Abstract: TH-PO1113
Biomarker Implementation: Evaluation of the Decision-Making Impact of CXCL10 Testing in a Pediatric Cohort
Session Information
- Transplantation: Clinical - Predictors of Outcomes - Biomarkers and Beyond
November 07, 2019 | Location: Exhibit Hall, Walter E. Washington Convention Center
Abstract Time: 10:00 AM - 12:00 PM
Category: Transplantation
- 1902 Transplantation: Clinical
Authors
- Lamarche, Caroline, BC Children Hospital, Vancouver, British Columbia, Canada
- Sharma, Atul Kumar, University of Manitoba, Winnipeg, Manitoba, Canada
- Sueyoshi, Tyki, University of British Columbia, Vancouver, British Columbia, Canada
- Blydt-Hansen, Tom D., University of British Columbia, Vancouver, British Columbia, Canada
Background
Children are at high risk for subclinical rejection and invasive kidney biopsy is currently used for active surveillance. Urinary CXCL10 hold the most promise as a biomarker for post-transplant monitoring of rejection. How it will influence clinical decision has never been tested. As such, our objective was to test whether CXCL10 can improve the clinical decision-making to identify organ rejection risk.
Methods
We first assembled a panel of experts to establish a minimum dataset for standard clinical decision-making for an indication biopsy. Clinical vignettes were then built from 15 prevalent pediatric kidney transplant recipients who had surveillance or indication biopsy and biobanked urine sample. Urine samples were tested for CXCL10 and reported as ratio to creatinine. Pediatric nephrologists were recruited review serial clinical vignettes and A) predict rejection risk and B) decide to biopsy; without then with urinary CXCL10 result and rejection diagnosis sensitivity/specificity information for different levels. Biopsy decisions were then correlated with the biopsy results. Inter-rater agreement (IRA) was assessed by Fleiss Kappa (K) for binary choice and interclass correlation (ICC) for probabilities.
Results
Eleven pediatric nephrologists were enrolled. IRA for choice to biopsy was fair both before (K=0.48, p<0.01) and after (K=0.43, p<0.01) incorporating CXCL10 data. ICC of probability assessment for rejection was poor before (0.28, p<0.01) and improved to fair (0.48, p<0.01) with addition of chemokine data (p=0.6 for difference). Clinicians did consider the CXCL10 in their decision making process and CXCL10/Cr correlated with the change in the estimated probability of rejection (r2=0.7756, p<0.0001), with 2.77 as the cutoff for a biopsy-based intervention. The decision-accuracy (majority ordering a biopsy when rejection was found or not ordering one when there was no rejection) improved from 8/15 (53.3%) cases to 11/15 (73.3%) with CXCL10 results. Using the cut-off value for CXCL10/Cr ratio of 2.77 would have been accurate in 12/15 cases (80%).
Conclusion
There is high variability in decision-making on biopsy indication. Urinary CXCL10/Cr improves probability estimates for risk of rejection. However, training may be required to assist nephrologists in using biomarker information for clinical decision-making.