Patients with SF3B1 Mutation Have Good Prognosis Even in the Presence of Other Poor Prognostic MDS Features and Have Better Outcomes During Treatment with Low Intensity Chemotherapy

Abstract 3831

The discovery of mutations in components of the RNA splicing machinery has underlined the importance of this pathogenetic mechanism in MDS biology. We and others reported that mutations in SF3B1 are frequent in refractory anemia with ring sideroblasts (RARS) and RARS associated with thrombocytosis (RARS-T). We recently reported good survival outcomes in a large cohort of SF3B1 mutant patients similar to the findings of Papaemmanuil et al, NEJM< 2011. Here, we aimed to evaluate the correlation between SF3B1 mutations in association with other poor prognostic clinical and pathologic factors in MDS. We focused on a large cohort of patients with several myeloid malignancies (n=340). Patients were grouped in MDS (n=176), MDS/MPN (n=88) and secondary AML (sAML) from a previous MDS (n=76). Diseases were classified according to the 2008 WHO classification. Survival outcomes (overall survival [OS], progression free survival [PFS] and event free survival [EFS]) were defined according to MDS IWG criteria. We studied a total of 340 patients with myeloid malignancies. median age, 69 years old (range 19–92), sex (Male, n=222; females, n=118). We performed direct sequencing for SF3B1 (exon 13–16) on a large cohort of MDS, MDS/MPN and sAML patients and found 11.8% SF3B1 mutants. The outcomes of SF3B1 mutant vs WT patients were compared based on the presence of poor clinicopathologic factors associated with MDS like RBC transfusion dependence, presence of SNP-A lesions, presence of acquired somatic uniparental disomy (AS-UPD), Age ≥ 60 years, presence of reticulin fibrosis in the bone marrow. We also assessed the effects of therapies in the prognostic effect of SF3B1 mutations. We previously reported that the presence of new SNP-A lesions in myeloid malignancies including MDS and AML are associated with poor prognostic outcomes. However, the good prognostic effects of SF3B1 mutation is still apparent even in the patients with new SNP-A abnormalities (OS: 40 vs 16 mos, p=.003; PFS: 40 vs 10 mos, p=.003, 40 vs 10 mos, p=.0007. However, when analysis is limited to acquired somatic uniparental disomy defects which we recently reported as the worst lesion among SNP-A abnormalities, the good prognostic effect of SF3B1 is lost (OS: 19 vs 9 mos, p=.29, PFS: 18 vs 7 mos, p=.20, EFS: 19 vs 8 mos, p=.17). Age is an important predictor of outcomes in MDS with higher age associated with worse outcomes. SF3B1 remained predictive of good outcomes in patients ≥ 60 years of age (OS: 40 vs 16, p=.002; PFS: 40 vs 11 mos, p=.003; EFS: 40 vs 10 mos, p=.0005). Persistent RBC transfusions are also associated with inferior survival in MDS, yet SF3B1 mutant patients continued to have good outcomes (OS 34 vs 13, p=.002; PFS: 27 vs 8 mos, p=.004; EFS: 14 vs 9 mos, p=.001). Reticulin fibrosis in the bone marrow is a characteristic feature of myeloproliferative neoplasms but their presence in MDS is associated with unfavorable results, SF3B1 mutant retain their good outcomes even in the presence of reticulin fibrosis in the BM (OS: 40 vs 15 mos, p=.09; PFS: 47 vs 12 mos, p=.04; EFS: 75 vs 13 mos, p=.008). Patients screened for SF3B1 were also further stratified according to treatments received. No patients with SF3B1 mutations underwent allogeneic hematopoietic stem cell transplantation and high intensity chemotherapy specifically induction chemotherapy or high dose cytarabine. However based on treatment with LIC, SF3B1 mutant patients have better survival outcomes compared to WT patients (OS: 61 vs 17 mos, p=.004; PFS: not reached [NR] vs 7 mos; p=.009; EFS: 61 vs 10 mos, p=.001). These observed survival outcomes remain significant when patients were stratified according to disease subtypes. MDS and MDS/MPN patients treated with LIC have better outcomes if they have SF3B1 mutation (OS: 61 vs 23 mos, p=.005; PFS: NR vs 16 mos, p=.01; EFS: 61 vs 16 mos, p=.001). A high number of TET2/DNMT3A mutations are found in SF3B1 mutants. TET2/DNMT3A mutations have been previously association with improved response to hypomethylating agents which is one of the possible reasons why SF3B1 mutants treated with LIC did better than their WT counterpart. In conclusion, SF3B1 retains its favorable prognostic effect even in the face of poor prognostic factors such as RBC transfusion dependence, presence of SNP-A lesions, Age ≥ 60 years, presence of reticulin fibrosis in the bone marrow except in the presence of AS-UPD. SF3B1 mutants treated with LIC also have better outcomes.