Dyserythropoiesis Evaluated By Red Score and Hepcidin/Ferritin Levels Predicts Response to Epoetin Zeta in Lower Risk MDS

INTRODUCTION: Erythropoiesis stimulating agents (ESA) are generally the first line of treatment of anemia in lower risk myelodysplastic syndrome (MDS) patients (pts) with a recent approval of epoetin alfa in this indication in EU. Predictive factors of response to ESA (E Hellström, BJH 2003, S Park et al, Blood 2008, V Santini et al, Blood 2013; O Kosmider et al., Haematologica 2016) include low RBC transfusion requirement, low serum EPO level, serum ferritin (SF) level<300ng/ml, low IPSS-R, and ≤2 somatic mutations. In a clinical trial using a biosimilar (epoetin zeta) in lower risk MDS pts, we investigated new biomarkers using flow cytometry (Ogata and Red Scores) (Mathis et al, Leukemia 2013), somatic mutations, GDF-15 and hepcidin levels.

METHODS: Main inclusion criteria for this multicenter prospective trial involving 16 French centers from 2013 to 2017 were: (1) IPSS low or int-1 MDS (2) ESA naive, (3) with Hb level< 10g/dL, transfusion dependent (TD) or not (4) without unstable cardiovascular disease, renal failure, folates, vitamin B12 or iron deficiency or active inflammatory status. Pts received epoetin zeta 40000 IU/week for at least 12 weeks and erythroid response (HI-E, IWG 2006 criteria) was assessed at W12. sEPO level, iron homeostasis markers, Ogata and Red Score (RS), GDF15, hepcidin, and molecular analysis by NGS on a 39-gene panel were determined at baseline (T0), and Ogata and RS were reassessed at W12.

RESULTS: 70 pts were included. Median age was 78 years (range 57-93), M/F: 31/39. At inclusion, WHO classification was 22 RCMD, 19 RCUD, 14 RARS, 4 RAEB, 6 CMML, 2 del 5q-, 3 unclassifiable, IPSS was 43 low (61.4%), 27 int-1 (38.5%), and R-IPSS was 13 very low (18.5%), 47 low (67%), 9 int (13%) and high (0.7%). Twenty (28.5%) pts were TD (median of 2 RBC/ 8weeks). ECOG was ≥1 in 30 pts.

Median Hb level was 9.5g/dL (IQR 8.9-10), Transferrin Saturation (TS) 39% (IQR 28-56), SF 411 ng/ml (IQR 258-831), sEPO level 49 U/L (IQR 25-122). Median Ogata score was 1 (IQR 1-2) at T0 and 2 (IQR 1-2) at W12. Median Red score (RS) was 7 (IQR 5-7) at T0 (with 90% of pts having RS ≥3), and 5 (IQR 4.75-7) at W12.

Median GDF15 level was 1971 pg/ml (IQR 1236-2860). GDF15 was lower in IPSS low vs int-1 pts, and correlated with TS (p<0.003). Hepcidin level was 27 ng/ml (IQR 14-42; normal 1.3-26.4). Median hepcidin/ferritin ratio (HFR) was 5.5 (IQR 1.9-11.85). HFR was inversely correlated with MCV, serum iron, TS, GDF15, and RS at T0 and at W12, and with the number of mutations.

Before treatment, NGS was performed in 68 pts. Most frequent mutations involved SF3B1 (n=26), TET2 (n=18), ASXL1 (n=14), SRSF2 (n=9), DNMT3A (n=9), U2AF1 (n=7), IDH2 (n=6), and EZH2 (n=6). No, 1, 2, 3, 4 or 5 mutations were detected in 12, 12, 26, 10, 7 and 1 pts respectively, and 26% of pts had >2 mutations.

The erythroid response (HI-E) rate was 47.6%. Median duration of response was 13 months. Three pts progressed to AML. Twelve pts had died. Median OS was 41 mo.

Responders had significantly lower baseline RS (median 5 in responders vs 7 in non-responders, p=0.01), higher HFR (median 9 vs 4.8, p=0.04), lower IPSS (p=0.01). GDF15 tended to be lower in responders (1822 pg/mL) vs non-responders (2180 pg/mL; p=0.47). Age, sex, ECOG, blood cell count (ANC, platelets, Hemoglobin), number and type of mutations, IPSS-R, SF, sEPO level, T0 and W12 Ogata score, previous RBC transfusions, decrease of RS and of Ogata score between T0 and W12 were not significantly associated with response. In multivariate analysis, only RS≥5 at T0 was significantly associated with worse HI-E (p= 0.007, OR 19, 95% IC: 2.1-476). HI-E was 39% in pts with RS≥5 vs 75% in pts with RS<5.

CONCLUSION: Erythroid response (HI-E) to epoetin zeta was 48% in lower risk MDS and was correlated with higher baseline hepcidin/ferritin ratio and lower Red score at T0. In multivariate analysis, only Red Score≥5 was associated with worse HI-E, in this population with very low median sEPO level.