coauthors have a patent pending on the test, which is also tied to the University of California, San Diego. Dr Vickers is named on a patent application for a statistical method to detect prostate cancer, which has been commercialized by OPKO Health Inc. He also receives royalties from sales of the test and has stock options in OPKO Health Inc.

BMJ. Published online January 10, 2018. Full text

Follow Medscape senior journalist Nick Mulcahy on Twitter: @MulcahyNick

For more from Medscape Oncology, follow us on Twitter: @MedscapeOnc


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.
Post as: