|
Activity Number:
|
440
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Wednesday, August 5, 2009 : 8:30 AM to 10:20 AM
|
|
Sponsor:
|
Section on Statistics in Epidemiology
|
| Abstract - #304255 |
|
Title:
|
How to Improve a Biomarker's Efficacy Using Longitudinal Data: Application to Prostate Cancer
|
|
Author(s):
|
Anna E. Kettermann*+ and Luigi Ferrucci and E. Jeffrey Metter and H. Ballentine Carter
|
|
Companies:
|
Johns Hopkins University and National Institute on Aging and National Institute on Aging and Johns Hopkins University
|
|
Address:
|
Department of Urology, Marburg 1, Baltimore, MD, 21287,
|
|
Keywords:
|
Longitudinal ; Survival Analysis ; Biomarker ; Cancer ; Reclassification
|
|
Abstract:
|
PSA is a widely recognized marker of prostate cancer. Our results indicate that a more efficacious marker could be obtained by using a man's entire PSA history. We examined the rate of change of PSA (PSAV) prior to cancer diagnosis in a cohort from the Baltimore Longitudinal Study of Aging. We found that PSAV at the transition point from linear growth to exponential growth of a subject's PSA was significantly more effective than PSA in early identification of lethal cancer. The transition point was identified using regression analysis. PSAV adjusted for age was a significant predictor of lethal cancer in the Cox regression model. Net reclassification improvement from a base model at the transition point containing age to the same model with PSAV demonstrated significant and large improvement in specificity and sensitivity.
|
