Nonparametric Percentile Curve Estimation for Prostate-Specific Antigen

Principal Investigator

Name
Oke Gerke

Degrees
MSc, PhD

Institution
Odense University Hospital

Position Title
Professor in Clinical Biostatistics in Diagnostic Research

Email
oke.gerke@rsyd.dk

About this CDAS Project

Study
PLCO (Learn more about this study)

Project ID
PLCO-1754

Initial CDAS Request Approval
Dec 4, 2024

Title
Nonparametric Percentile Curve Estimation for Prostate-Specific Antigen

Summary
Earlier research of our group considered the derivation of nonparametric percentile curve estimation for a nonnegative marker with excessive zeros, exemplified with the Agatston score for coronary calcification (see https://pubmed.ncbi.nlm.nih.gov/35782723/ ; https://pubmed.ncbi.nlm.nih.gov/34179988/). In preparation of my DMSc thesis defense on December 12, 2024, I would like to present pecentile curves for prostate-specific antigen in the same way as I derived these earlier for the Agatston score. The reason for doing so is to exemplify the technique with another marker (meaning prostate-specific antigen) that follows quite a different distribution than the Agatston score. In screening populations, the Agatston score is often highly right-skewed with an overexpression of zeros.

Aims

- Derive nonparametric percentile curves for all percentiles 1,2,3,...,99 for prostate-specific antigen in a male screening population
- Visualize 25%, 50%, 75%, 90% percentile curves as shown in Figure 3 of https://pubmed.ncbi.nlm.nih.gov/35782723

Collaborators

None.