The research team, led by Su-Hyeon Lee and Deok-Gie Kim, with corresponding authors Byungjin Hwang and Dong Jin Joo, took an unusually rigorous approach to a difficult clinical problem. They recruited 20 liver transplant recipients and collected paired samples of peripheral blood mononuclear cells, the circulating immune cells that include lymphocytes and monocytes, both before surgery and after transplantation. Ten of these patients went on to develop acute rejection and ten did not. Crucially, the two groups were matched for a battery of clinical variables that could otherwise confound the analysis, including age, sex, hepatocellular carcinoma status, autoimmune disease status, ABO blood group compatibility, and donor age. By controlling for these factors, the investigators could be more confident that the immunological differences they found were genuinely linked to rejection rather than to incidental patient characteristics.

Technically, the study is a tour de force of single-cell multi-omics. Rather than averaging gene expression across bulk tissue, the team used single-cell RNA sequencing to profile 88,940 high-quality individual cells, preserving the identity and state of each immune cell. They paired this transcriptomic readout with T cell receptor sequencing, which reads the unique genetic rearrangements that give each T cell its antigen-specific receptor. This combination allowed them to track not only which cell types were present and what genes they were expressing, but also how T cell clones expanded, contracted, and diversified over time. On top of the sequencing data, the researchers layered computational analyses of gene regulatory networks and intercellular communication, reconstructing the signaling conversations between immune cell populations that precede and accompany rejection.

The most striking finding is that the immune systems of future rejecters looked different before transplantation. Patients who later developed acute rejection already showed a distinct pre-transplant immune state compared with those who tolerated their new livers. Natural killer cells in the rejection group displayed an inflammatory activation profile with enhanced responsiveness to interferon, a family of signaling proteins central to antiviral and immune defense responses. Monocytes, the innate immune cells that patrol the blood and seed tissues, showed altered composition and transcriptional programs dominated by interferon-associated signatures. Together, these observations suggest a systemic, low-grade inflammatory readiness in the circulation of patients whose immune systems would later turn against the graft.

The adaptive immune side of the story was equally revealing. Before surgery, recipients who went on to reject showed a skewing of their CD8 T cell compartment toward effector memory states. Effector memory CD8 T cells are the rapid-response soldiers of the immune system: they circulate in the blood and peripheral tissues, and when they encounter their target antigen they can immediately deploy cytotoxic machinery without the delay required for naive T cells to differentiate. A pre-existing bias toward this compartment means the patient carries a standing army of cytotoxic cells primed for fast action, a configuration that could plausibly accelerate an attack on donor tissue once transplantation delivers a flood of foreign antigens.

After transplantation, the divergence between the two groups did not close; it widened in specific ways. Despite receiving standard immunosuppression, rejection patients maintained heightened overall immune activity. The team observed an expansion of KIR-expressing effector memory CD8 T cell populations in these patients. KIR, or killer cell immunoglobulin-like receptors, are molecules classically associated with natural killer cells that tune their reactivity to self, but their expression on CD8 T cells marks a subset with distinctive effector properties. The rejection group also showed increased expression of IRF1, a transcription factor that sits downstream of interferon signaling and orchestrates inflammatory gene programs in CD8 T cells. In other words, the post-transplant immune landscape of rejecters retained an activated, interferon-driven character that immunosuppression failed to fully extinguish.

To move from descriptive biology toward clinical utility, the researchers searched for genes whose expression reliably distinguished rejecting from non-rejecting patients across multiple time points. They identified 15 robust acute rejection-associated genes that were consistently detected before and after transplantation, and then validated a subset of these genes in liver biopsy tissue from the patients. This cross-validation is important: it links the circulating blood signature to the actual immunological battleground inside the graft, supporting the idea that a peripheral blood test could serve as a proxy for processes occurring in the transplanted liver. A blood-based biomarker panel built on such genes could, in principle, flag high-risk patients before rejection becomes clinically manifest, allowing clinicians to intensify monitoring or tailor immunosuppressive regimens without resorting to repeated biopsies.

The implications reach beyond liver transplantation. Acute cellular rejection is fundamentally a problem of immune recognition and memory, and the Yonsei study demonstrates that the relevant information is encoded in the composition and activation state of circulating immune cells. The finding that pre-transplant immune states predispose recipients to rejection challenges the traditional view of rejection as a purely post-transplant event driven by donor-recipient mismatch alone. Instead, the recipient’s baseline immune configuration, shaped by their disease history, viral exposures, and inflammatory milieu, appears to set the stage for how the immune system will respond to the graft. This reframing aligns with a broader trend in transplant immunology toward precision medicine, in which immunological risk stratification replaces one-size-fits-all immunosuppression.

