Date: Saturday, June 2, 2018
Session Time: 5:30pm-7:30pm
Presentation Time: 5:30pm-7:30pm
Location: Hall 4EF
Sub-clinical acute rejection (subAR) following KT is associated with poor long-term graft outcomes and can only be detected using invasive surveillance biopsies (SBx). Non-invasive biomarker detection of subAR is needed. The objective of this study is to develop a gene expression profile (GEP) that can detect subAR in the peripheral blood.
Precise clinical phenotypes (PCP) were used to define subAR in pts in a multi-center 24-mo observational study at centers that do routine SBx. Paired blood was shipped to a central lab and processed in batches using Affymetrix HT HG-U133+PM Peg Arrays. Following adjustment for batch effect, differential gene expression (DGE) analysis was performed to select potential classifiers. Random Forests (RF) and Gini importance were used to select the top model optimized for AUC. Bootstrap resampling was used to test for overfitting. Threshold selection was based on model performance metrics in the discovery cohort. We then validated the locked model/threshold of the subAR GEP on a separate prevalent cohort using paired samples.
Using DGE from 530 samples (subAR n=130; 79% borderline) collected from 250 KTR, we selected a RF model that consists of 61 probe sets mapping to 57 genes (AUC 0.85; 0.84 after bootstrap). We selected a predicted probability threshold of 0.375 based on best overall performance, favoring specificity and NPV (87% and 88%) over sensitivity and PPV (64% and 61%, respectively). We tested the locked model/threshold on DGE data from a separate cohort of 138 KTR who underwent SBx (subAR 42) at a single institution: NPV 78%; PPV 51%.
We have developed and validated a GEP that detects subAR post KT. Our study emphasizes important specifications: multi-center study with prevalent (subAR) populations; exclusive use of paired samples with central histology reads; strict oversight of PCP; application of blinded open source bioinformatics for discovery; threshold selection based on anticipated use of the biomarker; use of a locked model/threshold for validation.
CITATION INFORMATION: Whisenant T., Kurian S., Kandpal M., Zhao L., Ikle D., Armstrong B., Friedewald J., Heilman R., Poggio E., Marsh C., Baliga P., Bridges N., Odim J., Brown M., Charette J., Brietigam S., Sustento-Reodica N., Salomon D., Abecassis M. Bioinformatics Approach to the Development of a Novel Molecular Biomarker for Sub-Clinical Acute Rejection (subAR) in the Peripheral Blood Following Kidney Transplant (KT) Am J Transplant. 2017;17 (suppl 3).
To cite this abstract in AMA style:Whisenant T, Kurian S, Kandpal M, Zhao L, Ikle D, Armstrong B, Friedewald J, Heilman R, Poggio E, Marsh C, Baliga P, Bridges N, Odim J, Brown M, Charette J, Brietigam S, Sustento-Reodica N, Salomon D, Abecassis M. Bioinformatics Approach to the Development of a Novel Molecular Biomarker for Sub-Clinical Acute Rejection (subAR) in the Peripheral Blood Following Kidney Transplant (KT) [abstract]. https://atcmeetingabstracts.com/abstract/bioinformatics-approach-to-the-development-of-a-novel-molecular-biomarker-for-sub-clinical-acute-rejection-subar-in-the-peripheral-blood-following-kidney-transplant-kt/. Accessed June 4, 2020.
« Back to 2018 American Transplant Congress