Validation of response assessment according to international consortium for MDS/MPN criteria in chronic myelomonocytic leukemia treated with hypomethylating agents

The hypomethylating agents (HMA) azacitidine (AZA) 1,2 and decitabine (DAC) 3,4 are active in Chronic Myelomonocytic Leukemia (CMML) with overall response rates (ORR) of 40 – 70%, translating into median overall survivals (OS) of 12 – 22 months. Response to HMA in these CMML cohorts was mostly evaluated according to IWG-2006 criteria, 5 which do not assess improvement of myeloproliferative features nor quality of life. In myelodysplastic syndroms (MDS) patients treated with HMA, these criteria have limited impact in predicting OS. 6,7 Normalization of WBC and monocyte counts, regression of splenomegaly or other extra-medullary disease, and improvement of quality of life have been reported in CMML patients treated with HMA 1,8 and studies in MDS and primary myelo ﬁ brosis have shown that symptom improvement could be correlated to a prolonged OS. 9,10 To capture both MDS and myeloproliferative neoplasm (MPN) features in a singly response scale in CMML, an international expert panel, the MDS/MPN International Working Group, proposed new response criteria for MDS/MPN, hereafter referred to as overlap-MDS/MPN criteria. 11 These new criteria take in consideration bone marrow and peripheral blood blast reduction and improvement of cytopenias, but also account for correction of WBC, monocyte and peripheral immature myeloid cell counts (IMC), regression of splenomegaly and other extra-medullary disease. These criteria also assess correction of myelo ﬁ brosis, generally moderate in CMML, 12 and propose a provisional entity of ‘ clinical bene ﬁ t ’ solely based on improvement assessed with the MPN-SAF scoring system, 13 which has been developed in primary

The hypomethylating agents (HMA) azacitidine (AZA) 1,2 and decitabine (DAC) 3,4 are active in Chronic Myelomonocytic Leukemia (CMML) with overall response rates (ORR) of 40-70%, translating into median overall survivals (OS) of 12-22 months. Response to HMA in these CMML cohorts was mostly evaluated according to IWG-2006 criteria, 5 which do not assess improvement of myeloproliferative features nor quality of life. In myelodysplastic syndroms (MDS) patients treated with HMA, these criteria have limited impact in predicting OS. 6,7 Normalization of WBC and monocyte counts, regression of splenomegaly or other extramedullary disease, and improvement of quality of life have been reported in CMML patients treated with HMA 1,8 and studies in MDS and primary myelofibrosis have shown that symptom improvement could be correlated to a prolonged OS. 9,10 To capture both MDS and myeloproliferative neoplasm (MPN) features in a singly response scale in CMML, an international expert panel, the MDS/MPN International Working Group, proposed new response criteria for MDS/MPN, hereafter referred to as overlap-MDS/MPN criteria. 11 These new criteria take in consideration bone marrow and peripheral blood blast reduction and improvement of cytopenias, but also account for correction of WBC, monocyte and peripheral immature myeloid cell counts (IMC), regression of splenomegaly and other extra-medullary disease. These criteria also assess correction of myelofibrosis, generally moderate in CMML, 12 and propose a provisional entity of 'clinical benefit' solely based on improvement assessed with the MPN-SAF scoring system, 13 which has been developed in primary myelofibrosis and has never been validated in MDS/MPN. These new criteria remain to be validated.
To validate these overlap-MDS/MPN criteria in the most frequent entity amongst MDS/MPN, namely CMML, we updated clinical data from 79 CMML patients treated by AZA or DAC included in GFM CMML clinical trials (EudraCT No. 2008-000470-21) 4 or registry (PHRC MAD-06). 14 The cohort included 56 males and 23 females, with a median age of 72 years. At HMA onset, 57% of patients had CMML-1 and 43% had CMML-2. Splenomegaly was present in 40% of cases. Median Hb, WBC, ANC and platelets were 9.7 g/dl, 14.5 × 10 9 /l, 7.1 × 10 9 /l and 101 × 10 9 /l, respectively. CPSS prognosis score was low in 12%, intermediate-  (4-7)), without significant delay between criteria sets (paired t-test P = 0.43). According to IWG-2006, ORR was 57%. IWG-2006 responses included complete response (CR) in 20% of cases, marrow CR (mCR) with Hematological Improvement (HI) in 13%, mCR without HI in 10%, stable disease (SD) with HI in 14%. No patient achieved PR; 19% patients had SD without HI and 24% had progressive disease (PD). Regarding overlap-MDS/MPN criteria, the ORR was 71%. In our dataset, improvement of clinical symptoms was assessed retrospectively by reviewing patients' charts, instead of applying the MPN-SAF scoring system 13 as recommended. Similarly, we could not evaluate improvement of myelofibrosis, because the use of trephine biopsies is not part of the French guidelines for CMML. However, diffuse myelofibrosis is infrequent in CMML and can be suspected in case of dry tap. Overlap-MDS/MPN responses included CR in 13%, optimal marrow response (OMR) with clinical benefit (CB) in 18% (including CB in spleen size (CB-Spl) 6%, and CB in general symptoms (CB-Sym) 1%), OMR without CB in 15%, partial marrow response (PMR) in 1% and SD with CB in 24% (including CB-Spl 5%, and CB-Sym 1%). No patient achieved PR. Twenty-four percent of patients had SD, and 5% had progressive disease.
Overall response status at best response between IWG-2006 and overlap-MDS/MPN was concordant in 86% of cases, corresponding to a Cohen's Kappa 15 of.7, indicating a relatively good agreement between response criteria. Sources of discrepancies are summarized in Table 1  Eleven patients received ASCT, after a median of 6 cycles (IQR 3-13) of HMA, including 9 who had reached a response per overlap-MDS/MPN criteria (three OMR with CB, four OMR, two CB, including one CB-Spl), and only 4 of whom had reached a response per IWG-2006 criteria (3 HI and 1 mCR). One patient had spleen size reduction and was classified as stable disease per IWG-2006 criteria, and the remaining four patients were classified as progressive disease because of worsening cytopenias The small number of transplanted patients precludes analysis of post-transplant outcome, but these data suggest overlap-MDS/MPN responses encompass clinically meaningful improvement already considered for the decision to transplant in daily practice.
Finally, the impact on OS of dissociated responses was difficult to assess because of the small number of patients (n = 11), achieving a response per overlap-MDS/MPN but not per IWG-2006 criteria, and because five of them subsequently received allogeneic transplantation.
In conclusion, we report the first retrospective validation of these new overlap-MDS/MPN criteria in the setting of CMML treated with HMA. By taking into account improvement of myeloproliferative features and by allowing to classify patients with dissociated responses more easily, not considering isolated increase in bone marrow blasts or worsening of cytopenia as a progression, overlap-MDS/MPN criteria increase response rates as well as response duration.
These criteria remain to be evaluated in prospective studies, with a comprehensive evaluation of clinical symptoms and systematic cytogenetic examinations, to confirm their contribution in defining robust short-term endpoints for future CMML clinical trials.