Circulating MicroRNA Scores May Distinguish Cellular and Antibody-Mediated Heart Transplant Rejection With High Accuracy

Proposed Section Structure

This article is organized into the following sections: clinical background and unmet need; study design and methods; key diagnostic findings; prognostic implications; biological rationale and clinical interpretation; strengths and limitations; implications for practice and research; funding, registration, and citation.

Highlights

First, in a prospective 5-center cohort, circulating microRNA-based clinical rejection scores (CRS) showed excellent discrimination for biopsy-proven acute cellular rejection (ACR) and antibody-mediated rejection (AMR), with area under the receiver-operating characteristic curves of 0.93 and 0.92, respectively.

Second, using a CRS threshold of 65, both panels had very high negative predictive value: 100% for ACR and 99% for AMR. This is the kind of performance profile that is especially attractive for a screening test intended to reduce unnecessary endomyocardial biopsies (EMBs).

Third, elevated CRS values were not only cross-sectionally associated with rejection on concurrent biopsy, but also prospectively associated with a higher hazard of subsequent rejection, graft dysfunction, or death.

Fourth, the study addresses both major immunologic phenotypes of cardiac allograft rejection, a notable advance because many prior biomarker efforts have focused more narrowly on cellular rejection.

Clinical Background and Unmet Need

Surveillance after heart transplantation still depends heavily on EMB, an invasive procedure that is resource-intensive, uncomfortable for patients, and imperfect in practice. EMB samples only a small fraction of myocardium, creating a recognized risk of sampling error. Histopathologic interpretation also depends on expertise and may vary across centers, particularly for lower-grade findings and mixed rejection phenotypes.

The clinical problem is compounded by the biological heterogeneity of rejection. ACR is driven predominantly by T-cell-mediated injury, whereas AMR reflects antibody-mediated immune activation, endothelial injury, and microvascular inflammation. These phenotypes can overlap clinically, yet their diagnosis and management differ. A reliable blood-based assay that can detect either subtype at the time of routine surveillance would therefore address a major unmet need in transplant cardiology.

MicroRNAs are short noncoding RNAs that regulate post-transcriptional gene expression and can be measured in circulation. Because they are relatively stable in blood and may reflect cell-specific injury and immune activation pathways, they are biologically plausible candidates for transplant biomarker development. The present study by Goldberg and colleagues builds on previously identified microRNA signatures for ACR and AMR and tests their performance in a contemporary, multicenter prospective cohort.

Study Design and Methods

The investigation was conducted within GRAfT (Genomic Research Alliance for Transplantation), a 5-center, prospective, longitudinal cohort study. The cohort included 173 heart transplant recipients. Women comprised 29% of participants and 41% were Black, an important detail for evaluating diversity and early signals of generalizability. Median follow-up was 374 days after transplantation.

MicroRNA sequencing was performed on 922 blood samples, of which 720 were paired with EMB obtained at the same timepoint. The analysis leveraged previously identified microRNA panels associated with ACR (12 microRNAs) and AMR (17 microRNAs). These were incorporated into logistic regression models to generate separate ACR and AMR clinical rejection scores scaled from 0 to 100.

The primary diagnostic assessment was the area under the receiver-operating characteristic curve (AUC) for detection of biopsy-defined ACR and AMR. The authors also examined clinically useful operating characteristics at a prespecified or selected CRS threshold of 65, including sensitivity, specificity, and negative predictive value. Beyond concurrent diagnosis, the study used adjusted Cox proportional hazards models to test whether higher CRS values predicted a composite long-term outcome consisting of subsequent ACR or AMR by EMB, allograft dysfunction, or death.

Key Results

Diagnostic performance for acute cellular rejection

There were 14 episodes of ACR among the biopsy-paired samples. The median ACR CRS was 78 in samples with ACR versus 42 in those without ACR, a highly significant difference (P<0.001). The AUC for the ACR CRS was 0.93, indicating excellent discrimination.

