Therapeutic Strategies in Patients with Atypical CML (aCML) and Unclassified MDS/MPN (MDS/MPN-U). a Single Center Report

Introduction: The WHO defined overlapping MDS/MPN group includes two rare entities: BCR-ABL negative aCML and MDS/MPN-U, defined by at least 3 months persistent WBC >13G/l or platelets >450G/l with at least 1 cytopenia, significant dysplasia (>10%) in at least 1 lineage, no circulating monocytosis (<1G/l or less <10% WBC) and <20% peripheral and/or BM blasts (Wang et al. Blood 2014). aCML is distinguished from MDS/MPN-U by >10% immature circulating precursors. Median overall survival (OS) in aCML and MDS/MPN-U ranges from 12 to 22 months, without any standard treatment clearly defined. We previously found Azacitidine (AZA) to yield responses in BCR-ABL negative MPN including a small subset of MDS/MPN patients (Thépot el al. Blood 2011).

Methods: 20 patients, with WHO defined aCML or MDS/MPN-U, were diagnosed in our center from November 2003 to January 2014. Their characteristics and response to treatments administered, including AZA (using IWG 2006 criteria for MDS and evolution of MPN characteristics, such as WBC and splenomegaly (SM) were analyzed.

Results: Median age was 73.5 years (range 54-89), M/F 9/11, WHO diagnosis aCML (N=13) and MDS/MPN-U (N=7). Median WBC was 23.4G/l (13.2-96.4), median circulating myeloid precursors were 13.5% (4-66), median ANC was 13.1G/l (3.8-33.8), median Hb level was 10.6g/dl (7.2-14.5) and median platelets 125G/l (27-900). Median BM blast count was 3% (2-12). Five patients (25%) had SM at diagnosis (4 aCML; 1 MDS/MPN-U). Karyotype was normal in 9 and abnormal in 11: (isolated +8 (N=5), del20q (N=3), del5q (N=1), t(3 ;5);+8 (N=1) and tetrasomy 8 (N=1)). Of 14 patients analyzed, only 3 (21.4%) were JAK2-V617F mutated (2 aCML; 1 MDS/MPN-U). Median OS of the entire cohort was 26 months and median OS of aCML and MDS/MPN-U patients was 22 months and 32 months respectively (p=0.71). Time to disease progression (defined as the appearance of marrow blasts >5%, or of a new cytopenia (ANC<1.5G/l; Hb<10g/dl or transfusion dependency, Plt<100G/l) was 24 months. 11 patients (55%) progressed, including 4 to AML. At diagnosis, 8 patients received hydroxyurea (HU), 2 patients AZA because of transfusion dependent anemia and excess of blasts and 10 patients best supportive care only (BSC). At disease progression, 7/11 patients received AZA and 4 other therapies (intensive chemotherapy (IC), n=1, HU alone, n=1, BSC, n=2). HU yielded no significant ORR in the 9 patients treated except for control of WBC. The patient receiving IC had treatment failure. Six of the 9 patients treated with AZA also needed concomitant HU. Median BM blast count at AZA onset was 12% (5 aCML, 3 MDS/MPN-U, 1 AML) and median number of AZA cycles was 6 (3-11) with an ORR according to IWG 2006 criteria of 22.2% (2/9 patients) with 1 CR and 1 mCR+HI-P. In those 2 patients, WBC normalized despite HU discontinuation and a third patient experienced a significant decrease of SM while HU could be also discontinued. In this cohort of aCML and MDS/MPN-U, only 1/7 patients with chromosome 8 abnormalities responded to AZA. Median OS of patients receiving AZA was 12 months. All patients treated with AZA died from disease progression at the closing date of our study.

Conclusions: Treatment options for aCML and MDS/MPN-U, other than supportive care and HU are currently mostly limited to investigational agents. The role, if any, of hypomethylating agents, in these diseases, is not established. In this single center cohort, 2 marrow responses (1 CR and 1 mCR+HI-P) with AZA were observed, yielding similar modest ORR as reported for proliferative CMML treated by AZA (Drummond et al. Leukemia 2014). Due to the rarity of these diseases, and their genetic heterogeneity, multicenter prospective trials are needed to assess any potential clinical benefit of hypomethylating agents.