Myelodysplastic neoplasms (MDS) are heterogeneous clonal disorders with increased risk of transformation to acute myeloid leukemia. The HMGA2 gene, a chromatin-binding regulator suppressed by the let-7 miRNA family, has been implicated in malignant transformation and fibrosis. In this study, HMGA2 was investigated in patients with MDS, while four members of the let-7 family (let-7a, let-7b, let-7c, let-7d) were specifically assessed in those with bone marrow fibrosis. HMGA2 showed no association with established prognostic scores or survival, while in MDS with bone marrow fibrosis, distinct let-7 patterns emerged: let-7a associated with cytopenias and reduced platelet counts, let-7c with blast percentage and ineffective hematopoiesis, while let-7d displayed a potentially protective profile with higher platelet count and fewer cytopenias. A novel positive correlation between HMGA2 and let-7d was also identified. These findings suggest that let-7 miRNAs may serve as prognostic biomarkers and therapeutic targets in fibrotic MDS, whereas HMGA2 appears less predictive.

Background-Objectives: Although considered standard of care for patients with low-/intermediate-1 risk MDS and isolated del(5q), lenalidomide is not widely used in patients exhibiting additional cytogenetic abnormalities, on top of del(5)q. The aim of this study was to provide real-world evidence for the efficacy of lenalidomide in patients with del(5q), with or without additional cytogenetic abnormalities. Methods: Patients with MDS exhibiting del(5q) in the Greek National Myelodysplastic Syndromes Registry were analyzed if they had received at least one lenalidomide dose and detailed response assessment/follow-up was available. Results: Among 238 patients analyzed, 153 (64.3%) had del(5q) syndrome (Group-I), 34 (14.3%) had an isolated del(5q) abnormality but were not 5q− syndrome (Group-II), 26 (10.9%) had del(5q) plus only one additional cytogenetic abnormality (Group-III), and 25 (10.5%) had del(5q) plus >1 additional abnormality (Group-IV). Among 218 (91.6%) evaluable patients, a major response was achieved by 146 (67.0%) patients, 114/146 (78.1%) in Group-I, 18/31 (58.1%) in Group-II, 10/20 (50.0%) in Group-III, and 4/21 (19.0%) in Group-IV. Overall, hematological response was seen in 177/218 (81.2%) patients, even among those with an excess of bone marrow blasts/frank acute myeloid leukemia. Duration of response was comparable between the four patient groups. A complete cytogenetic response was achieved by 38.0% overall, more commonly in Group-I (42.3%) and -III (35.7%). Transfusion-independent patients and those with a higher MCV or lower marrow blast cells at baseline had a higher probability of achieving a major response. With multivariate analysis, factors associated with overall survival were age, performance status, transfusion dependence, and marrow blast cell percentage at treatment start, as well as time from initial diagnosis to lenalidomide start. Conclusions: Lenalidomide was highly effective in patients with the del(5)q syndrome and also in those with isolated del(5)q, other than del(5)q syndrome, or those exhibiting del(5)q plus only one additional cytogenetic abnormality, not affecting chromosome 7.

Luspatercept has been approved for the treatment of anaemia in transfusion-dependent (TD) patients with lower risk (LR) myelodysplastic syndromes (MDS) after erythroid-stimulated agent (ESA) failure, according to the results of the MEDALIST trial. In this multicentre retrospective study, we report efficacy and safety data of luspatercept administered in 98 TD LR-MDS patients after ESA failure. The percentage of patients that stopped luspatercept due to adverse events was comparable to that reported in the MEDALIST study. Furthermore, we observed that 44.3% patients who had completed 24 weeks of follow-up achieved transfusion independence lasting longer than 8 weeks, compared to 38% in the MEDALIST trial. These positive results may be attributed to the inclusion of patients with lower transfusion needs in our study. All responses were observed within 8 months since luspatercept onset and many were long-lasting, even in the high-transfusion burden patient group. In addition, response to luspatercept and the presence of less than two mutations independently predicted for longer overall survival. Overall, our results confirm luspatercept's safety and efficacy in TD LR-MDS patients who have experienced ESA failure in a real-world setting.

Introduction Patient medication adherence is essential for the effectiveness of therapeutic interventions. Patients' perceptions of the necessity of the prescribed treatments and concerns about the potential adverse effects are modifiable psychological factors that can influence adherence to treatment recommendations. Primary immune thrombocytopenia (ITP) is defined by immune-mediated platelet destruction, resulting in thrombocytopenia and an elevated risk of bleeding. Therapeutic strategies focus on controlling acute bleeding and mitigating long-term hemorrhagic risk. However, little is known about adherence behavior or the influence of medication beliefs on adherence in patients with ITP. Therefore, we investigated medication adherence and beliefs about medicines in this population and examined the relationship between them. Methods This multicenter, cross-sectional study included 135 adult patients with ITP recruited from 12 hematology outpatient clinics across Greece. Eligible participants had been diagnosed with chronic or persistent ITP according to internationally recognized criteria and had been self-administering prescribed treatment for their condition for at least one month prior to study enrollment. Participants were asked to complete a demographic questionnaire, the five-item Medication Adherence Report Scale (MARS; score range: 5–25), and the Beliefs about Medicines Questionnaire Specific (BMQ; score range: 5–25 per subscale). Clinical data were collected by treating physicians. All variables were summarized using descriptive statistics. Multiple linear regression analysis was performed to identify factors associated with the total MARS score. Statistical significance was set at α = 0.05. Results The mean (SD) age of the participants was 57.1 (17.6) years; 62.2% were female and 80.7% reported not living alone. The median (min, max; IQR) duration of the disease and the duration of the treatment were 43 (3, 350; 97) and 10 (1, 156; 25) months, respectively. The most common route of medication administration was oral (79.9%) and the most common dosing schedule was daily or more than once/week (79.3%). The median (min, max; IQR) platelet count was 112.000 (1.000, 450.0000; 101.000). Overall, 91.1% of patients were in chronic phase and were receiving thrombopoietin receptor agonists. The mean (SD) total MARS score was 23.3 (2.5). Regarding specific MARS items, 94.1% of patients reported that they never or rarely forget to take their ITP medication and 91.8% reported that they never or rarely take less medication than prescribed. In contrast, 38.4% of participants reported that they always, often or sometimes adopted one or more non-adherent behaviors. The mean (SD) total scores of the BMQ necessity and BMQ concerns subscales were 19.5 (4.0) and 14.8 (4.6), respectively. The mean (SD) necessity-concerns differential (NCD) score was 4.7 (5.2). Regarding individual necessity items, 88.2% of participants were in agreement with the fact that their medicines protect them from becoming worse and 82.2% were in agreement with the fact that their health, at present, depends on their medicines. Regarding specific concern items, 66.7% of the participants were in agreement with the fact that they sometimes worry about long term effects of their medication and 48.1% were in agreement with the fact that having to take medicines worries them. Multiple linear regression models adjusted for age, gender, living conditions, disease duration, treatment duration and route of administration demonstrated that total BMQ necessity subscale score was significantly positively associated with total MARS score (β = 0.143, p = 0.017). Conclusions This study found a high level of medication adherence and strong necessity beliefs about medicines among patients with ITP. The positive NCD score suggested that patients' beliefs about the necessity of their medication were stronger than their concerns. However, more than half of participants were concerned about the long-term effects of their medications. A positive association between necessity beliefs and adherence was also observed, suggesting that patients with stronger necessity beliefs about their treatment tend to be more adherent to their prescribed regimen. Longitudinal and interventional studies are required to explore whether strengthening patients' beliefs about the necessity of their prescribed medication could improve adherence behavior in ITP patients.

