Allogeneic stem cell transplantation for myelofibrosis in the modern era: Single-center outcomes with DIPSS risk...

To describe the post-transplant outcomes in patients with myelofibrosis stratified by the Dynamic International Prognostic Scoring System (DIPSS) risk at transplantation, and to identify clinical factors associated with overall survival (OS) after allogeneic hematopoietic stem cell transplantation (allo-HSCT).

We retrospectively analyzed the data of 42 patients with myelofibrosis who underwent allo-HSCT at Samsung Medical Center between 2014 and 2023. OS was estimated using the Kaplan-Meier method and compared using the log-rank test. Exploratory Cox proportional-hazard regression analyses were performed to assess independent associations with OS.

At transplantation, 28 patients (66.7%) had Intermediate (Int-1/2) and 14 (33.3%) had High DIPSS risk. OS differed significantly by DIPSS risk group (log-rank P = 0.0034): the median OS was 15.2 months (95% confidence interval [CI], 7.2-NR months) in the Intermediate-risk group versus 6.4 months (95% CI, 3.3-35.1 months) in the High-risk group. All 14 patients in the High-risk group (100%) and 14 of 28 patients in the Intermediate-risk group (50%) died during follow-up. The 1-, 2-, and 3-year OS rates in the Intermediate- and High-risk groups were 57.1% vs. 28.6%, 50.0% vs. 14.3%, and 50.0% vs. 7.1%, respectively. In an exploratory multivariable analysis adjusted for hematopoietic cell transplantation-specific comorbidity index and donor type, DIPSS High-risk disease remained independently associated with inferior OS (adjusted hazard ratio [HR], 2.91; 95% CI, 1.27-6.63; P = 0.011). No other clinical variable, including age (≥ 60 vs. < 60 years; P = 0.16 by log-rank), achieved statistical significance.

DIPSS High-risk disease at transplantation was associated with uniformly poor post-transplant survival in this single-center cohort. These hypothesis-generating findings underscore the importance of transplant timing before progression to high-risk disease, and warrant validation in larger multicenter studies.