Accession Number:



Metabolomic Profiling of Prostate Cancer Progression During Active Surveillance

Descriptive Note:

Technical Report,30 Sep 2011,29 Oct 2017

Corporate Author:

Johns Hopkins School of Medicine Baltimore United States

Personal Author(s):

Report Date:


Pagination or Media Count:



This final report covers activities during the entire project period and including the no cost extension period, the completion of laboratory analyses and most statistical analyses. Delays in the project resulted in some changes to project aims delays were due to personnel changes and loss of key personnel due to illness, difficulty obtaining appropriate samples from Pathology Department, and lengthy evaluation of HRPO issues. However, the overall objectives were met, with requisite samples obtained from 100 active surveillance patients who had biopsy upgrading within 3 years, and 100 who were free of upgrading for5 years. 850 named metabolites were identified in the serum samples, with 169 significantly elevated or decreased in cases compared to controls. In urine there were 691 named metabolites, with 169 significantly elevated or decreased in cases compared to controls. Because of the time element inherent in the case and control definition, there were differences in sample age between cases and controls, creating a potential bias. Analyses were restricted to metabolites that were not correlated with sample age, excluding those with a p-value for correlation 0.10, resulting in 637 serum metabolites and 545 urine metabolites remaining in the analysis pipeline. Dimension reduction was achieved by excluding metabolites with low expression, and those with mean fold difference between the 25th and 75th percentiles, and non-significant univariate Wilcoxon rank sum test. Regularized logistic regression analysis with an elastic net penalty was applied to the reduced metabolite sets, identifying 14 serum metabolites and 9 urine metabolites significantly associated with Gleason score upgrading, with AUCs of 0.793 and 0.711, respectively. There was little overlap between the metabolites identified in both sample matrices.

Subject Categories:

  • Medicine and Medical Research

Distribution Statement: