Biopsy Decisions Guided by New Predictor

Pam Harrison

March 22, 2018

COPENHAGEN — A new tactic can help physicians decide whether or not to refer patients with prostate-specific antigen (PSA) levels high enough to suggest clinically significant prostate cancer, new research shows.

"There is no real strategy advising general practitioners how to refer patients to a specialist for a biopsy," said Jan Verbeek, MD, from the Erasmus Medical Centre in Rotterdam, the Netherlands.

"Some physicians will refer patients to a specialist if their PSA is high, but others feel that the PSA might be high for some other reason and don't refer," he told Medscape Medical News. "We developed a prediction tool to help physicians decide if a patient should really be worried about prostate cancer or not."

Men who can benefit from treatment "should be referred, but if a patient's life expectancy is limited, and there is no elevated risk of having a potentially aggressive prostate cancer, then the treatment benefit is limited to none. That is not a patient you should refer, in our view," Verbeek explained during a poster session here at the European Association of Urology 2018 Congress.

The variables upon which the prediction tool is based are readily available to general practitioners: PSA level, free-PSA percentage, patient age, estimated life expectancy, and comorbidities.

We developed a prediction tool to help physicians decide if a patient should really be worried about prostate cancer or not.

Verbeek and his colleagues analyzed data on 19,553 men 55 to 74 years of age who were initially screened as participants in the European Randomized Study of Screening for Prostate Cancer (ERSPC) from 1993 to 1999.

To adjust for healthy screen effect, the investigators used contemporary life-expectancy tables and two prognostic factors: Charlson Comorbidity Index score and self-rated health.

Survival curves for men with clinically significant prostate cancer who had not been treated were derived from the Surveillance, Epidemiology, and End Results (SEER) program, stratified by age and comorbidities.

They estimated the effect that treatment had on clinically significant prostate cancer using results from the Prostate Cancer Intervention Versus Observation Trial (PIVOT).

Prediction models for both clinically significant prostate cancer and life expectancy were evaluated with the area under the receiver operating characteristic curve.

Two Examples

The prediction model can be demonstrated with an example of two men, each with a PSA of 4.0 ng/mL and a free-PSA of 17%.

The first man is 65 years of age, has a Charlson Comorbidity Index score of 0, and good health status. "We predict a 10% chance of clinically significant prostate cancer on biopsy," Verbeek reported.

But, he asked, if this patient is referred to a urologist, "does he benefit from that referral?"

On the basis of data from the ERSPC, SEER database, and the PIVOT study, this patient's life expectancy would be 15 years if clinically significant prostate cancer is undetected and untreated. But if his cancer is detected and treated, "his life expectancy is 17 years. He lives 2 additional years with treatment," Verbeek explained. "So this patient should be referred for a biopsy."

The second man is 75 years of age, has a Charlson Comorbidity Index score of 2, and poor health status. His life expectancy is estimated to be only 5 years.

The likelihood that clinically significant prostate cancer will be detected on biopsy is higher — at 17% — because he is older. However, the 10-year mortality risk is virtually identical if a clinically significant prostate cancer is treated or if it is not (81% vs 82%).

Therefore, the mortality benefit of treatment at 10 years is only 1% in this older, sicker man.

"In our opinion, this patient should not be referred to a specialist for a biopsy," Verbeek said.

The investigators hope that the proposed tool — which they say should be available soon — will help primary care physicians triage patients for timely referral for biopsy, or not, reducing unnecessary testing and overdiagnosis.

Relying on statistics to make treatment decisions for individual patients is challenging because individual patients very often deviate from statistical predictions, said Hendrik Van Poppel, MD, PhD, from the University of Leuven in Belgium.

Statistics Tricky

The mean life expectancy today in Belgium is 81 years for men, 82 years for women. "If I have a patient who is 75, then I could say, well, he's only going to survive 7 more years, so it's not worthwhile sending him for biopsy," he told Medscape Medical News. Instead, "we may do nothing at all or we may just watch him."

But although statistics indicate that the patient will die at 81, he could live to be 92. "Then his disease progresses and he's too old to be treated aggressively. This is the problem with all the statistics we have," Van Poppel pointed out.

MRI, Van Poppel suggested, can help physicians avoid doing unnecessary biopsies simply because it only detects significant tumors and not insignificant cancers.

"We know that men need to have an MRI when their first set of biopsies is negative yet there's still a suspicion that the patient has prostate cancer, which probably means it was missed on biopsy the first time," he explained.

However, if nothing is detected on subsequent MRI, "nothing needs to be done," he said.

In contrast, if a significant tumor is present on MRI, the image can direct the operator directly to the tumor, where a precise biopsy sample, pinpointed by the MRI, can be taken.

"This allows some patients to avoid biopsies altogether," he said.

However, Van Poppel acknowledged, the use of MRI prior to standard biopsy is less common in the United States than it is in Europe, so some men might well be sent for a biopsy when none is required.

The study was supported by a grant from the Prostate Cancer UK Foundation. Verbeek and Van Poppel have disclosed no relevant financial relationships.

European Association of Urology (EAU) 2018 Congress: Abstract 277. Presented March 17, 2018.

Follow Medscape on Twitter @Medscape


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.