At a CRS threshold of 65, the ACR panel achieved 79% sensitivity and 97% specificity. Most strikingly, the negative predictive value was 100%. Although the number of rejection events was modest, this result suggests the test may be especially useful to rule out clinically meaningful ACR in low-score samples.

Diagnostic performance for antibody-mediated rejection

There were 25 episodes of AMR. The median AMR CRS was 75 in samples with AMR compared with 53 in those without AMR (P<0.001). The AUC for the AMR CRS was 0.92, again in the excellent range.

Using the same CRS threshold of 65, the AMR panel showed 84% sensitivity and 86% specificity, with a 99% negative predictive value. AMR has been a particularly challenging target for noninvasive diagnostics, so this level of performance is clinically notable.

Potential biopsy-sparing implications

Both the ACR and AMR CRS were below 65 in 589 tests, representing 82% of evaluable paired assessments. This observation is highly relevant to implementation. If validated in independent cohorts and embedded into a safe clinical algorithm, a large proportion of routine surveillance encounters might be managed without immediate biopsy, reserving EMB for higher-risk or discordant cases.

Prognostic significance

The value of the assay extended beyond same-day diagnosis. After adjustment, each 10-point increase in ACR CRS was associated with a 42% increase in the hazard of the composite outcome of future rejection, allograft dysfunction, or death (hazard ratio [HR], 1.42; 95% CI, 1.20-1.69; P<0.001). For the AMR CRS, each 10-point increase was associated with a 45% increase in hazard (HR, 1.45; 95% CI, 1.14-1.86; P=0.003).

This finding suggests that the microRNA signal may capture a state of ongoing alloimmune activity or graft vulnerability even when the paired biopsy does not yet show overt rejection, or when pathology underestimates disease burden because of spatial sampling limitations.

Why These Findings Matter Clinically

The major practical strength of this study is that it addresses the real decision point clinicians face: what to do at the time of surveillance biopsy. A blood test with very high negative predictive value can be clinically valuable even if it is not perfect as a stand-alone rule-in test. In transplant surveillance, ruling out rejection safely is often more important than maximizing positive predictive value, because the current default is repeated invasive testing.

The separate assessment of ACR and AMR is also important. These syndromes differ immunologically and therapeutically. A biomarker strategy that can distinguish rejection subtype is likely to be more useful than a single omnibus “rejection yes/no” assay. It may also prove useful in patients with nonspecific symptoms, equivocal pathology, or concern for mixed rejection.

In addition, the prognostic association with later graft dysfunction or death raises the possibility that microRNA profiles are not merely diagnostic markers but also indicators of the allograft’s immunobiologic trajectory. This could eventually support risk-adapted surveillance intervals or earlier therapeutic intervention.

Biological Plausibility

The results are biologically credible. Circulating microRNAs can derive from activated immune cells, injured endothelial cells, and damaged cardiomyocytes. ACR is characterized by lymphocytic infiltration and myocyte injury, whereas AMR more strongly involves endothelial activation, complement engagement, and microvascular inflammation. Distinct microRNA signatures for these processes are therefore plausible. The ability of separate panels to identify ACR and AMR supports the idea that peripheral blood can capture meaningful differences in underlying graft immunopathology.

From a translational perspective, the study also fits with the broader trend in transplantation toward minimally invasive, molecularly informed surveillance. Gene-expression profiling and donor-derived cell-free DNA have already demonstrated that blood-based assays can improve post-transplant monitoring. MicroRNA assays may add complementary information, particularly if they are better able to capture rejection subtype or vascular injury biology.

Strengths of the Study

Several features strengthen confidence in the findings. The multicenter prospective design reduces the risk that the results are entirely center-specific. Samples were obtained longitudinally and paired with biopsy, which is the appropriate clinical reference standard in current practice. The investigators used previously identified microRNA sets rather than performing a purely exploratory discovery analysis in the same dataset, which helps reduce overfitting concerns. The study also reported clinically interpretable measures such as sensitivity, specificity, and negative predictive value at a usable threshold.

