Myelodysplastic syndromes (MDS) are myeloid neoplasms that present as refractory cytopenia and typically occur in elderly people . Because of population aging, MDS represent one of the most common hematologic malignancies in Western countries nowadays, their annual incidence exceeding 20 per 100,000 persons over the age of 60 years .
MDS are highly heterogeneous from a clinical point of view, ranging from conditions with a near-normal life expectancy to forms approaching acute myeloid leukemia (AML) . This clinical heterogeneity reflects different somatic mutations responsible for clonal proliferation of myelodysplastic stem cells [3-5]. In particular, spliceosome mutations, implying abnormalities of mRNA splicing in the pathogenesis of MDS, have variable prognostic relevance, depending on the mutated splicing factor [6-9].
From a practical point of view, the definition of individual risk requires the use of prognostic models. In 1997, Greenberg et al. developed the International Prognostic Scoring System (IPSS) , which has rapidly become a benchmark for clinical decision-making, clinical trials, and drug approval. Despite its usefulness, this scoring system has weaknesses , and has been recently revised with the development of the IPSS-R . Following the introduction of the World Health Organization (WHO) classification of MDS , a WHO classification-based Prognostic Scoring System (WPSS) was defined, based on WHO categories, cytogenetic abnormalities, and transfusion-dependency [14, 15]. WPSS proved to be able to provide dynamic prognostic information at any time of the disease course, and to predict the outcome of allogeneic hematopoietic stem cell transplantation (HSCT) .
Despite recent therapeutic progress, the only curative treatment for MDS patients remains allogeneic HSCT, which is considered as a conventional therapeutic option until the age of 65 in eligible patients. Its efficacy, however, is considerably limited by morbidity and mortality, resulting in a long-term survival rate of about 30% . Several issues must be taken into account when considering allogeneic HSCT and evaluating its benefits in the individual patient with MDS, and a crucial question is timing of transplantation . Considering the clinical course of MDS, the optimal timing of allogeneic HSCT would be a disease stage that provides the best overall life expectancy, accounting for both pretransplantation and posttransplantation survival. In fact, patients at early stages, especially those with a somatic mutation of SF3B1, may experience long periods with stable disease after diagnosis [2, 6], and the risks of morbidity and mortality related to allogeneic HSCT would be unacceptably high for many of them.
However, a number of studies have shown that advanced disease stage at transplantation is associated with inferior overall survival [16, 17]. In particular, a recent study of the European Group for Blood and Marrow Transplantation clearly showed that advanced disease stage at transplantation was the major independent variable associated with an inferior 4-year overall survival in MDS patients 50 years or older . However, a previous decision analysis by the International Bone Marrow Transplant Registry (IBMTR) concluded that, whereas immediate transplantation was associated with maximal life expectancy in patients with intermediate-2- and high IPSS risk, for those with low and intermediate-1 IPSS-risk delayed transplantation offered optimal survival benefit . It was therefore concluded that the optimal timing of transplantation was at the time of disease progression from lower to higher IPSS risk groups. This study has substantially influenced clinical practice despite a number of intrinsic limitations . In particular, the IBMTR analysis considered patients at the time of MDS diagnosis, ignoring changes in their disease status that frequently occur before transplantation or leukemic evolution.
To overcome the above limitations, we carried out an ad hoc decision analysis in MDS patients aged up to 65 years, classified according to WHO criteria and stratified according to either the IPSS or WPSS. We used a continuous-time multistate Markov model to describe the natural history of disease and evaluate the effect of allogeneic HSCT on survival at different stages of disease.