51˶

Another Prostate Cancer Screening Candidate Outperforms Standard PSA Testing

— Multiparametric Stockholm3 avoided far more unnecessary biopsies without sacrificing sensitivity

MedpageToday

SAN FRANCISCO -- A multiparametric blood test for prostate cancer showed potential to avoid more than half of unnecessary biopsies without sacrificing accuracy, a large prospective study showed.

In a comparison against the current PSA testing standard of ≥4 ng/mL, the Stockholm3 biomarker, which incorporates a PSA cutoff of 15 ng/mL with other proteins and genomic information, would have spared 56% of men from biopsies for grade group (GG) 1 or benign disease. The standard PSA cutoff would have avoided 19% of unnecessary biopsies, decreasing to 10% with a cutoff of ≥3 ng/mL.

Sensitivity, specificity, and predictive values with the multicomponent test either approximated or surpassed those achieved with conventional PSA testing, reported Scott Eggener, MD, of the University of Chicago, at the Genitourinary Cancers Symposium.

"Stockholm3 has the potential to reduce unnecessary harms of prostate cancer screening," said Eggener. "Stockholm3 has attractive characteristics in this diverse cohort, and [demonstrated] successful recruitment of a large cohort of men from traditionally underrepresented minorities. The Stockholm3 should be available in the early part of 2024."

During a discussion that followed the presentation, a viewer following online asked how the Stockholm3 compared with other tests, such as the and the .

"It's a busy space, the secondary biomarker space, and the question always comes up is how they compare head to head," said Eggener. "They're all based on good science. All of them have good studies showing their value. For a variety of reasons that you all probably know, it's rare for them to go head to head, so we really don't know."

In response to another question, Eggener noted that the 16% rate of MRI use in the study was lower than might be expected in a population-based investigation, probably because more than half of the blood specimens were archival from a biobank.

"It would be interesting to see if prospective routine use of MRI would change the operating characteristics," he said.

A final question related to investigators' ability to recruit a racially/ethnically diverse population, as Asian, Black, and Hispanic patients accounted for a majority of the blood specimens. Eggener said investigators sought out participating sites that have diverse patient populations.

Stockholm3: The Basics

The Stockholm3 incorporates proteins (total and free PSA plus three cancer-related proteins -- GDF15, KLK2, and PSP94), polygenic risk using 101 single-nucleotide polymorphisms, and patient clinical data. The test has a multiplex algorithm that uses the multiparametric data to predict the likelihood of GG ≥2 prostate cancer. A risk score ≥11 is considered indicative of prostate cancer.

An conducted in Sweden showed use of the test reduced benign biopsies and diagnosis of clinically insignificant prostate cancers. In a subsequent study, use of the test reduced MRI procedures and referrals for prostate biopsy. Racial and ethnic representation in prostate cancer trials is typically poor, Eggener noted. Investigators in the sought to address that issue by evaluating Stockholm3 performance metrics in a diverse North American population, involving 17 centers in the U.S. and Canada.

Eligible participants were ages 45-75, had no known prostate cancer but a clinical indication for prostate biopsy, and self-identified as Asian, Black/African American, Hispanic/Latino nonwhite, or white. All participants had blood drawn prior to prostate biopsy.

The primary endpoint was GG ≥2, and the key secondary endpoints were GG 1 prostate cancer or benign disease. A primary objective was to evaluate the relative sensitivity of Stockholm3 (PSA ≥15) and PSA ≥4 for detecting GG ≥2. A second objective was to compare the relative specificity of Stockholm3 and PSA ≥4 for avoiding GG 1 cancer or benign disease.

SEPTA included 912 men recruited from the 17 participating sites plus 1,217 men who had been biopsied and had biobanked blood. Study participants had a median age of 63 and a median PSA of 6.1 ng/mL. A fourth of the patients had an abnormal digital rectal exam, 20% had a family history of prostate cancer, and 20% had a prior negative biopsy. The median Stockholm3 score was 17, and 16% of the patients had undergone MRI-targeted biopsies.

Key Findings

Investigators achieved the ethnic diversity they sought, as 16% of patients were Asian, 24% were Black, 14% were Hispanic, and 46% were white. Biopsy results showed benign disease in 57% of cases, GG 1 prostate cancer in 14%, and GG ≥2 in 29%. Analysis of biopsy findings by ethnicity showed that Asian and Hispanic men had more benign disease (69% and 60%, respectively) and Black men had fewer benign diagnoses (49%). Detection of GG 1 was similar across the groups (11-17%). Black men had more GG ≥2 (37%) and Asians had less (21%).

Biopsy results showed that a PSA ≥3 ng/mL detected 97% of GG ≥2, followed by PSA ≥4 (92%) and Stockholm3 (88%). Use of Stockholm3 avoided almost three times as many unnecessary biopsies as a PSA cutoff of ≥4 ng/mL and almost six times more than a cutoff of ≥3.

As compared with PSA cutoffs of ≥3 or ≥4 ng/mL, the Stockholm3 test achieved the following values:

  • Specificity: 56% vs 10% vs 19%, respectively
  • Negative predictive value (NPV): 92% vs 91% vs 85%
  • Sensitivity: 88% vs 97% vs 92%
  • Positive predictive value (PPV): 44% vs 31% vs 31%

Lowering the Stockholm3 cutoff to ≥11 ng/mL led to slightly better NPV, sensitivity, and PPV, but at the expense of 18% more biopsies for low-risk and benign disease.

Disclosures

The SEPTA study was sponsored by the Karolinska Institute.

Eggener disclosed relationships with A3P Biomedical (Stockholm3 developer), Candel Therapeutics, Cellvax, Janssen, MetasTx, and OptumHealth.

Primary Source

Genitourinary Cancers Symposium

Vigneswaran HT, et al "SEPTA: Stockholm3 validation study in a multi-ethnic cohort for prostate cancer detection" GuCS 2024; Abstract 262.