You are here

Integrating biomarkers across omic platforms and biomaterials to stratify patients with indolent and aggressive prostate cancer


Keefe Murphy, Brendan Murphy, Susie Boyce, Louise Flynn, Sarah Gilgunn, Colm O’Rourke, Cathy Rooney andHenning Stockmann, Anna Walsh, Stephen Finn, Richard O’Kennedy, John O’Leary, Stephen Pennington, Antoinette Perry, Pauline Rudd, Radka Saldova, Orla Sheils, Denis Shields, William Watson

Publication Type: 
Refereed Original Article
Classifying indolent prostate cancer represents a significant clinical challenge.We investigated whether integrating data from different omic platforms couldidentify a biomarker panel with improved performance compared to individ-ual platforms alone. DNA methylation, transcripts, protein and glycosylationbiomarkers were assessed in a single cohort of patients treated by radicalprostatectomy. Novel multiblock statistical data integration approaches wereused to deal with missing data and modelled via stepwise multinomial logisticregression, or LASSO. After applying leave-one-out cross-validation to eachmodel, the probabilistic predictions of disease type for each individual panelwere aggregated to improve prediction accuracy using all available informa-tion for a given patient. Through assessment of three performance parame-ters of area under the curve (AUC) values, calibration and decision curveanalysis, the study identified an integrated biomarker panel which predictsdisease type with a high level of accuracy, with Multi AUC value of 0.91(0.89, 0.94) and Ordinal C-Index (ORC) value of 0.94 (0.91, 0.96), which wassignificantly improved compared to the values for the clinical panel alone of0.67 (0.62, 0.72) Multi AUC and 0.72 (0.67, 0.78) ORC. Biomarker integra-tion across different omic platforms significantly improves prediction accu-racy. We provide a novel multiplatform approach for the analysis,determination and performance assessment of novel panels which can beapplied to other diseases. With further refinement and validation, this panelcould form a tool to help inform appropriate treatment strategies impactingon patient outcome in early stage prostate cancer
Digital Object Identifer (DOI): 
Publication Status: 
Date Accepted for Publication: 
Wednesday, 13 June, 2018
Publication Date: 
Molecular Oncology
National University of Ireland, Dublin (UCD)
Open access repository: