EHA 2026 | Results from the EPO-Pretar trial: impact of early vs delayed ESA initiation in patients with LR-MDS

So this was a phase 3 GFM trial aiming at answering the question, do we have to initiate ESA early in low-risk MDS patients with anemia or after waiting for the transfusion threshold? So we designed the prospective trial to answer this question and there were two arms in the early arm the ESA was introduced when the hemoglobin was between 9 and 10.5 grams per deciliter and in the late arm ESA was introduced when the hemoglobin dropped below 9 grams per deciliter near the transfusion threshold. So therefore we enrolled 84 patients, 43 in arm A and 41 in arm B, and the median delay of initiation of EPO was 0.2 months in the early arm and 16 months in the delayed arm. Both groups were well balanced in terms of age, gender, IPSS-R, IPSS-M. And the primary endpoint was transfusion dependence, the delay of transfusion dependence. And finally, there was no difference in the early arm versus the delayed arm for transfusion dependence. Nevertheless, the erythroid response rate was higher in the early arm, 85% versus 55%. This was clearly significant. And the duration of response was numerically longer in the early arm, 32 months versus 20 months. So then after we evaluated the molecular IPSS to see whether it could predict response rate, and both groups were well balanced in terms of IPSS-M and gene mutations. But when we evaluated the ESA-IPSS-M score, which was published last year by Marco Raddi and Valeria Santini, so this score integrates endogenous erythropoietin level and IPSS-M. And there is quite a complex formula to calculate that. But to be simple, when the score is high, the better the response and the longer is the delay of transfusion dependence. So in the early arm, it was striking to see that this score could predict the patients with a higher response rate and with a longer delay of transfusion dependence. So to summarize, early initiation of ESA did not delay transfusion dependence but it leads to a higher response rate to ESA and a simple molecular score which integrates endogenous erythropoietin level can predict a subset of patients who will benefit from this early initiation of ESA. So these are the results of our poster today. And we analyzed the transcriptomic cytokine profile and the site of analysis of these patients included in the EPO-PRETAR trial. And the take-home message is that transcriptomic signatures showed that the early arm responders have an increase of gene expression in the interferon gamma signature. And in the late arm, there are some signatures correlated with inflammation. And these results were confirmed by a cytokine profile, where early arm responders secrete interferon gamma with T-cells and NK cells and whereas in late arm non-responders you have an inflammatory milieu with IL-6 and IL-1. So therefore it seems that the early responders have an adaptive, favorable immune microenvironment, whereas the non-responders have an inflammatory milieu which impedes ESA to lead to response.So this was a phase 3 GFM trial aiming at answering the question, do we have to initiate ESA early in low-risk MDS patients with anemia or after waiting for the transfusion threshold? So we designed the prospective trial to answer this question and there were two arms in the early arm the ESA was introduced when the hemoglobin was between 9 and 10.5 grams per deciliter and in the late arm ESA was introduced when the hemoglobin dropped below 9 grams per deciliter near the transfusion threshold. So therefore we enrolled 84 patients, 43 in arm A and 41 in arm B, and the median delay of initiation of EPO was 0.2 months in the early arm and 16 months in the delayed arm. Both groups were well balanced in terms of age, gender, IPSS-R, IPSS-M. And the primary endpoint was transfusion dependence, the delay of transfusion dependence. And finally, there was no difference in the early arm versus the delayed arm for transfusion dependence. Nevertheless, the erythroid response rate was higher in the early arm, 85% versus 55%. This was clearly significant. And the duration of response was numerically longer in the early arm, 32 months versus 20 months. So then after we evaluated the molecular IPSS to see whether it could predict response rate, and both groups were well balanced in terms of IPSS-M and gene mutations. But when we evaluated the ESA-IPSS-M score, which was published last year by Marco Raddi and Valeria Santini, so this score integrates endogenous erythropoietin level and IPSS-M. And there is quite a complex formula to calculate that. But to be simple, when the score is high, the better the response and the longer is the delay of transfusion dependence. So in the early arm, it was striking to see that this score could predict the patients with a higher response rate and with a longer delay of transfusion dependence. So to summarize, early initiation of ESA did not delay transfusion dependence but it leads to a higher response rate to ESA and a simple molecular score which integrates endogenous erythropoietin level can predict a subset of patients who will benefit from this early initiation of ESA. So these are the results of our poster today. And we analyzed the transcriptomic cytokine profile and the site of analysis of these patients included in the EPO-PRETAR trial. And the take-home message is that transcriptomic signatures showed that the early arm responders have an increase of gene expression in the interferon gamma signature. And in the late arm, there are some signatures correlated with inflammation. And these results were confirmed by a cytokine profile, where early arm responders secrete interferon gamma with T-cells and NK cells and whereas in late arm non-responders you have an inflammatory milieu with IL-6 and IL-1. So therefore it seems that the early responders have an adaptive, favorable immune microenvironment, whereas the non-responders have an inflammatory milieu which impedes ESA to lead to response.