Prostate Specific Antigen-Growth Curve Model to Predict High-Risk Prostate Cancer.

Area of Science:

• Urology
• Oncology
• Biostatistics

Background:

• Prostate cancer (PrCA) is a significant health concern, with high-risk forms requiring early detection.
• Prostate-specific antigen (PSA) is a commonly used biomarker, but its interpretation can be challenging.
• Predictive models for high-risk PrCA are crucial for timely intervention.

Purpose of the Study:

• To evaluate the efficacy of a prostate-specific antigen (PSA)-derived growth curve in predicting high-risk prostate cancer (PrCA).
• To assess the PSA rate of change as a predictor for aggressive PrCA.
• To compare the predictive performance of PSA growth curves against single PSA measurements.

Main Methods:

• Utilized data from 38,340 men in the Prostate, Lung, Colorectal, and Ovarian Cancer Screening Trial (PLCO) to develop a PSA growth curve model.
• Employed non-linear mixed regression to model PSA growth and estimate the PSA rate of change.
• Validated the model using clinical data from 680,390 veterans, calculating the PSA rate one year prior to diagnosis or exit.

Main Results:

• A PSA rate threshold of 0.37 ng/ml/year demonstrated high sensitivity (97.2%) and specificity (97.3%) for detecting high-risk PrCA in the PLCO cohort.
• In the veteran cohort, the same threshold yielded 95.5% sensitivity and 85.2% specificity.
• An adjusted threshold of 0.99 ng/ml/year was optimal for African-American men in the VA data, with 89.1% sensitivity and 80.0% specificity.
• PSA rate proved to be a superior predictor of high-risk PrCA compared to the last single PSA value.

Conclusions:

• PSA growth curves show promise in predicting high-risk PrCA, as demonstrated in the PLCO data.
• Application of the model in the veteran cohort indicated lower specificity, highlighting the need for further refinement.
• Prospective validation is essential, particularly for African-American men, who are at higher risk for aggressive PrCA.