Older patients with acute myeloid leukemia or myelodysplastic syndromes who undergo reduced-intensity stem cell transplants face a persistent threat: relapse. A retrospective study from two Korean transplant centers now reports that pairing a sensitive sequencing-based residual disease test with donor chimerism assessment at three months after transplantation sharply separates patients into groups with dramatically different outcomes, from an 85.6 percent two-year survival rate down to 13.6 percent.
The research, published in Blood Research, analyzed 113 patients with AML or MDS who were aged 60 or older, or 55 or older if judged unfit for standard-intensity conditioning because of serious comorbidities, major organ dysfunction, impaired performance status, or a high hematopoietic cell transplantation comorbidity index score. All received reduced-intensity conditioning allogeneic stem cell transplants at Soonchunhyang University Bucheon Hospital or Soonchunhyang University Seoul Hospital between January 2018 and May 2025. The median age was 63 years, roughly two-thirds of patients had AML, and nearly half received grafts from haploidentical donors.
Because eight patients died before day 90, before a three-month sample could be obtained, the investigators performed a landmark analysis restricted to the 105 patients who survived to that time point, with survival time counted from day 90 onward. This design minimizes a statistical distortion known as immortal time bias, which arises when patients must survive long enough to have a biomarker measured. The landmark cohort contained 29 deaths, which constrained how many variables the multivariate models could accommodate.
Residual disease was measured with targeted next-generation sequencing, a technique the authors argue offers superior sensitivity to quantitative PCR or flow cytometry. Bone marrow or peripheral blood samples collected three months after transplantation were sequenced to a mean depth of approximately 1,000-fold, and positivity was defined as a variant allele frequency of at least 0.2 percent for a mutation present at diagnosis at 3 percent or higher. Forty-nine patients were tested with an earlier 45-gene panel from a national research project and 56 with a later 74-gene institutional panel, with the same threshold applied throughout. Chimerism was assessed by short tandem repeat analysis of 16 loci, with complete donor chimerism defined as the absence of recipient-specific alleles above the assay’s 1 percent detection limit.
The results were striking. At three months, 28 of the 105 evaluable patients, or 26.7 percent, tested positive for measurable residual disease, while 82 patients, or 78.1 percent, had achieved complete donor chimerism. In univariate analysis, residual disease positivity predicted inferior overall survival with a hazard ratio of 6.373, and complete donor chimerism predicted improved survival with a hazard ratio of 0.356. On multivariate analysis restricted to four prespecified variables, residual disease positivity remained an independent predictor of death with a hazard ratio of 3.315, whereas complete donor chimerism showed only a trend toward benefit that did not reach statistical significance, with a hazard ratio of 0.494 and a P value of 0.051. A formal interaction test between the two markers did not reach significance, and in that expanded model residual disease retained significance while chimerism did not.
When patients were divided into four groups based on both markers, the differences were dramatic. Two-year overall survival ranged from 85.6 percent among patients who were residual disease negative with complete donor chimerism to 13.6 percent among those who were both residual disease positive and had mixed chimerism. Two-year cumulative relapse incidence followed the same gradient: zero percent, 8.4 percent, 53.7 percent, and 69.7 percent across the four groups, with the double-high-risk group faring worst. The authors concluded that residual disease status, rather than chimerism alone, was the dominant determinant of relapse and survival, while loss of complete donor chimerism further identified a subset of residual disease positive patients at the highest risk.
Chimerism dynamics added nuance. Among 23 patients with mixed chimerism at three months, 15 subsequently converted to complete donor chimerism, and the conversion rate depended heavily on concurrent residual disease status: 11 of 12 residual disease negative patients converted, compared with only 4 of 11 residual disease positive patients. Patients who converted had superior two-year overall survival compared with those who remained mixed, at 48.2 percent versus 22.6 percent. Donor type, whether matched related, matched unrelated, or haploidentical, showed no significant association with early chimerism status.
A complicating factor in sequencing-based monitoring is the interpretation of mutations in DNMT3A, TET2, and ASXL1, genes commonly mutated in age-related clonal hematopoiesis that can persist after transplantation without signaling relapse. Current recommendations favor excluding patients whose positivity rests solely on these so-called DTA mutations. In a sensitivity analysis, the researchers reclassified five such patients as residual disease negative; the graded pattern of relapse and survival across subgroups persisted, and residual disease positivity remained independently associated with inferior survival. Notably, among the 28 residual disease positive patients, the five with DTA-only positivity did not differ in survival from the other 23, although the authors caution that three of the five carried complex karyotype or TP53 mutations, poor-risk features that may have driven their outcomes and limited the comparison.
The study also examined whether post-transplant interventions could alter the trajectory. Roughly 70 percent of residual disease positive patients received therapy, including azacitidine, venetoclax, FLT3 inhibitors, or donor lymphocyte infusion, typically initiated two to three months after transplantation. However, the effect on survival remained inconclusive, with no statistically significant differences in the subgroup comparisons. The authors attribute this to limited treatment duration, a median of four cycles, small subgroup sizes, and the possibility that preemptive therapy triggered by later molecular relapse still introduces residual timing bias that only a time-dependent analysis anchored at therapy initiation could fully resolve.
The investigators emphasize that their findings are observed associations from a retrospective, two-center study without external validation, and that prospective confirmation is required before the approach can inform clinical practice. Other limitations include heterogeneity in disease type and status at transplant, evolving sequencing platforms across the study period that could not be formally stratified, and reliance on unfractionated rather than lineage-specific chimerism testing, meaning the results reflect overall donor engraftment rather than donor T-cell activity. Surveillance was also constrained by reimbursement rules that limited sequencing-based testing to scheduled bone marrow evaluations at three and twelve months, which may have missed clinically silent relapses. Still, the authors argue that routine incorporation of sequencing-based residual disease monitoring alongside donor chimerism analysis at the three-month landmark offers older transplant recipients a practical early warning system, allowing interventions such as immunosuppression tapering, donor lymphocyte infusion, or targeted maintenance to begin before overt hematologic relapse occurs.
Subject of Research: Prognostic value of combined next-generation sequencing residual disease and donor chimerism monitoring after reduced-intensity transplantation in older patients with AML or MDS
Article Title: Post-transplant monitoring of NGS-based measurable residual disease and donor chimerism as a prognostic tool in older patients with AML/MDS
Article References: Kim, S. H., Yun, J., Kim, K. H., Yoon, S. Y., Kim, C. K., & Park, S. K. (2026). Post-transplant monitoring of NGS-based measurable residual disease and donor chimerism as a prognostic tool in older patients with AML/MDS. Blood Research, 61(1), Article 68. https://doi.org/10.1007/s44313-026-00187-6
Image Credits: AI Generated
DOI: 10.1007/s44313-026-00187-6
Keywords: acute myeloid leukemia, myelodysplastic syndromes, allogeneic stem cell transplantation, measurable residual disease, donor chimerism, next-generation sequencing, reduced-intensity conditioning, relapse risk, older patients, prognostic biomarkers, donor lymphocyte infusion, clonal hematopoiesis
News Source: Nathaniel Bowman. (October 27, 2026). Blood test duo at three months flags relapse risk after transplants in older leukemia patients. Scienmag.
