Pathway Analysis of Molecular Mutations Can Modify Morphologic, Cytogenetic and Prognostic Risk Stratification Schemes in Myelodysplastic Syndromes (MDS), Myelodysplastic Syndromes/Myeloproliferative Neoplasms (MDS/MPN), and Secondary Acute Myeloid Leukemia (AML)

Abstract 3791

Molecular mutations of genes important in MDS pathogenesis have been identified using next generation sequencing technologies. These mutations have variable effects on prognosis: results from individual studies are conflicting, and the impact of combining mutations based on pathogenetic pathways has not been explored. Furthermore, most studies do not take into account the types of therapies received by patients (pts). We hypothesized that analyses of mutations incorporating pathways would yield more meaningful results, especially if studied in the context of therapies. We studied 340 pts with MDS (N=176), MDS/MPN (N=88) and sAML (N=76). Median age was 69 years (range, 19–92), 118 (35%) were female. Direct sequencing of 14 genes important in MDS pathogenesis originally identified using whole exome sequencing and involved in methylation (TET2/IDH1/2/DNMT3A), histone modification (ASXL1/EZH2/UTX), signaling (CBL/KRAS, NRAS, JAK2), transcription (RUNX1/TP53) and RNA splicing (SRSF2/U2AF1) was performed. Pts were classified by 2008 WHO criteria. Prognosis was assigned by IPSS (low=67, Int-1=85, Int-2=56, high=64). Karyotypes by metaphase cytogenetics (MC) and SNP-A were stratified by IPSS cytogenetic risk group (Low=82, intermediate=87, poor=106). Treatment received included high intensity (cytarabine+ anthracycline, high dose ARA-C,) (N=75), low intensity (N=124) (hypomethylating therapy, IMiDs, low dose ARA-C), and supportive therapy (N=141) (growth factors, hydroxyurea).

Categorical variables were analyzed using X² or Fischer-exact test. Overall survival (OS) and progression-free survival (PFS) were defined using IWG criteria and analyzed by Kaplan-Meier and log-rank statistics. P-values0.05 were considered statistically significant. Univariate and multivariate analyses were performed by Cox Proportional Hazards, covariates included were age, BM blasts, cytopenias, karyotype, jointly with molecular pathway data. A total of 258 mutations were detected in 158 pts. Mutations involving genes key to methylation (41 vs 37%, p=.01), and histone modification (52 vs 25%, p=.008) were more common in MDS/MPN compared to MDS pts. Mutations in methylation (37 vs 23%, p=.05), transcription (38 vs 14%, p=.01) and splicing (38 vs 17%, p=.007) were more common in MDS compared to sAML pts, and mutations of genes involved in methylation (41 vs 23%, p=.02), signaling (60 vs 7%, p=.005), histone modification (52 vs 23%, p=.004), transcription (49 vs 13%, p=.001) and splicing (45 vs 17%, p=.0005) were more frequent in MDS/MPN compared to sAML pts. Worse outcomes were noted in pts with mutations in histone modification (OS: 17 vs 21 mos, p=.04), transcription (PFS: 7 vs 13 mos, p=.02) and splicing (OS: 11 vs 23 mos, p=.01) genes. This was evident in pts with MDS who had transcription factor-related gene mutations (OS: 9 vs 30 mos, p=<.0001), splicing (PFS: 8 vs 21 mos, p=.008) and in MDS/MPN and MDS pts with histone pathway (27 vs 31 mos, p=.04) and splicing mutations (7 vs 21 mos, p=.0001). Of note, this was clearly mentioned in WHO lower-risk MDS pts with histone pathway (OS: 30 vs 39 mos, p=.01; PFS: 21 vs 32 mos, p=.05) and splicing mutations (PFS: 15 vs 28 mos, p=.003). Pts with good-risk karyotype defects had poor survival if they had histone mutations (OS: 17 vs 35 mos, p=.005; PFS: 11 vs 20 mos, p=.002) while those with intermediate-risk karyotype had worse outcomes if they had splicing mutations (OS: 10 vs 35 mos, p=.001; PFS: 7 vs 35 mos, p=.0001). The survival effects were retained even after accounting for presence of new SNP-A lesions. By IPSS, lower-risk pts had worse outcomes if they had mutations in methylation (OS: 36 vs 18 mos, p=.02), histone modification (28 vs 37 mos, p=0.02) or splicing (OS: 7 vs 17 mos, p=.04; PFS: 14 vs 28 mos, p=.04). Therapeutically, LIC treated pts had better OS if they had methylation mutations (23 vs 8 mos, p=.0009) but worse outcomes if they had transcription factor (7 vs 11 mos, p=.02) or splicing (8 vs 13 mos, p=.05) mutations. Mutations in transcription factors (OS: HR=2.9 CI: 1.58–5.13, p=.0007; PFS: HR=3.3 CI: 1.87–5.63, p=<.0001) and splicing (OS: HR=2.1, CI=1.22–3.76, p=.008; PFS: HR=2.1, CI=1.27–3.71, p=.005) remained independent predictors of poor outcomes in MDS by multivariate analysis. In conclusion, the presence of molecular pathway mutations can modify morphologic, cytogenetic and prognostic scoring schemes in MDS, MPN, AML and other hematological cancers.