Next-Generation Sequencing in Myelodysplastic Syndromes: Prognostic Interaction Between Adverse Mutations and IPSS-R

Background

In a seminal report (NEJM 2011;30:2496), ASXL1, TP53, EZH2, ETV6 and RUNX1 mutations in myelodysplastic syndromes (MDS) were shown to carry an adverse prognostic impact that was independent of the international prognostic scoring system (IPSS) (Blood 1997;89:2079). Two recent studies (Leukemia20 May 2016; doi: 10.1038/leu.2016.138; Leukemia. 2014;28:241) have proposed mutation-enhanced prognostic models that include the revised IPSS (IPSS-R) (Blood 2012;120:2454 ). In the current study, we applied targeted next generation sequencing (NGS) in order to examine the prognostic interaction between adverse mutations and IPSS-R.

Methods

The study population was recruited from our institutional database of patients with primary MDS, based on availability of archived DNA from the time of diagnosis. The diagnosis of MDS and leukemic transformation (LT) was made according to WHO criteria (Blood. 2009;114:937). Targeted capture assays were carried out, on bone marrow DNA specimens, for the following genes: TET2, DNMT3A, IDH1, IDH2, ASXL1, EZH2, SUZ12, SRSF2, SF3B1, ZRSR2, U2AF1, PTPN11, Tp53, SH2B3, RUNX1, CBL, NRAS, JAK2, CSF3R, FLT3, KIT, CALR, MPL, NPM1, CEBPA, IKZF, and SETBP1.Specific variants were deemed as mutations if they are associated with a hematologic malignancy (as identified by COSMIC database), or if they have not been associated with a dbSNP.

Results

Patient characteristics and type and number of mutations:

179 patients were evaluated (median age 73 years; 68% males). IPSS-R risk distribution was 11% very high, 18% high, 16% intermediate, 39% low and 16% very low. At least one mutation was seen in 139 (78%) patients including 17% with one mutation, 29% with 2, 20% with 3, 10% with four, 2% with 5 and one patient with 6 mutations. The most frequent mutations were ASXL1 (30%), SF3B1 (20%), TET2 (17%), U2AF1 (16%) and SRSF2 (16%).

Mutation clusters and clinical correlates:

ASXL1 mutations were associated with the absence of SF3B1 (p=0.007), U2AF1 (p=0.01) and SRSF2 (p=0.003) mutations; SF3B1 with absence of U2AF1 (p=0.004) and SRSF2 (p=0.004) mutations; and U2AF1 with absence of SRSF2 mutations (p=0.01). Clinical correlates of ASXL1 mutations included older age, lower hemoglobin, and lower risk karyotype; SF3B1 with higher leukocyte count, higher platelet count, lower bone marrow blast percentage and lower risk karyotype; TET2 with older age and higher hemoglobin.

Univariate overall and leukemia-free survival analysis before and after adjusting for IPSS-R:

In univariate analysis, ASXL1, RUNX1 and TP53 mutations adversely and SF3B1 favorably affected survival; only ASXL1 mutations retained significance when analysis was adjusted for IPSS-R (HR 1.5, 95% CI 1.0-2.1; p=0.03). For leukemia-free survival, univariate analysis identified SRSF2, which was near-significant for overall survival (p=0.06), and RUNX1 mutations as adverse and SF3B1 mutations as favorable risk factors; SRSF2 (HR 4.1, 95% CI 1.6-10.2) and SF3B1-unmutated (HR 5.9, 95% CI 1.1-31.5) retained their significance when adjusted for IPSS-R. A borderline significance (p=0.07) indicating inferior survival for patients with three or more mutations was fully accounted for by the significant (P<0.0001) association between the presence of ASXL1 mutations and higher number of mutations.

Multivariable analysis with individual IPSS-R variables:

When the five IPSS-R variables (hemoglobin, platelets, absolute neutrophil count, bone marrow blast percentage and karyotype) were analyzed (both as continuous and IPSS-R-defined categorical variables) with each one of the aforementioned mutations with significant effect on overall or leukemia-free survival, only ASXL1 retained its significance for overall survival (HR 1.6, 95% CI 1.1-2.3) and SRSF2 (HR 5.2, 95% CI 2.1-13.3) for leukemia-free survival. Other IPSS-R variables that retained their significance for survival, in the presence of ASXL1 mutation status as a covariate, included hemoglobin level, platelet count and karyotype.

Conclusions: In our cohort of 179 molecularly-annotated patients with newly-diagnosed primary MDS, ASXL1 and SRSF2 mutations were identified as IPSS-R-independent risk factors for overall and leukemia-free survival, respectively. The current study also suggests the need to re-evaluate currently established risk factors, in the context of prognostically-relevant molecular information.