Allogeneic hematopoietic stem cell transplantation (HSCT) is the only potentially curative therapy for myelodysplastic syndromes (MDS). However, the procedure has a significant risk of morbidity and mortality. Prior studies have shown that early HSCT improves survival in patients with higher-risk MDS based on the International Prognostic Scoring System (IPSS) and the Revised IPSS (IPSS-R), while deferring HSCT (with medical management as needed) is preferred for patients with lower-risk MDS.1-3 The recently developed IPSS-Molecular (IPSS-M), which incorporates molecular data and information regarding loss of heterozygosity at TP53, has been shown to improve prognostication compared to the IPSS-R. It remains unknown whether using the IPSS-M to select patients with MDS for early HSCT will improve outcomes.4
In this study, Cristina A. Tentori, MD, and colleagues employed mathematical modeling to develop a decision support system (DSS) to determine the ideal timing of HSCT for patients with MDS by using the IPSS-M score and age. The study used retrospective observational data from 7,118 adults with a diagnosis of MDS, divided into a training cohort with 4,627 patients and a validation cohort with 2,491 patients. First, the authors validated that IPSS-M categories independently predict both overall survival and relapse after HSCT, even when accounting for other factors known to affect HSCT outcomes, including age, disease status, and type of conditioning regimen. Next, the research team developed cause-specific flexible parametric survival models to study transition hazards for important events: evolution to acute myeloid leukemia, death without HSCT, relapse after HSCT, and death after HSCT. Based on these models, a DSS using age and IPSS-M score was created to estimate the restricted mean survival time (RMST) over an eight-year period based on the time to transplant. The methods are described in detail in a companion methodological paper.5
In both the training and validation cohorts, immediate HSCT was associated with an improved RMST in patients with moderate-high-, high-, and very-high-risk IPSS-M scores. Immediate HSCT was also associated with a longer RMST in patients with both monoallelic and biallelic inactivation of TP53. In contrast, a delayed HSCT approach was associated with an improved RMST in patients with low- and moderate-low-risk IPSS-M scores. These results demonstrate that IPSS-M scores should be incorporated to help in deciding whether patients with MDS should be offered an early HSCT. The authors next compared the results of the DSS using IPSS-M versus IPSS-R scores. By incorporating molecular data, the transplant recommendation (immediate vs. delayed HSCT) changed in 17% of cases, with a significant improvement in RMST (1.2 years; p=0.001) expected by using the IPSS-M compared to the IPSS-R.
In Brief
This study supports incorporating the IPSS-M to determine whether an individual with MDS will benefit from an immediate versus deferred approach to HSCT. In addition, a significant improvement in patient outcomes is expected by using the IPSS-M compared to the IPSS-R, demonstrating that molecular data adds important prognostic information that can help with HSCT decision-making. It should be noted, however, that the DSS recommends immediate HSCT for any patient where the lower bound of the 95% confidence interval (CI) surrounding the optimal timing of transplant is less than 12 months, even if the size of the CI is large, which should be taken into consideration when making decisions for individual patients.6 Also of note, patients over age 70 were not included in this analysis. As HSCT becomes more accessible to fit elderly patients, there will be a need for analyses of HSCT outcomes in individuals over age 70 with MDS that incorporate the IPSS-M.
Disclosure Statement
Dr. McMahon indicated no relevant conflicts of interest.
