Title : Heart transplant rejection: From endomyocardial biopsy to molecular testing and pediatric considerations
Abstract:
Background: Heart transplantation remains the definitive therapy for end-stage heart failure, but allograft rejection continues to be a major cause of morbidity and graft dysfunction. Traditional surveillance has relied on endomyocardial biopsy (EMB), which provides direct histologic assessment but is invasive, carries procedural risks, and is subject to interobserver variability.
Objectives: Emerging noninvasive surveillance tools, including donor-specific antibodies (DSAs), donor-derived cell-free DNA (dd-cfDNA), and gene expression profiling (GEP), were systematically reviewed, with particular emphasis on pediatric patients.
Results: Studies of donor-derived cell-free DNA (dd-cfDNA), including D-OAR/DEDUCE, GRAFT, and SHORE, consistently show high negative predictive value for excluding rejection. In the AlloMap Registry, a threshold of 0.25 for moderate-to-severe acute cellular rejection yielded 83% accuracy, 58% sensitivity, and 93% specificity; combined gene expression profiling (GEP) and dd-cfDNA testing achieved a 97% negative predictive value. GRAFT and DEDUCE, including Prospera-based testing, reported similarly strong rule-out performance. GEP studies (CARGO, CARGO 2, IMAGE, and eIMAGE) also demonstrated high negative predictive value, and randomized trials showed outcomes comparable to biopsy-based surveillance. AlloMap remains the only FDA-approved assay and is listed as a class IIa recommendation in the 2010 ISHLT guidelines. Recent guideline updates in 2023 increasingly support molecular biomarkers, particularly in pediatric practice: they are considered reasonable in infants and younger children and are now recommended as primary noninvasive screening tools in older children and adolescents. Interpretation of molecular assays requires caution in settings such as infection, systemic inflammation, ischemia, trauma, recent transfusion, other organ or stem-cell transplantation, and HIV. In addition, three studies of artificial intelligence (AI) in histopathology reported diagnostic accuracies of 71%–98% for acute cellular rejection, antibody-mediated rejection, and cardiac allograft vasculopathy.
Conclusion: The field is moving toward a hybrid surveillance model that combines histology, molecular profiling, and risk stratification to detect rejection more accurately and less invasively. Pediatric studies further suggest that integrating molecular methods with risk prediction models may meaningfully reduce biopsy frequency without compromising rejection detection. Future progress will likely depend on stronger validation of these tools, integration with AI-assisted interpretation, and tailored application across adult and pediatric transplant populations.