There are, of course, important caveats. The study involved 20 carefully matched patients, a sample size appropriate for discovery but too small to support immediate clinical deployment of a diagnostic test. The 15-gene signature and the cellular features identified will need validation in larger, independent, and ideally multi-center cohorts before they can inform treatment decisions. It also remains to be seen whether the signatures generalize across different donor types, immunosuppressive protocols, and causes of liver failure. Nonetheless, the methodological framework, integrating single-cell transcriptomics, T cell receptor repertoire tracking, regulatory network inference, and tissue validation, provides a template that other transplant programs can adopt to build the evidence base needed for clinical translation.

For patients awaiting a liver, the promise is tangible: a pre-operative blood draw that reveals whether their immune system is primed for conflict, and post-operative monitoring that catches rejection in its earliest molecular stages rather than after tissue damage has occurred. For the field of transplant immunology, the study adds a detailed cellular and molecular atlas of how rejection unfolds in the human circulation, from interferon-primed natural killer cells and interferon-signature monocytes to KIR-expressing effector memory CD8 T cells and IRF1-driven inflammatory programs. As single-cell sequencing becomes faster and cheaper, the vision of routine immune profiling before and after organ transplantation moves closer to reality, and with it the possibility of preventing rejection rather than merely treating it.

Subject of Research: Single-cell RNA sequencing and T cell receptor repertoire analysis of immune signatures associated with acute liver allograft rejection

Article Title: Integrative single-cell RNA-seq and TCR repertoire analysis reveals distinct immunological features of acute liver allograft rejection

Article References: Lee, S.-H., Kim, D.-G., Cho, Y., Min, E.-K., Lee, J. G., Kim, M. S., Park, S., Hwang, B., & Joo, D. J. (2026). Integrative single-cell RNA-seq and TCR repertoire analysis reveals distinct immunological features of acute liver allograft rejection. Genome Medicine. https://doi.org/10.1186/s13073-026-01753-4

Image Credits: AI Generated

DOI: 10.1186/s13073-026-01753-4

Keywords: liver transplantation, acute rejection, single-cell RNA sequencing, T cell receptor repertoire, CD8 T cells, natural killer cells, monocytes, interferon signaling, IRF1, biomarkers, immunosuppression, Genome Medicine

The study, published in the Journal of Translational Medicine, drew on data from the PET-BRIGHT clinical trial, in which thirty-five patients with metastatic non-small cell lung cancer received the standard combination of carboplatin, pemetrexed, and the immunotherapy pembrolizumab. Each patient underwent a FDG-PET/CT scan and a blood draw at two moments: before treatment began, and again after just three weeks, a single cycle of therapy. From these timepoints the researchers extracted three families of biomarkers: imaging metrics that quantify how much glucose metabolically active tumor tissue consumes, measures of T-cell receptor diversity that capture the breadth of the immune system’s anti-cancer repertoire, and a panel of inflammatory cytokines circulating in the blood.

The choice of these three modalities reflects a growing recognition that no single biomarker can capture the complexity of immunotherapy response. FDG-PET imaging measures uptake of a radioactive glucose analog, revealing where and how aggressively tumor cells are metabolizing energy, with summary metrics such as standardized uptake value and total lesion glycolysis summarizing whole-body tumor burden. T-cell receptor sequencing reads out the diversity of immune recognition, on the theory that a richer repertoire of tumor-targeting T cells signals a stronger response to immune checkpoint blockade. Cytokines, the small signaling proteins through which immune cells communicate, offer a systemic window into inflammatory states that may either support or suppress anti-tumor immunity.

Against these signals, the researchers benchmarked the current clinical standard: PD-L1 tumor proportion score, an immunohistochemistry measurement that guides treatment selection in lung cancer today. The results were sobering for the status quo. PD-L1 achieved an area under the receiver operating characteristic curve of just 0.58, barely better than a coin flip at distinguishing responders from non-responders. This weak discrimination, the authors note, is precisely what motivates a multimodal, uncertainty-aware approach, because patients and clinicians currently must commit to months of therapy armed with a biomarker that carries limited predictive weight.

The machine learning pipeline itself was deliberately conservative, a design choice appropriate for a small cohort. Using nested leave-one-out cross-validation with automated feature selection, the team identified a single representative biomarker per modality, guarding against the overfitting that plagues models trained on dozens of features and only a handful of patients. Classification was performed with class-balanced logistic regression, an interpretable method chosen over more opaque architectures, and performance was assessed both by discrimination, how well the model separated responders from non-responders, and by calibration, whether predicted probabilities matched observed outcomes as measured by the Brier score.

