Abstract: FR-PO786
Serial Testing of Blood Gene Expression and Donor-Derived Cell-Free DNA for Predicting Future Kidney Allograft Failure
Session Information
- Transplantation: Clinical - Biomarkers
November 04, 2022 | Location: Exhibit Hall, Orange County Convention Center‚ West Building
Abstract Time: 10:00 AM - 12:00 PM
Category: Transplantation
- 2002 Transplantation: Clinical
Authors
- Guo, Kexin, Northwestern University Feinberg School of Medicine, Chicago, Illinois, United States
- Kushner, Alexis J., Northwestern University Feinberg School of Medicine, Chicago, Illinois, United States
- Rebello, Christabel, Northwestern University Feinberg School of Medicine, Chicago, Illinois, United States
- Kinsella, Bradley M., Northwestern University Feinberg School of Medicine, Chicago, Illinois, United States
- Sinha, Rohita, Eurofin Viracor, Lee Summit, Missouri, United States
- Zhao, Lihui, Northwestern University Feinberg School of Medicine, Chicago, Illinois, United States
- Friedewald, John J., Northwestern University Feinberg School of Medicine, Chicago, Illinois, United States
- Park, Sookhyeon, Northwestern University Feinberg School of Medicine, Chicago, Illinois, United States
Background
Both gene expression profiles (GEP) and donor-derived cell-free DNA (dd-cfDNA) have been established as non-invasive biomarkers to detect subclinical and clinical acute rejection in a kidney allograft. However, the clinical impact of combined serial testing of GEP and dd-cfDNA on kidney allograft survival has not been well studied. We hypothesized that we could predict future kidney allograft survival using serial testing of combined GEP and dd-cfDNA.
Methods
We analyzed 261 subjects from a previously reported multicenter, prospective observational study. Multiple serial samples of GEP and dd-cfDNA were collected throughout the study period. Graft failure was defined as returning to dialysis or re-transplant. We used a joint model to predict future allograft survival using serial GEP and dd-cfDNA and allograft failure. The study cohort was randomly divided into 70 and 30%, training and testing sets, respectively. We assessed the model performance with the area under the receiver operating characteristic (AUROC).
Results
Of 261 subjects, 182 (70%) were used for training the model. A total of 16 cases of allograft failure were observed in the training set. In the training set, the AUROC to predict graft failure at 4-year post kidney transplant (KT) was 0.83 using GEP and dd-cfDNA until 1 year KT (Fig 1A). When we used GEP and dd-cfDNA data from up to 2 years post KT, the AUROC improved to 0.88 (Figure 1B). For the validation set, 7 graft failures were observed from 79 subjects. The performance remains stable in the validation set. The AUROCs were 0.67 and 0.83, using up to 1-year and 2-year serial GEP and dd-cfDNA data, respectively.
Conclusion
The combination of GEP and dd-cfDNA tests can be used to predict future graft failure.
Funding
- Private Foundation Support