Decision Analysis of Allogeneic Stem Cell Transplantation in Patients with Myelodysplastic Syndrome Stratified According to the Revised International Prognostic Scoring System (IPSS-R)

Myelodysplastic syndromes (MDS) are heterogeneous myeloid neoplasms that range from indolent conditions to forms very close to acute leukemia (AML). The International Prognostic Scoring System (IPSS) is the benchmark for clinical decision-making. Recently, the International Working Group for Prognosis in MDS (IWG-PM) defined 5 new cytogenetic risk groups on a large cohort of patients representing the basis for the revised IPSS (IPSS-R) together with refined categories for marrow blasts and cytopenias (Blood. 2012;120:2454-65). IPSS-R categorization enables more accurate definition of the prognosis of MDS patients, in particular those in low or intermediate-1 IPSS risk groups. Despite recent progress, the only curative treatment for MDS remains allogeneic stem cell transplantation (allo-SCT), which however involves a non-negligible risk of mortality/morbidity. Allo-SCT is not considered in patients with very low IPSS-R risk, whose median survival is 9 years. Conversely, eligible patients with high or very high IPSS-R risk, whose survival is 1–2 years, should be offered immediate transplantation. Uncertainty exists about the optimal timing of allo-SCT in the low and intermediate IPSS-R risk groups, and to address this issue we performed an ad hoc decision analysis. We analyzed two cohorts of MDS patients (age 18-70 years): i) 477 patients diagnosed with MDS at the Policlinico S. Matteo, Pavia, Italy between 2000–2010, who mostly received best supportive care; ii) 488 patients who received allo-SCT in the same period reported to the Gruppo Italiano Trapianto di Midollo Osseo (GITMO) registry. We adopted a continuous time multi-state Markov approach to model the natural history of the disease. Each IPSS-R risk group is represented by a state, and transition is allowed to the next state (to AML from very high risk) or to death. A transition intensity (i.e., instantaneous risk of moving to another state) is then estimated for each possible transition. Allo-SCT is modeled as a time-dependent covariate with 3 levels: i) no allo-SCT; ii) transplantation done no more than 3 months before; iii) transplantation done more than 3 months before. The effect of allo-SCT on survival was estimated by hazard ratios (HR) with respect to the no allo-SCT level. We examined two different policies: policy A) perform allo-SCT after t months since entering the low IPSS-R risk state or immediately in case of disease progression before the planned delay time t; policy B) perform allo-SCT after t months since entering the intermediate IPSS-R risk state or immediately in case of disease progression before the planned delay time t. For each policy, the expected survival of patients with different age at diagnosis was calculated. According to the model, a patient with a low IPSS-R risk is expected to spend 2.13 years in this risk group, and 2.39 years in the intermediate one. The cumulative incidence of progression from low to a higher IPSS-R risk was 27% at 3 years, while that of progression from intermediate to a higher risk was 39% at 3 years. Overall, expected survival times from diagnosis were higher under policy B vs policy A. Expected survival times were up to three years greater under policy B with respect to policy A for younger patients. When the waiting time before transplantation was progressively increased from 0 to 60 months, there was an increase in expected survival under policy A (in younger patients this increases by about two further years if transplant is delayed until 60 months in low IPSS-R risk state). Conversely, under policy B, delaying transplant further reduces expected survival, by up to about three years with a delay of 60 months after entering intermediate risk state or further progression to high risk state. In summary, this study suggests that a delayed transplantation strategy is advisable for patients with early disease, i.e. very low and low IPSS-R risk, whereas no further delay is advised for patients belonging to intermediate or higher risk categories. When indicated, setting the timing of intervention immediately after disease progression makes such a strategy easy to implement in the clinical practice. These results provide clinicians with a base of evidence to maximize the effectiveness of allo-SCT in MDS patients.