Prognosis In Patients with MDS or AML and Bone Marrow Blasts Between 10% and 30% Is Not Associated with Blast Count but Depends on Cytogenetic and Molecular Genetic Parameters

Since the WHO classification in 2001, myelodysplastic syndromes (MDS) and acute myeloid leukemia (AML) are separated by 20% of blasts in the bone marrow (BM). However, whether this arbitrary threshold can be justified from biologic or genetic aspects is in continuous debate. Design of study: We analyzed clinical outcomes and cytogenetic/molecular genetic features in 292 pts with 10–30% of BM blasts as assessed by cytomorphology. According to WHO criteria (2008), 111 pts were diagnosed as MDS with 10–19% of myeloblasts, while 181 had a diagnosis of AML with 20–30% of myeloblasts (“AML with recurrent genetic abnormalities” were excluded).

A total of 292 pts (f/115; m/177; median, 68.4 yrs; 20.4–88.3 yrs) were investigated by MGG staining of BM, chromosomal banding analysis, FISH, and PCR for the FLT3-ITD, NPM1, and MLL-PTD.

Median age did not differ significantly between the MDS and AML cohorts (68.8 yrs versus 68.2 yrs; n.s.; T-test). Aberrant karyotypes (MDS: 45/111; 40.5%; vs. AML: 79/181; 43.6%) and cytogenetic risk groups (classified according to MRC cytogenetic criteria; Grimwade, 2010) were similar in MDS and AML cohorts: intermediate karyotypes: MDS: n=86/111; 77.5%; vs. AML: 140/181; 77.3%; unfavorable karyotypes: MDS: 25/111; 22.5%; vs. AML: 41/181; 22.7% (Fisher's exact test; n.s.). Subsequently, the different molecular markers were compared between the different categories. No significant differences were detected between MDS and AML cohorts regarding the NPM1mut (total cohort: n=55/292; 18.8%; MDS: 23/111; 20.7%; vs. AML: 32/181; 17.7%; n.s.), FLT3-ITD (total: 12/292; 4.1%; MDS: 3/111; 2.7%; vs. AML: 9/181; 5.0%; n.s.), and MLL-PTD (total: 18/292; 6.2%; MDS: 5/111; 4.5%; vs. AML: 13/181; 7.2%; n.s.). Overall survival (OS) was first compared between the two above defined morphological subgroups according to thresholds of 20% of blasts. No significant differences in the median OS were observed (MDS: 27.3 mo. vs. AML: 43.7 mo.; p=0.687). This was followed by analysis of OS according to 4 blast categories: 1) 10–14% of blasts: n=61, median OS: 34.9 mo.; 2) 15–19%: n=50: 21.5 mo.; 3) 20–24%: n=98: 43.7 mo.; and 4) 25–30%: n=83: median not reached). When comparison was performed between all of the subgroups, no significant differences were detected (p=0.422). In contrast, when the total cohort was subdivided according to cytogenetics, pts with unfavorable karyotypes (KTs) (n=66/292; 22.6%; median OS, 14.1 mo.) had significantly worse outcomes than those with intermediate karyotypes (226/292; 77.4%; 43.7 mo.; p<0.001). The cytogenetic risk group had a significant prognostic impact also when MDS (unfav. KTs: med. OS, 13.1 mo.; intermed. KTs: 34.9 mo.; p=0.002) and AML (unfav. KTs: med. OS, 19.0 mo.; intermed. KTs: 43.7 mo.; p=0.046) were separately investigated. Focusing again on the total cohort, an NPM1 mutated status conferred a better prognosis when compared to NPM1 wildtype pts (median not reached versus 27.3 mo.; p=0.031), while FLT3-ITD and MLL-PTD had no significant influence on outcomes (FLT3-ITD positive: 14.7 mo. vs. FLT3 wildtype: 34.9 mo.; p=0.896; MLL-PTD positive: median not reached; MLL-PTD negative: 34.4 mo.; p=0.496). Furthermore, OS was better in the NPM1mut/FLT3wt subgroup (n=46; median OS, not reached) when compared to all other molecular subgroups (n=246; median OS, 27.3 mo.; p=0.016). In univariate analysis for the total cohort, age (p=0.001), cytogenetic risk group (p=0.001), the NPM1mut/FLT3wt status (p=0.022) were prognostically relevant, whereas the blast percentages (10-19% vs. 20–30%; p=0.687), WBC (p=0.496), hemoglobin (p=0.071), presence of FLT3-ITD (p=0.896), and MLL-PTD (p=0.499) had no relevance. In multivariate analysis, only age (p=0.008) and the cytogenetic risk group (p=0.004) were significantly associated with OS.

1) The separation of patients demonstrating 10–30% of bone marrow myeloblasts into MDS and AML does not distinguish between groups with different prognosis. 2) Cytogenetic and molecular genetic risk factors occur with comparable frequencies in MDS and AML subsets. 3) Karyotype and NPM1 mutation status are powerful parameters for risk categorization and treatment stratification in both subgroups with borderline blasts.

No relevant conflicts of interest to declare.