The headline innovation is the application of conformal prediction, a statistical framework that converts a model’s raw output into prediction sets with formal guarantees. Rather than declaring that a patient will respond or not, the conformal framework targets a specified reliability level, in this case eighty percent, and returns either a confident single-class prediction or a set containing both possibilities, signaling that the model cannot confidently classify that patient. The proportion of confident singleton predictions therefore becomes a direct, interpretable measure of uncertainty at the level of the individual patient, not merely the population average that conventional accuracy metrics provide.

The performance results chart a coherent story about how biomarker information evolves over treatment. Before therapy began, T-cell receptor diversity was the strongest single modality, reaching an AUROC of 0.80, followed by PET imaging at 0.76, while cytokines lagged at 0.54. Combining PET and T-cell data through late fusion, in which separately trained modality-specific models are merged, pushed pre-treatment discrimination to 0.85, and raised the proportion of confident singleton predictions from 78 percent to 89 percent. After the three-week blood draws and scans were added, the hierarchy shifted: cytokines became the strongest single modality at 0.78 while T-cell metrics were attenuated at 0.66, and late fusion of PET and cytokines achieved the highest overall discrimination of 0.86. Thirteen of twenty-two model and timepoint combinations beat a permutation-based null at the conventional significance threshold, and empirical coverage landed at 78 percent and 82 percent, close to the eighty percent reliability target the conformal framework was set to honor.

That dynamic shift in which modality carries the most predictive weight is itself scientifically interesting. It suggests that pre-treatment immune repertoire breadth anticipates who will benefit from checkpoint blockade, but that once therapy begins, the inflammatory milieu measured in plasma becomes the more informative readout of whether the treatment is taking hold. A static biomarker panel frozen at baseline would miss this evolution entirely, and the study’s longitudinal design, with its week-three reassessment, demonstrates how a decision support system could update its confidence as new evidence accumulates over the first critical weeks of therapy.

The authors are appropriately careful about the limitations of what they have built. Thirty-five patients from a single center constitute a proof-of-concept feasibility study, and the findings are explicitly framed as hypothesis-generating rather than practice-changing. Small cohorts make any machine learning model vulnerable to optimistic bias, which is why the permutation testing and conservative feature selection matter so much here. Before such a system could inform real treatment decisions, it would need validation in larger, multi-institution cohorts spanning different tumor types, treatment regimens, and demographic populations, along with calibration checks to confirm that the conformal guarantees hold outside the development data.

Even so, the prototype points toward a future in which clinical AI systems do something more honest than issuing confident verdicts. By pairing multimodal biomarker fusion with formal uncertainty quantification, the framework identifies not only which patients are likely to respond, but also, and perhaps more valuably, those for whom no confident prediction can be made, flagging them for closer monitoring or alternative strategies. In a disease where the current standard biomarker performs barely better than chance, knowing the limits of prediction may prove as clinically important as the predictions themselves. The interface prototype developed alongside the statistical framework shows how such uncertainty could be surfaced directly to oncologists, turning a black-box risk score into a transparent, caveat-aware recommendation that clinicians can weigh against the realities of each patient’s care.

Subject of Research: Uncertainty-aware multimodal prediction of chemoimmunotherapy response in metastatic non-small cell lung cancer

Article Title: Toward uncertainty-aware clinical decision support for treatment response prediction in metastatic NSCLC: integrating FDG-PET, T-cell repertoire, and cytokines with conformal prediction

Article References: Yaseen, F., Hippe, D. S., Cui, S., Fu, J., Kim, Y., Grassberger, C., Deng, L., Ye, T., Kinahan, P. E., Zeng, J., Gennari, J. H., & Bowen, S. R. (2026). Toward uncertainty-aware clinical decision support for treatment response prediction in metastatic NSCLC: integrating FDG-PET, T-cell repertoire, and cytokines with conformal prediction. Journal of Translational Medicine. https://doi.org/10.1186/s12967-026-09030-z

Image Credits: AI Generated

DOI: 10.1186/s12967-026-09030-z

Keywords: metastatic non-small cell lung cancer, FDG-PET, T-cell receptor repertoire, cytokines, conformal prediction, clinical decision support, immunotherapy, PD-L1, biomarkers, machine learning, uncertainty quantification, treatment response prediction