International MDS Risk Analysis Workshop (IMRAW)/IPSS Re-Analyzed: Impact of Depth of Cytopenias on Clinical Outcomes in MDS.

The IPSS, created for evaluating clinical outcomes of MDS patients with IMRAW data combined from seven large risk-based studies, has been validated through multiple studies of MDS patients and has been proven useful in clinical decision-making. To determine whether the depth of cytopenias (in addition to their number) could further refine the prognostic ability of the IPSS, we evaluated this parameter in the 815 patients from the IMRAW database and assessed their correlation to patients’overall survival (OS) and risk of evolution to acute myeloid leukemia (AML). Univariate analyses of platelet (plt), hemoglobin (Hb), and absolute neutrophil count (ANC) depth levels were correlated with IPSS risk groups, FAB categories [refined as: RA, RARS, RAEB-1, RAEB-2, RAEB-T, CMML, del(5q) alone], cytogenetic groups, bone marrow blast percent (%), and age groups (≤60, >60 years).Our results showed that all cytopenic categories had statistically significant rank correlations with IPSS and marrow blast % (Tables 1 & 2, Kendall rank coefficient τ, p<0.0001; ANC data not shown) and also FAB categories (Kruskall test of correlation, p<0.0001). For example, as disease risk progressed from IPSS low to high, % patients with plt <20000/μl increased from 1 to 13%, and % patients with Hb ≤8g/dl increased from 11 to 47%. Similar trend was observed as marrow blast % increased from <5% to 21-30%: % patients with plt <20000/μl increased from 4 to 14%, and % patients with Hb ≤8g/dl increased from 23 to 45%. Within the FAB subtypes, RARS and del(5q) subgroups had the lowest % of patients with plt <20000/μl (0 and 4%) compared to other subgroups, while RAEB-T subgroup had the highest % of these patients (10%); CMML had the lowest % of pts with Hb ≤8g/dl (15%), while del(5q) subgroup had the highest % of these patients (44%). Patients with either plt <20000/μl or Hb ≤8g/dl had increased representation within higher clinical risk categories. Patients with more severe cytopenias had lower OS and higher risk of AML evolution (Table1, Log rank chi-squared (X²) test of association, p<0.0001). The correlation was not found with cytogenetic or age groups. In multivariate analysis, only Hb categories had additive prognostic value to the IPSS for evaluation of OS, but not AML evolution, with greatest prognostic value in int-1 and int-2 categories (Cox model, X²=81, p<0.0001). Our data demonstrated associations between depth of cytopenias and clinical subgroups in MDS which correlated with clinical outcomes, thus refining IPSS. In parallel with prior reports suggesting that red blood cell transfusion dependency was associated with poor clinical outcomes in MDS, these data indicate that the depth of the Hb level per se at MDS diagnosis has additive predictive value beyond the IPSS for OS.