The inclusion of prognostic modeling is another strength. Diagnostic biomarkers are more compelling when they also relate to subsequent hard outcomes, and here both ACR and AMR scores were associated with later rejection, graft dysfunction, or death.

Important Limitations

The findings should still be interpreted with appropriate caution. First, the number of rejection events was relatively small, especially for ACR with only 14 episodes. Small event counts can inflate uncertainty around performance estimates, even when AUC values are high. Second, biopsy itself is an imperfect reference standard. If microRNA scores detect injury missed by biopsy, standard diagnostic metrics may both underestimate and complicate interpretation of true assay performance.

Third, the abstract does not provide the full details needed to assess calibration, subgroup performance, or robustness across time from transplant, immunosuppression regimens, hemodynamic status, infection, renal dysfunction, or other inflammatory confounders. These issues matter because blood-based immune biomarkers can be influenced by nonrejection processes.

Fourth, implementation questions remain unresolved. Sequencing-based microRNA analysis may be less immediately deployable than simpler assays unless translated into standardized, reproducible clinical platforms. Preanalytic variables, normalization methods, turnaround time, and interlaboratory reproducibility will be central to clinical adoption.

Finally, the study supports potential reduction in EMB use, but it does not by itself establish a safe biopsy-free management strategy. That will require prospective interventional trials comparing biomarker-guided surveillance with standard biopsy-based protocols.

Implications for Practice and Research

For clinicians, the most immediate takeaway is that circulating microRNA CRS appears promising as a noninvasive screening tool for both ACR and AMR after heart transplantation. At present, the data are most supportive of a rule-out role, particularly because of the very high negative predictive values. In the near term, such an assay could be tested as an adjunct to existing surveillance frameworks, helping to identify which patients are least likely to benefit from routine biopsy.

For researchers and health systems, the next steps are clear. Independent external validation is essential, ideally in larger cohorts with more rejection events and broader demographic and clinical diversity. Comparative studies against other blood-based tools, including donor-derived cell-free DNA and gene-expression profiling, would clarify whether microRNA testing is additive, superior, or best used in combination. Implementation studies should also evaluate cost, logistics, patient acceptability, and whether biomarker-guided surveillance can safely reduce biopsy burden without increasing missed rejection or adverse outcomes.

There is also an opportunity to explore whether serial changes in CRS, rather than a single threshold, better reflect dynamic alloimmune risk. Longitudinal trajectories may prove especially useful in patients with previous rejection, donor-specific antibodies, hemodynamic instability, or graft dysfunction of uncertain cause.

Conclusion

This multicenter prospective study provides strong evidence that circulating microRNA-based clinical rejection scores can identify both acute cellular rejection and antibody-mediated rejection at the time of EMB with excellent diagnostic discrimination. Equally important, higher scores predicted a greater subsequent risk of rejection, allograft dysfunction, or death. The combination of noninvasiveness, subtype specificity, and high negative predictive value makes this approach highly attractive for future transplant surveillance strategies.

Still, the work should be viewed as an important translational advance rather than a final practice-changing step. Before routine adoption, the assay will need broader validation, standardization, and testing in biopsy-sparing clinical pathways. If those steps are successful, circulating microRNAs could become a meaningful part of precision monitoring after heart transplantation.

Funding and ClinicalTrials.gov

ClinicalTrials.gov identifier: NCT02423070.

The abstract provided does not specify detailed funding information. Readers should consult the full published article for complete funding and disclosure statements.

Citation

Goldberg JF, Bagchi P, Mercado A, Shah KB, Najjar SS, Tchoukina I, Rodrigo ME, Hsu S, Jang M, Kong H, Marboe CC, Berry GJ, Valantine HA, Agbor-Enoh S, Shah P, GRAfT Investigators. Identification of Heart Transplant Rejection Subtypes With Circulating MicroRNAs. Circulation. Heart failure. 2026-01-12;19(5):e013141. PMID: 41521923. URL: https://pubmed.ncbi.nlm.nih.gov/41521923/