A Phase II Trial of Eltrombopag in Chronic Myelomonocytic Leukemia (CMML) with Thrombocytopenia

Context: Thrombocytopenia is a frequent complication in CMML associated with worse prognosis (Itzykson JCO 2013, Patnaik Leukemia 2013). Thrombopoietin (TPO) protects HSCs from DNA damage (de Laval Cell Stem Cell 2013). The TPO analog eltrombopag (ELT) has shown efficacy in MDS (Oliva Lancet Hematol 2017). The limited data on ELT in CMML is restricted to a small cohort (n=7) with high risk features, where 4 pts had disease progression during ELT (Ramadan Clin Lymphoma Myeloma Leuk 2016). We report interim results of a multicenter phase 2 trial investigating the efficacy, toxicity and biomarkers of ELT in 25 CMML with platelets (PLT) < 50 x10⁹/L (NCT02323178).

Methods: Key inclusion criteria were: WHO 2008 defined HMA-naive CMML, PLT < 50 x10⁹/L, marrow blasts ≤ 5%, IPSS low/int-1 in MD-CMML, and in MP-CMML no or 1 severity criteria (Hb < 10 g/dL, ANC > 16 x10⁹/L, abnormal karyotype, extramedullary disease) ,spleen size < 16 cm. ELT was started at 100 mg/d with escalating doses to 300 mg/d. Platelet Response (HI-P) was evaluated at 12 weeks, according to IWG 2006 criteria, and responders could continue treatment. Gene mutations were analyzed by NGS at inclusion. DNA repair was explored by counting gH2AX foci in CD34⁺ cells obtained at inclusion, evaluation and study exit.

Results: We report on the first 19 evaluable pts (median age 79y, M/F 15/4), all CMML-1 by WHO 2008, including 14 MD-CMML and 5 MP-CMML, with a low/int-1/int-2 CPSS (Such Blood 2013) in 61%, 33% and 6% of pts respectively (resp). Median PLT count was 29 x10⁹/L (IQR 19 - 43.5) and 10 (53%) pts had PLT transfusion dependency (PLT-TD). TET2, RUNX1, ASXL1, SRSF2, PHF6, and RAS pathway mutations were present in 17, 10, 6, 7, 4 and 7 of 18 pts sequenced, resp.

The median maximal dose of ELT received was 200 mg/d (IQR: 150 - 300 mg/d). Clinical grade ≥3 toxicities, all cardiovascular or pulmonary, were noted in 5 pts in the first 12 weeks, and biological grade ≥3 toxicities in 6 pts, without grade ≥3 ALT/AST elevation. Six (32%) pts went off study before week 12 because of lack of efficacy (n=2), presence of in circulating blasts (n=1), progression to CMML-2 (n=1), pancreatic cancer (n=1) and patient decision (n=1). All were considered non-responders. At 12 weeks, 12 (63%, 95%CI: 38 - 84) pts had achieved HI-P, including 3/5 MP-CMML and 9/14 MD-CMML (p=1). Responses were obtained in 7/10 pts with baseline PLT-TD and 5/9 pts without PLT-TD (p=0.65). The mutational profile of responders was comparable to that of non-responders except that none of the 4 PHF6^mut pts responded (p=0.01). Responders at 12 weeks were receiving a median of 125 mg/d of ELT, compared to 225 mg/d in non-responders.

Responders received ELT for a median of 21 weeks (IQR 15 - 38). With a median follow-up of 18 months (95%CI: 16.8 - NR), 11 of the responders discontinued ELT because of loss of efficacy (n=2), patient decision (n=1), death from unrelated causes (n=2), liver toxicity (n=2), renal failure (n=1) and disease progression (n=3). Median response duration was 8 months (95%CI: 5 - NR), with one pt still on treatment with a response lasting more than 14 months. The 12-month cumulative incidence of AML was 19% (95%CI: 4 - 41%) and 12-month progression-free and overall survival (OS) were 41% (95%CI: 23 - 73) and 65% (95%CI: 45-92%). By comparison, in the 21 CMML pts with PLT < 50 x10⁹/L and ≤5% BM blasts from our previous CMML cohort (Itzykson JCO 2013), none of whom received ELT, the 12-month estimates for cumulative incidence of AML and OS were 10% (95%CI 0 - 23%) and 71% (95%CI 47 - 86%), resp. An updated analysis including all 25 pts will be presented.

At protocol screening, a higher number of gH2AX foci were seen in CMML bone marrow CD34+ cells (n=14) compared to age-matched controls (n=5, p<0.01), reflecting higher levels of baseline DNA damage. However, in the 5 pts for whom serial bone marrow samples were obtained during ELT, gH2AX foci were reduced in CD34⁺ cells from all 5 pts, suggesting that in vivo exposure to ELT protected CMML stem/progenitor cells from DNA damage.

Conclusion: Our interim results suggest that ELT is manageable in CMML pts with no excess of marrow blasts and thrombocytopenia and results in a high response rate, with no obvious increase in disease progression, though prolonged responses seem infrequent. ELT could potentially protect CMML stem/progenitor cells from DNA damage. Larger pt numbers and longer follow-up are however required to determine precisely the role of ELT in thrombocytopenic CMML pts.