Longitudinal Modelling of Prostate-Specific Antigen (PSA) Trajectories and Their Association with Prostate Cancer Diagnosis in the PLCO Trial

Principal Investigator

Name
CiarÃ¡n Courtney O'Toole

Institution
University of Limerick

Position Title
PhD Researcher

Email
otoole.ciaran@ul.ie

About this CDAS Project

Study
PLCO (Learn more about this study)

Project ID
PLCO-2025

Initial CDAS Request Approval
Mar 16, 2026

Title
Longitudinal Modelling of Prostate-Specific Antigen (PSA) Trajectories and Their Association with Prostate Cancer Diagnosis in the PLCO Trial

Summary
Prostate cancer screening strategies increasingly emphasise risk-adapted approaches informed by individual patient characteristics and biomarker trajectories. Prostate-specific antigen (PSA) remains central to early detection, yet interpretation of PSA changes over time remains challenging. Understanding the longitudinal behaviour of PSA, including its variability, velocity, and doubling time, may improve prediction of prostate cancer diagnosis and guide clinical decision-making.
This project aims to leverage the extensive longitudinal PSA data available within the Prostate, Lung, Colorectal and Ovarian (PLCO) Cancer Screening Trial to explore PSA dynamics prior to prostate cancer diagnosis. By analysing repeated PSA measurements among participants, we will model PSA trajectories and assess their association with subsequent prostate cancer detection, stage, and grade.
Using mixed-effects, we will quantify how individual PSA slopes and baseline levels relate to the likelihood and timing of prostate cancer diagnosis, and whether the pattern differ across key patients groups, such as age, baseline PSA, race, and family history. The findings will inform methodological strategies for modelling longitudinal biomarkers in prostate cancer and provide empirical evidence for developing future risk calculators that incorporate PSA trends.
This analysis forms part of broader PhD research focused on improving prostate cancer risk prediction in the Irish clinical context, where longitudinal PSA data are limited. Insights gained from PLCO will guide the design and validation of models using real-world data collected from Irish Rapid Access Prostate Clinics, supporting development of an MRI-integrated prostate cancer risk calculator.

Aims

• To model individual longitudinal PSA trajectories among PLCO participants and assess their predictive value for prostate cancer and clinically significant disease.
• To estimate the association between PSA trajectory parameters (baseline PSA, velocity, acceleration) and subsequent prostate cancer diagnosis.
• To compare PSA dynamics between participants diagnosed with prostate cancer and those without, investigated for differences across age, baseline PSA, and screening arm.
• To evaluate the predictive performance and calibration of models for prostate cancer diagnosis over a defined follow-up period.
• To provide methodological insights and parameter estimates to inform development of an MRI-integrated risk calculator using Irish clinical data.

Collaborators

Ciarán Courtney O'Toole University of Limerick
Amirhossein Jalali University of Limerick