Blood Test Detects Heart Transplant Rejection in Asian Patients, First Prospective Study Finds

For the tens of thousands of people living with a donor heart, the most feared complication is acute rejection, the immune system’s assault on the transplanted organ. Diagnosing it has long required endomyocardial biopsy, an invasive procedure in which a flexible catheter snips tiny fragments of heart muscle for pathologists to examine under the microscope. Now, a prospective study from Fuwai Hospital in Beijing has delivered the first rigorous validation of a blood-based alternative for Asian heart transplant recipients, establishing a specific genetic threshold above which rejection is likely and below which, the researchers caution, the danger is not necessarily over.

The study, published in Transplant International, enrolled 80 heart transplant recipients and analyzed 299 plasma samples using a domestically developed next-generation sequencing assay. The technology rests on a simple biological premise: when donor heart cells die, they release their DNA into the recipient’s bloodstream. Because donor and recipient genomes differ, this donor-derived cell-free DNA, or dd-cfDNA, can be quantified as a fraction of total circulating DNA. Rising levels signal graft injury, potentially long before symptoms or echocardiographic changes appear.

What makes the Chinese assay distinctive is its genetic architecture. Most commercial dd-cfDNA tests, including the widely used AlloSure platform, distinguish donor from recipient DNA by genotyping single nucleotide polymorphisms, single-letter genome variations. The new assay instead targets 48 insertion-deletion markers, or Indels, stretches of DNA that are either present or absent entirely. These markers were selected for minor allele frequencies of 0.4 to 0.6 in East Asian populations, a range that maximizes the probability that any given donor-recipient pair will differ at enough sites for reliable quantification. The researchers argue that Indel-based detection may overcome the single-base resolution limits of short-read sequencing and reduce background noise, while the multiplexing capacity of sequencing surpasses digital PCR by assessing dozens of loci simultaneously, protecting against single-marker failure.

To interpret any dd-cfDNA measurement, clinicians first need to know what normal looks like. The team established this with a reference cohort of 42 clinically stable recipients, monitored at one, two, four, eight, 24, and 48 weeks after transplantation. Levels peaked in the first week, a median of 0.29%, reflecting the expected wave of ischemia-reperfusion injury from the surgery itself, then declined steadily. A stable trend emerged around week four, and by week eight levels had reached their nadir of 0.19%, remaining low through one year. Statistical modeling using generalized estimating equations confirmed significant declines from weeks one and two to week eight, with no meaningful change thereafter. The practical implication is that dd-cfDNA signals obtained before eight weeks post-transplant should be interpreted cautiously, since early levels are confounded by surgical injury.

The diagnostic core of the study involved 41 patients undergoing biopsy for clinical suspicion of rejection, yielding 42 biopsies paired with 56 dd-cfDNA samples. Pathologists graded the specimens according to International Society for Heart and Lung Transplantation criteria while remaining blinded to the blood results. The findings were striking: patients with biopsy-proven acute rejection had a median dd-cfDNA of 0.87%, nearly four times the 0.23% seen in those without rejection, a difference with a P value below 0.001. Levels rose in parallel with histological severity. Biopsies showing significant acute cellular rejection, grade 2R or higher, averaged 0.81%, while mild grade 1R rejection did not differ from rejection-free tissue. Antibody-mediated rejection showed an even steeper gradient, with grade 2 or higher AMR reaching 1.04%, consistent with the biology of humoral rejection, in which antibodies attack graft endothelium, a rich source of released DNA.

Receiver operating characteristic analysis identified 0.33% as the optimal composite threshold for detecting acute rejection, yielding an area under the curve of 0.874, sensitivity of 81.0%, and specificity of 85.7%. For antibody-mediated rejection alone, a higher cutoff of 0.69% performed best, with 80.0% sensitivity and 92.9% specificity. Notably, the 0.33% threshold is numerically higher than the 0.20% to 0.25% cutoffs established in predominantly Western cohorts using SNP-based assays. The researchers attribute this discrepancy to a combination of possible ethnic variations in cell-free DNA metabolism and immunological responses, fundamental methodological differences between Indel-based and SNP-based platforms, and cohort selection biases. Without head-to-head comparison on identical samples, the relative contributions of biology versus technology cannot be disentangled, a limitation the authors acknowledge openly.

