Upfront vs posttherapy allogeneic stem cell transplantation for high-risk myelodysplastic syndrome – a real-world analysis in the era of molecular risk stratification

Abstract

Abstract Background: For patients with high-risk myelodysplastic syndromes (MDS) or borderline acute myeloid leukemia (AML), the timing of allogeneic stem cell transplantation (ASCT) remains a matter of debate. Higher bone marrow (BM) blast counts (>10%) are associated with worse outcomes, often prompting cytoreductive therapy before transplantation. Recent EBMT registry data (Scheid et al., Blood 2024) challenge this approach, showing minimal benefit from blast downstaging using the IPSS-R. We aimed to validate these findings in a single-center cohort using the newer IPSS-M molecular risk classification. Methods: We retrospectively analyzed 129 consecutive patients with high-risk MDS or borderline AML undergoing first ASCT between 01/2013 and 07/2024. Patients were grouped as frontline ASCT (n = 88) or pre-treated with chemotherapy or hypomethylating agents prior to ASCT (n = 41). Median age was 62 vs. 64 years, median BM blasts at diagnosis were 12% vs. 10% and 12% and 7% before transplant. Median time from diagnosis to ASCT was 6.7 vs. 23.1 months, respectively. The median IPSS-R score was 5 in both groups, while median IPSS-M values were 0.64 in the upfront group and 0.90 in the pre-treatment group. Conditioning regimens were predominantly reduced-intensity (Flu/Treo or Flu/Bu). OS and progression-free survival (PFS) were estimated by Kaplan–Meier and compared with the log-rank test. Results: With a median follow-up of 23 months, median OS was 123.9 months for frontline ASCT vs. 44.4 months for pre-treated patients (p = 0.17). Median PFS was 71.0 vs. 27.7 months (p = 0.096). The proportion of patients whose IPSS-M improved from diagnosis to transplant was 26.1% in the frontline group and 34.1% in the pre-treated group, whereas worsening occurred in 29.5% and 31.7%, respectively. No major differences were observed in engraftment kinetics, graft source (predominantly PBSC), donor type, or HLA match distribution. GvHD-relapse-free survival was increased in the pre-treated patients group (p = 0,029). Pre-transplant changes in IPSS-M, whether improvement, stability, or worsening, were not associated with significant differences in OS, PFS, or GRFS. Conclusion: In this real-world cohort, pre-transplant cytoreduction was not associated with superior survival compared to upfront transplantation, despite modest improvements in IPSS-M scores in some pre-treated patients. These findings add to growing evidence that, in high-risk MDS and borderline AML, delaying ASCT for blast reduction may not confer meaningful clinical benefit. Early referral and transplantation should be considered, particularly in the molecularly defined high-risk population.