S169: CLINICAL AND MOLECULAR MARKERS FOR PREDICTING RESPONSE TO ROMIPLOSTIM TREATMENT IN LOWER-RISK MYELODYSPLASTIC SYNDROMES

Background: In about half of patients with lower-risk (LR) myelodysplastic syndromes (MDS), thrombocytopenia is present at the time of diagnosis, being associated with shortened survival and a higher risk of progression to acute myeloid leukemia (AML). Romiplostim (ROM), a thrombopoietin receptor agonist (TPO-RA), has shown safety and clinical efficacy in prospective trials in LR-MDS. Post-hoc analyses have demonstrated hematologic improvement of platelets (HI-P) after ROM treatment contingent on endogenous thrombopoietin (TPO) levels and platelet transfusion events (PTE) (Sekeres et al. BJH 2014). Aims: The prospective EUROPE multicenter phase 2 trial (NCT02335268) investigated the predictive value of biomarkers like endogenous TPO levels, PTE and molecular markers on the clinical efficacy of single-agent ROM treatment within the `European Myelodysplastic Syndromes Cooperative Group` (EMSCO) network. Patients with IPSS low or intermediate 1 risk were eligible if baseline bone marrow blast count was <5% (central morphology) and platelet count was ≤30 Gpt/L or ≤50 Gpt/L in case of a bleeding history. Methods: According to a previously published model of response to TPO-RA (Sekeres et al. BJH 2014), patients were assigned into two different cohorts at the time of screening based on previous PTE and centrally assessed TPO serum levels (cohort A: TPO<500 ng/l and PTE<6 units/past year; cohort B: TPO>500 ng/l, and/or PTE36 units/past year). The primary efficacy endpoint was the rate of HI-P according to IWG 2006 criteria lasting for 38 weeks. ROM was initiated at a dose of 750 μg weekly by subcutaneous injection, the dose was adjusted based on the patient`s platelet counts. Results: From 2015 to 2019, a total of 77 patients were included at 29 different trial sites in Germany, France and the Czech Republic. Regarding the primary endpoint, 32 out of 77 (42%) responded (HI-P) with a numerically higher response rate in cohort A (47%, n=24) vs. cohort B (31%, n=8) (p=0.2953). At 16 weeks of ROM treatment, three (4%) and seven (9%) patients had additional neutrophil (HI-N) and erythroid (HI-E) responses, respectively. None of the patients achieved trilineage responses (HI-P, HI-E and HI-N). Median duration of response was significantly longer for patients in cohort A (351 days) compared to cohort B (315 days) (p=0.006, log-rank-test). Mutated SRSF2 was significantly more frequent in responders (41%) compared to non-responders (16%) (p=0.018, Fisher’s exact test). In patients with an SRSF2 mutation, the probability to achieve HI-P was 65% compared to 33% inpatients with SRSF2 wildtype (Figure 1A). Comparing responders vs. non-responders, we found no significant changes of variant allelic burden of variants detected pre- and post-ROM (Figure 1B). Finally, we developed a response prediction model to ROM therapy with the aim to improve personalized patient stratification in the future. The percentage of correctly predicted HI-P was highest for the model, which included the variables platelet count, SRSF2 mutation status and the hemoglobin level using the threshold of 11.4 g/dl and resulted in an overall accuracy of 70 % for a correct ROM response prediction (Figure 1A). Image:Summary/Conclusion: In conclusion, this prospective study confirms the efficacy and overall safety of ROM in this subgroup of LR-MDS patients with thrombocytopenia. To avoid overfitting of variables and to confirm our results, the here presented response prediction model needs to be validated in an external independent cohort. * U.P. and L.A. contributed equally to this study as senior authors