The study also demonstrated that dd-cfDNA tracks treatment response. In 14 patients with paired measurements before and two weeks after anti-rejection therapy, median levels fell from 0.98% to 0.38%, a statistically significant decline that paralleled substantial drops in NT-proBNP, a marker of cardiac strain. When stratified by rejection type, the reduction was significant only in the cellular rejection subgroup; antibody-mediated rejection showed a more modest, non-significant decline, suggesting that resolution of humoral graft injury may require a longer observation window than two weeks. Left ventricular ejection fraction, meanwhile, did not change significantly over the same period, hinting that structural recovery lags behind the molecular resolution of injury.

Perhaps the most biologically revealing finding concerns graft function. Across the entire cohort, dd-cfDNA showed a graded inverse correlation with left ventricular ejection fraction: patients with severely reduced function below 50% had the highest levels, those with preserved function above 60% the lowest. Yet when the analysis was stratified by rejection status, the inverse association persisted only in the rejection subgroup and vanished among rejection-free patients. This suggests that dd-cfDNA specifically reflects immune-mediated myocardial injury rather than serving as a general marker of cardiac dysfunction. Discordant cases reinforced the point: among 12 patients with elevated dd-cfDNA despite preserved ejection fraction, 11 had biopsy-proven rejection, indicating that molecular injury can precede functional deterioration, while patients with reduced function but low dd-cfDNA typically had no rejection on biopsy, their dysfunction attributable to non-immunological factors such as altered loading conditions or arrhythmias.

The researchers are careful about what their threshold can and cannot do. Because their for-cause biopsy cohort had a high rejection prevalence of roughly 60%, the positive predictive value of 89.5% is inflated relative to routine surveillance settings, while the negative predictive value of 75.0% falls well short of the 94% or higher figures reported in low-prevalence surveillance studies. In practical terms, a dd-cfDNA level at or above 0.33% should trigger comprehensive evaluation, but a low value should never be used in isolation to defer biopsy. The authors frame the test as better suited to a rule-in than a rule-out strategy, an adjunctive tool within multimodal monitoring rather than a replacement for histology. They also note that infection, which was significantly more frequent in the rejection group, may independently elevate dd-cfDNA and could not be adjusted for given the sample size.

The study’s limitations are substantial and the authors state them plainly: a single center, modest sample size, only nine antibody-mediated rejection events, no protocol biopsies in the reference cohort, and no direct benchmarking against commercial assays. Internal validation through 1,000 hierarchical bootstrap resamples produced point estimates clustered near 76% but with confidence intervals spanning roughly 37 to 40 percentage points, underscoring the preliminary nature of the findings. Still, as the first prospective evaluation of dd-cfDNA in a Chinese heart transplant population using a domestically developed platform, the study fills a genuine gap. External validation in multicenter cohorts is required before routine clinical implementation, but the trajectory is clear: the era of monitoring a transplanted heart through a simple blood draw, calibrated to the genetics of the population it serves, is drawing measurably closer.

Subject of Research: Heart Transplantation

Article Title: Prospective evaluation of donor-derived cell-free DNA for noninvasive detection of acute rejection in an Asian heart transplant population

Article References: Prospective evaluation of donor-derived cell-free DNA for noninvasive detection of acute rejection in an Asian heart transplant population. (n.d.). https://doi.org/10.3389/ti.2026.16415

Image Credits: AI Generated

DOI: 10.3389/ti.2026.16415

Keywords: heart transplantation, donor-derived cell-free DNA, acute rejection, liquid biopsy, next-generation sequencing, transplant immunology, Prospective, evaluation, donor-derived, cell-free, noninvasive, detection