Acute cellular allograft rejection remains a major cause of limited long-term graft survival in cardiac transplantation. A critical challenge in trying to reduce the incidence of acute cellular rejection arises from the difficulty of accurately and efficiently making diagnoses. The primary approach requires taking small pieces of heart muscle tissue, which is highly invasive and costly, and suffers from sampling error and inter-observer grading variability. Replacing the biopsy with a simple blood test would be of great value to patients and substantially reduce healthcare costs.
Approach and Methods
We previously used high-throughput, untargeted transcriptomic profiling in blood samples of heart transplant patients to identify 9 mRNAs and 5 proteins whose combined expression discriminated patients undergoing acute cellular rejection from those who were not. We now validated the mRNA targets on a clinically-amenable NanoString nCounter platform. The performance of the novel assay, HEARTBIT, as well as that of a proteogenomic ensemble including 5 proteins, was assessed by cross-validation.
In cross-validation the area under the receiver operating characteristic curve (AUC) of the transcriptomic signature was 0.81, with 47% specificity at ≥ 90% sensitivity. Addition of 5 proteins to the mRNA panel using ensembling resulted in an enhanced performance (AUC of 0.86, with 65% specificity at ≥ 90% sensitivity, in cross-validation).
Summary and Conclusion
Here, we demonstrate 1) successful translation of biomarker signatures from untargeted high-throughput screening onto a compact, clinically-amenable technological platform, and 2) promising utility of our novel assay, HEARTBIT, for improved detection of acute cellular rejection.