Comparison of five prognostic scoring systems, the French-American-British (FAB) and World Health Organization (WHO) classifications in patients with myelodysplastic syndromes: Results of a single-center analysis.

We retrospectively studied 89 consecutive patients diagnosed with primary myelodysplastic syndrome (MDS) over a period of 10 years to (1) identify prognostic factors for overall survival (OS) and leukemia-free survival (LFS); (2) to assess and compare the Bournemouth-, Spanish-, Düsseldorf-, Lille-, and the International prognostic scoring systems (IPSS); and to (3) compare the French-American-British (FAB) and World Health Organization (WHO) classifications. The median age of patients was 63 years (range, 26-85). Karyotype analyses were done in 85 patients (96%). Median OS was 3 years; 67 patients (75%) have died, and 28 (31%) had progression to acute myeloid leukemia (AML). Major independent prognostic variables for both OS and LFS (multivariate analysis) were percentage of bone marrow (BM) blasts (P < 0.0001), and in patients with cytogenetic data available, cytogenetic risk groups by Lille-score (OS, P = 0.031/LFS, P = 0.002) and IPSS (OS, P = 0.024). All five prognostic scoring systems successfully discriminated risk groups as regards OS and LFS, but in patients with cytogenetic data available, the major independent prognostic score for OS (P < 0.0001) and LFS (P = 0.006) was the IPSS. The FAB and WHO classifications also successfully discriminated between risk groups. The new WHO subgroups [refractory cytopenia with multilineage dysplasia (RCMD), with (RCMD-RS) or without ringed sideroblasts] showed a significantly (P = 0.0454) different prognosis for OS, but not for LFS (P = 0.0839), in comparison to the subgroups having erythroid dysplasia only (RA/RARS). Risk stratification into refractory anemia with excess blast-I (RAEB-I) and RAEB-II tended to yield different prognoses for OS and LFS. The 5q-minus syndrome strongly predicted for a good prognosis. In patients treated with the demethylating agent decitabine (n = 24), IPSS "poor risk" cytogenetics were unable to predict for the expected worse prognosis when compared to "intermediate-risk" cytogenetics. In conclusion, we confirm in a single-center patient cohort that the use of the WHO classification improves the predictive value of the FAB classification and that, in patients with cytogenetic data available, the IPSS can be used for clinical decision-making.