PF670 MOLECULAR CLASSIFICATION OF PATIENTS WITH POLYCYTHEMIA VERA OR ESSENTIAL THROMBOCYTHEMIA WHO DEVELOP RESISTANCE OR INTOLERANCE TO HYDROXYUREA

Background: A recent study (Grinfeld et al, NEJM 2018) has proposed a genomic classification of myeloproliferative neoplasms (MPN) that predicts patients’ outcomes. There is no information on the potential value of such classification to identify patients with polycythemia vera (PV) and essential thrombocythemia (ET) who may have at a higher risk to develop treatment resistance, particularly to hydroxyurea (HU). Aims: To perform a molecular characterization of PV or ET patients who develop resistance/intolerance to HU by means of next generation sequencing (NGS), and to evaluate the impact of specific non‐driver mutations on the overall survival and the risk of transformation to myelofibrosis (MF) or myelodysplastic syndrome/acute myeloid leukemia (MDS/AML). Methods: NGS studies were performed in 58 patients (PV n = 37, ET n = 21) with resistance or intolerance to HU (ELN criteria). Patients were hierarchically classified in the eight genomic subgroups reported by Grinfeld et al . Group 1: TP53 disruption or aneuploidy ( TP53 mutation, Chr17pLOH or Chr5‐/Chr5q‐). Group 2: ≥1 chromatin or spliceosome mutation ( EZH2 , IDH1 , IDH2 , ASXL1 , PHF6 , CUX1 , ZRSR2 , SRSF2 , U2AF1 , KRAS , NRAS , GNAS, CBL , Chr7/7qLOH, Chr4q/LOH, RUNX1 , STAG2 and BCOR ). Group 3: CALR mutation. Group 4: MPL mutation. Group 5: homozygous JAK2 mutation. Group 6: heterozygous JAK2 mutation. Group 7: MPN with other driver mutation. Group 8: MPN with no known driver mutation. Results: Median time to resistance was 48 months (range: 1–287) corresponding to thrombocytosis >600x10 9 /L despite high doses of HU (n = 8, 14%), need for phlebotomies despite high doses of HU (n = 6, 10%), uncontrolled myeloproliferation (n = 3, 5%), progressive splenomegaly (n = 2, 3%), cytopenias at the lowest dose of HU to achieve response (n = 37, 64%), and extrahematological toxicity (n = 2, 3%). Genomic characterization of these patients showed a higher frequency of cases with TP53 disruption/aneuploidy or with chromatin‐spliceosome mutations, and a lower frequency of MPN with heterozygous JAK2 mutation, especially for PV patients, compared to previously reported data (Fig. 1). The genomic subgroup changed from diagnosis to time of resistance in 7 out of 29 cases (24%) with available paired samples: Group 1 (two new cases with TP53 mutation), Group 2 (3 new cases due to acquisition of mutations in ZRSR2 , PHF6 plus ZRSR2 , and STAG2 , respectively), Group 5 (1 case changed from heterozygous to homozygous JAK2 ), and Group 7 (1 case with loss of JAK2 ). Cytopenia at any dose of HU to achieve response was more frequently observed in cases with TP53 disruption/aneuploidy or with chromatin‐spliceosome mutations than in the other subgroups (82% vs. 47%, p = 0.007). With a median follow‐up of 7.6 years from HU start (range: 1–28), 26 patients died and 16 and 11 patients progressed to MF and MDS/AML, respectively. There was a tendency towards inferior survival in MPN patients with TP53 disruption/aneuploidy whereas no differences in MF transformation were observed according to molecular classification. All MDS/AML progressions were observed in the TP53 disruption/aneuploidy or with chromatin‐spliceosome mutations groups (7‐year probability of 44% vs. 0% in other genomic subgroups, p  < 0.001, Fig. 1). Summary/Conclusion: Treatment with HU is inadequate for PV and ET patients with TP53 disruption/aneuploidy or chromatin‐spliceosome mutations due to the development of cytopenia at the lowest responding dose and a high rate of progression to MDS/AML. These data favor a change towards a molecular classification of MPN and support a different therapeutical approach for high risk genetic groups.