Skip to Content


New Prognostic Classifications of MDS

Leukemia Insights - Winter 2008

One of the main problems in treating patients with MDS is the difficulty of its diagnosis and therefore classification. At M. D. Anderson, we have documented that in approximately 15% to 30% of patients referred to us there is a discrepancy with the outside diagnosis. In most instances, patients have acute myelogenous leukemia (AML) instead of MDS but patients with lower risk disease are also seen, as well as others with other conditions such as lymphomas, metastatic solid tumors and even no evidence of any pathological change at all.

Therefore proper diagnosis is key when assessing these patients. Diagnosis of MDS requires a team of expert pathologists for the manual counting of blasts in the bone marrow, cytogenetics, and in some cases flow cytometry. Once the diagnosis is made, the question is what is the best approach: observation or initiation of therapy? Obviously this decision depends on the “risk” of the patient. Until recently most of the currently available classifications (FAB, WHO, or even the IPSS) were not accurate tools to predict the clinical behavior of patients with MDS, particularly those with lower risk disease. In the past at M. D. Anderson, many patients with MDS (those with blasts less than 10%) were referred back for “observation” only and therapy was recommended at the time of progression. We recently reviewed this practice in an analysis of the prognosis of patients with low or intermediate-1 risk disease by the IPSS referred to our center over the last 25 years1. We observed that most of these patients had a poor prognosis with a median survival in general less than 24 months (figure 1). Importantly, only 10% of these patients transformed to AML and it appears that most patients died from MDS related complications (i.e. infections, bleeding). This data indicates that “early intervention” in MDS may be beneficial and that specific therapeutic alternatives are needed for this group of patients that constitutes close to 70% of all patients with MDS.

Figure 1. Survival of Patients with lower risk MDS based on MDACC low risk model

MDS Survival Graph

Another important issue is the fact that technically, the IPSS score was applied at the time of initial diagnosis, a moment that is not always the time when therapeutic decisions are made. To overcome this problem, we have developed a new comprehensive classification2 for all patients with MDS that we hope will become standard in MDS (Table 1). The other major advantage of this new system is that it does not require WHO assessment of the disease. A table to predict overall and 4-year survival is shown in table 2.

Table 1. M. D. Anderson Low Risk MDS Model

Adverse Factor


P value

Assigned Score

Unfavorable cytogenetics


< 0.0001


Age >/= 60 years


< 0.0001


Hgb < 10(g/dL)


< 0.0001


Plt < 50 x 109/L
50-200 x 109/L




BM blasts >/= 4%




© 2014 The University of Texas MD Anderson Cancer Center