Molecular warning-driven proactive intervention strategies for antibody-mediated rejection in kidney transplantation

The traditional “post-injury salvage” model, which relies on a rise in serum creatinine followed by biopsy confirmation, frequently leads to irreversible graft injury. This narrative review integrates the immunopathological mechanisms of antibody-mediated rejection (ABMR) with the limitations of current diagnostic approaches, and proposes a shift in the trigger for clinical decisions from “organ dysfunction” to “molecular signal abnormality.” We organize the review around a Decision Algorithm that follows the logic of pathological mechanisms and diagnostic gaps, biological interpretation of molecular signals, risk stratification, biopsy and treatment decisions, and dynamic monitoring of treatment response. Key biomarkers include donor-derived cell-free DNA (dd-cfDNA), functional donor-specific antibodies (DSA) with C1q-binding status and IgG subclass, and urinary CXCL10. We emphasize that dd-cfDNA should be interpreted as a dynamic injury signal rather than a static threshold, incorporating serial changes, patient-specific baseline, magnitude of change, and duration of elevation. The algorithm explicitly distinguishes pre-existing, de novo , and persistent DSA, and incorporates DSA origin, time of appearance, persistence, changes in MFI, and complement-binding status. Non-HLA antibodies are presented as complementary or exploratory markers in DSA-negative cases, with uncertainties in assay standardization, clinical thresholds, and causal relationship to tissue injury. We present a four-tier risk stratification system with corresponding monitoring and treatment decision pathways. Clinical action plans based on defined biomarker thresholds are specified, with explicit labeling of which thresholds are supported by prospective/multicenter evidence and which are proposed as a future management framework. Targeted therapies—IL-6 blockade, complement-targeted therapy, and FcRn inhibitors—are graded by evidence level (randomized trials, observational studies, case series, or hypothesis-generating proposals) and should be used individually according to biomarker-driven stratification, preferably within clinical trials. Although challenges remain in standardization, cost, accessibility, and the lack of prospective interventional validation, an integrated diagnostic framework combining multi-omics, machine learning, and molecular biopsy holds promise for achieving personalized and prospective post-transplant immune management. This approach may interrupt the chain of injury before fibrosis develops and improve long-term graft survival.