Abstract 2811

Among chronic myeloproliferative neoplasms (MPN) myelofibrosis (MF) has the worst prognosis. The International Prognostic Score System-IPSS reliably discriminates four categories of patients (pts) with different survival (OS); however, the possibility to identify pts with very high-risk disease, based on intrinsic characteristics of the underlying disease, would be particularly useful for therapeutic choices that include early stem cell transplantation. Several studies have addressed the prognostic value of molecular abnormalities harbored by MF pts, but a part from a negative role of low JAK2V617F allele burden on survival independently reported by the Mayo Clinic and our own group, none of them impacted significantly on survival. Here we have genotyped at diagnosis 370 pts with primary MF (PMF) and 148 with post-polycythemia vera/essential thrombocythemia MF (PPV/PET-MF) for mutations of EZH2 and ASXL1. EZH2 is the catalytic component of polycomb repressive complex 2 involved in trimethylation of H3 lysine 27, while ASXL1 is a member of enhancer of trithorax and polycomb group required for repression of HOX genes through deubiquitination of histone H2A. All 20 EZH2 exons were screened by high-resolution melting followed by direct sequencing whereas direct sequencing was used for ASXL1 exon 12 genotype. Pts median FU was 39 mo (1–340); 128 pts died (25.9%), 81 (18.3%) because of leukemia (AL). Among PMF pts survival differed significantly according to the IPSS category and was also predicted by the lowest V617F quartile.

EZH2 mutation. We found 22 PMF (6%), 1 PPV-MF (1.2%) and 6 PET-MF (9.4%) mutated pts, harboring 25 different mutations, 20 exonic (heterozygous in 83%), 5 intronic. Concurrent JAK2 V617F, MPL W515L/K, IDH1/2, TET2, CBL and ASXL1 mutations were found in 41.4%, 0%, 0%, 6.9%, 3.4% and 20.7% of EZH2 -mutated pts. In multivariate analysis EZH2 mutation was associated with leukocytosis (12.3±13.2 vs 18.7±11.5×10e9/L) in PMF pts, but no additional clinical/hematological correlation was discovered. EZH2 mut pts clustered preferentially in the IPSS high-risk category (52.6%; P=.002). More EZH2 mut than wt PMF pts died (51.9% vs 24.4%; P<.001); median OS was shorter in EZH2 mut pts (32 vs 137 mo, P=.001). Leukemia occurred in 31.8% of EZH2 mutated and 17.6% of wt pts; leukemia-free survival (LFS) was shorter in EZH2 mut PMF pts (25.7 vs 63.6 mo; P=.028).

ASXL1 mutation. In a total number of 334 evaluated patients we found 38 different mutations in 49 pts (14.7%), 35 PMF (15%), 5 PPV-MF (8.6 %) and 9 PET-MF (21.4%). Mutations were frameshift, non sense and missense mutations, heterozygous in 46 of 49 cases. Co-expression of JAK2 V617F, MPL W515L/K and EZH2 mutation was found in 55.1%, 4.1% and 12.2%, respectively. Co-espression of ASXL1 and IDH1/2, TET2 or CBL was observed in two, one and one of the evaluable subjects, respectively. ASXL1 mutation was associated with leucocytosis (P=.006), a blast count >1% (43.8% vs 16.6%, P<.0001) and constitutional symptoms (P=.027). ASXL1 mutated pts clustered preferentially in the IPSS high-risk category (46.4%; P<.0001). OS was significantly shortened in ASXL1 mutated PMF patients (median= 87.6 mo vs 264; P=.004); leukemia occurred in 34.9% of ASXL1 mutated and 15.2% of wt pts (P=.002). LFS was shorter in ASXL1 mutated PMF patients [137 mo than wt 264 mo, (P=.01)].

According to EZH2 and ASXL1 mutational status, median OS was significantly shorter in double EZH2/ASXL1 mutated PMF pts compared with single EZH2 or ASXL1 mutated or WT patients (34 vs 116 vs 198 vs 272 mo, respectively. P<.0001). On multivariate analysis, OS was predicted by IPSS score (P<.001), ASXL1 (P=.016) and EZH2 mutational status (P=.011). Leukemia-free survival (LFS) was shorter in double EZH2/ASXL1 mutated PMF pts compared with single EZH2 or ASXL1 mutated or WT patients (25 vs 153 vs 138 vs 264 mo; P=.002), but the low number of cases (n=6) prevented multivariate analysis.

No impact of EZH2 or ASXL1 mutations on OS or LFS was seen in PPV/PET-MF.

Overall, these results showed that mutations in genes involved in epigenetic gene regulation such as EZH2 and ASXL1 negatively impact on OS and LFS in pts with PMF, suggesting that screening for these mutations might be useful for risk stratification and treatment decisions.

Disclosures:

Barosi:Novartis: Membership on an entity's Board of Directors or advisory committees. Vannucchi:Novartis: Honoraria.

Author notes

*

Asterisk with author names denotes non-ASH members.

This icon denotes a clinically relevant abstract

Sign in via your Institution