Abstract
Background
Despite refinement on prognostic stratification in myelodysplastic syndromes (MDS), clinical heterogeneity, particularly in the group of lower risk disease (LR-MDS), characterizes these neoplastic conditions. Besides clinical and biological known prognostic parameters, recent progress in the mutational analysis of patients (pts) with MDS supports the evidence of the role of certain somatic mutations associated not only with overall survival but also with response to azacitidine (AZA) treatment [1-3]. However, mutational profile when leukemic progression occurs, and response to AZA, particularly in the group of lower-risk disease is uncertain. A recent analysis from Chronic Myeloid working group of the International Cancer Genome Consortium detected ASXL1 and RUNX1 as most frequent mutations in 3 out 7 LR-MDS patients with mutational data available at progression [4].
Methods
Retrospective analysis of a large cohort (N=437) of patients with primary LR-MDS [defined as either an IPSS score of low/intermediate-1 and/or favorable cytogenetic categories (good/intermediate by IPSS or very good/good/intermediate by IPSS-R)] followed until progression to acute myeloid leukemia (AML) with mutational analysis available at the time of leukemic progression. The Ion AmpliSeq™ AML Cancer Research Panel targeting 19 genes implicated in AML (entire coding regions: CEPBA, DNMT3A, GATA2, TET2, TP53; hotspot regions: ASXL1, BRAF, CBL, FLT3, IDH1, IDH2, JAK2, KIT, KRAS, NPM1, NRAS, PTPN11, RUNX1, WT1) was used to detect driver mutations in an "Ion S5™ Sequencer". Nonsense, frameshift and splicing mutations with frequency >3% were considered for analysis. Somatic confirmed mutations in COSMIC database were considered pathogenic and mutations not described in COSMIC but with a pathogenic prediction, were considered probably pathogenic.
Results
After 103 months median follow-up for the entire cohort (95% CI: 81-124), 77 patients (17.6%) progressed to AML. From those, data on somatic mutations at the time of leukemic progression was available in 16 patients. Median time from diagnosis of MDS to leukemic progression was 10.2 months (range: 2-98 months). Mean number of somatic mutations detected per patient was 2 (range: 1-5), being RUNX1 the most frequently mutated (5 pts; 31%), followed by TET2 and CBL (found in 4 pts each; 25%) and TP53, CEBPA, DNMT3A, GATA2, ASXL1, BRAF, IDH2, JAK2, KRAS, NPM1 and WT1 (detected in 3 pts, each; 18%). According to WHO classification, refractory cytopenia with multilineage dysplasia was the most frequent MDS subtype and 12/16 pts had intermediate o high risk by the MDA score for LR-MDS. Karyotype was diploid in 11/16 cases at diagnosis. Four pts acquired additional cytogenetic abnormalities at AML progression (pts#2, 10, 14 and 15 in table). Interestingly, 3 out of those 4 had TP53 mutation at progression, all of them with complex karyotype at AML evolution, with 1 remaining pt presenting TET2, ASXL1 and CBL mutation. Nine pts received AZA at AML progression. Among AZA treated pts, 3/9 experienced long lasting and complete responses (2 pts with TET2 and CBL and 1 pt with TP53 mutations) while pts with remaining mutations identified at leukemic progression did not or lost response promptly.
Conclusion
This study, although lacking molecular data at diagnosis, provides a heterogeneous mutational landscape at AML progression, with some of them at a higher frequency than expected at MDS diagnosis (>30% pts had RUNX1 and 25% pts had CBL mutation at progression compared to 10-20% and 1% reported in MDS, respectively). Whether these represents ancestral mutations or were present as small subclones and expanded over time needs to be clarified in larger studies. Risk models including mutation data should evaluate this information, not only predicting survival, but also, the probability of leukemic progression.
References
[1] Bejar R, et al. N Engl J Med . 2011;364:2496-2506
[2] Bejar R, et al. Blood. 2015;126:907
[2] Itzykson R, et al. Leukemia. 2011:1147-1152
[3] Papaemmanuil E, et al. Blood . 2013; 122:3616-3627
No relevant conflicts of interest to declare.
