TOPLINE:

In a study involving more than 11,000 men enrolled in the Prostate, Lung, Colorectal, and Ovarian (PLCO) cancer screening trial, about one quarter of elevated prostate-specific antigen (PSA) results (≥ 2.5, 3.0, or 4.0 ng/mL) dropped below biopsy threshold the following year. More than half of men with at least one elevated result had subsequent normalization. Men with persistently elevated PSA values had a low likelihood of normalization and could proceed directly to diagnostic evaluation without repeat testing.

METHODOLOGY:

• PSA testing often leads to unnecessary biopsies due to limited specificity and intraindividual variability. Guidelines recommend repeat testing before biopsy, but the benefit is not uniform.
• Researchers conducted a retrospective, multicenter cohort study analyzing data of 11,176 men (median age, 60 years) enrolled in the PLCO screening trial who received annual PSA testing over 6 years (between 1995 and 2006), were not diagnosed with prostate cancer during this period, and had at least two PSA measurements taken 1 year apart.
• To evaluate variability, researchers calculated two metrics: the probability that an elevated PSA measurement would fall below the same threshold at the subsequent annual test (PSA-level analysis) and the proportion of patients with at least one elevated PSA result who experienced such a drop during follow-up (patient-level analysis).
• Researchers evaluated three predictors — current PSA level, prior PSA level above the threshold, and prior PSA level below the threshold — using univariable and multivariable logistic regression.
• The primary outcome was a PSA measurement at or above one of three biopsy thresholds of interest (2.5, 3.0, and 4.0 ng/mL) that decreased below the threshold at the subsequent annual measurement. Overall, 2700 patients were included at a PSA threshold of 2.5 ng/mL, 1928 at 3.0 ng/mL, and 952 at 4.0 ng/mL.

TAKEAWAY:

• Overall, 22% of PSA measurements ≥ 2.5 ng/mL decreased below this threshold the following year. Corresponding rates were 25% for the threshold of 3.0 ng/mL and 30% for the threshold of 4.0 ng/mL.
• At patient level, 54% of men with at least one PSA measurement ≥ 2.5 ng/mL had a subsequent PSA level below this threshold; the rates of PSA normalization increased to 58% and 62% for thresholds of 3.0 and 4.0 ng/mL, respectively.
• A multivariable model using the log-transformed difference between the current PSA level and the threshold, along with indicators of whether prior PSA levels were above or below the threshold, achieved high discrimination (area under the curve > 0.78) across training, testing, and validation sets.
• Patients with PSA levels persistently above a given biopsy threshold (risk score, -3) had a low probability (< 10%) of the PSA level decreasing below the threshold on the subsequent annual test. This probability increased to approximately 20% for a score of -1, approximately 30% for a score of 0, and 50%-60% for a score of 2.

IN PRACTICE:

In this study, “significant intraindividual variability in PSA levels was observed, with many elevated values falling below the threshold at the next yearly measurement,” the authors wrote. “These findings suggest the utility of guideline recommendations to confirm elevated PSA results in most patients before performing further diagnostic evaluation and that patients with a prior PSA score above a given biopsy threshold and no recent PSA scores below that threshold could proceed to further diagnostic evaluation without repeat testing,” they concluded.

SOURCE:

The study, led by Nicholas Pickersgill, MD, Memorial Sloan Kettering Cancer Center, New York City, was published online in JAMA Oncology.

LIMITATIONS:

The study lacked data on factors influencing PSA, was limited to annual testing, and excluded men later diagnosed with prostate cancer. The PLCO cohort may not represent all men undergoing PSA screening.

DISCLOSURES:

The study was supported by grants from National Cancer Institute, National Institutes of Health. Several authors reported receiving grants or personal fees from and having other ties with various sources. Full disclosures are noted in the original article.

This article was created using several editorial tools, including AI, as part of the process. Human editors reviewed this content before publication.