Introduction. Luspatercept has been recently approved for the treatment of red blood cell (RBC) transfusion-dependent (TD) Lower-Risk (LR) Myelodysplastic Syndrome (MDS) patients and can be administered either in the frontline or after erythropoietin failure. Our aim is to report its efficacy and safety in this study population in the everyday clinical practice using the well-defined IWG 2018 criteria for hematological response. Methods. We retrospectively studied LR, TD MDS patients from 15 Greek and 2 Turkish sites since October 2020. Transfusion burden was defined as high (HTB), intermediate (ITB) and low (LTB) if they had received ≥8, 3-7 or 1-2 RBC transfusions in the last 16 weeks before luspatercept onset respectively. Response was defined as transfusion independence (TI) for at least 8 weeks. Therefore, date of response was set at 8 weeks after the date of last recorded transfusion or after the date of luspatercept onset, if no more RBC transfusions were administered afterwards. If relapses were noted before the completion of 16 weeks since these timepoints, those responses were not considered meaningful and were excluded. All variables were reported as median (5-95% range). Fisher's exact test and Wilcoxon signed-rank test were performed for comparison of discrete and continuous variables respectively and the Kaplan-Meier method with log-rank test was used for time-to-event data. Results. A total of 98 patients, 59 male and 39 female, aged 75 ± 8 years were studied. 80 patients had bone marrow ring sideroblasts (RS) and 18 did not. In addition, 95 patients had prior erythropoietin failure and 3 were erythropoietin naive. There were 48 HTB, 37 ITB and 13 LTB patients. Out of 54 patients with available molecular results, 39 were SF3B1 positive and 15 were SF3B1 negative. 74 patients received maximum luspatercept dose (1.75 mg/kg), 12 received intermediate dose (1.33 mg/kg) and 12 received minimum dose (1 mg/kg). Median number of luspatercept doses was 14 (5-95% range: 4- 54). Median follow-up time was 13.1 (4.2-39,8) months and 20 deaths were observed within this time period. During follow-up, 57 patients stopped treatment, 8 due to adverse events (2 due to musculoskeletal pain, and the rest due to vertigo, hypertension, orthostatic hypotension, soft tissue infections, bleeding diathesis and one undetermined case), 8 because of death unrelated to treatment, 2 proceeded to transplantation and 36 stopped due to inadequate response. Notably, 3 patients stopped because of persistent hemoglobin levels>11g/dL. Out of 98 patients, 42 responded to treatment, 25% in the HTB, 57% in the ITB and 69% in the LTB group respectively (p=0.002). All LTB who became TI also exhibited >1,5g/dL increase in hemoglobin. There was 48.7% response in the SF3B1 positive vs 20% response in the SF3B1 negative group (p=0.051). HTB remained significant after controlling for SF3B1 status (Mantel-Haenszel p=0.002). On the other hand, there was no difference in response in the non-RS group (33,3%) compared with the RS group (45%, p=0.366). 29 and 33 patients who responded had ferritin and lymphocyte levels available upon response respectively. There was a significant decrease in ferritin upon response vs baseline (691 (52, 4375) vs 880 (93, 6291) ng/dL, p=0.008) and a significant increase in lymphocytes (1700 (748, 3902) vs 1635 (453, 2972) cells/uL p=0.05). For the 42 responders, time to response was 14.8 (8, 30) weeks and duration of response was 44.9 (4.3, 170.4) weeks at the time of data collection. 10 out of 42 patients lost response after having experienced 59 (10.3- 142) weeks of TI. Overall, patients who achieved response had better chances of survival (log-rank p=0.048). Conclusions. Luspatercept is tolerable and effective in the treatment of LR-TD MDS patients. Response depends mainly on transfusion burden at baseline and the presence of SF3B1 mutations. Most responses occur within the first six months of treatment and they are long lasting. Response to luspatercept treatment favorably affects overall survival in TD LR MDS patients.

Nonarteritic anterior ischemic optic neuropathy (NA-AION) is the most common form of acute painless, usually unilateral, optic neuropathy in the elderly population. Systemic risk factors include diabetes mellitus, arterial hypertension, sleep apnea syndrome, and cardiovascular disease. Alpha 52-year-old man developed sequential, bilateral NA-AION, involving a worsening, severe vision compromising attack in the secondly affected eye. Thrombophilia testing revealed compound heterozygosity for the C6777T mutation of the MTHFR gene and the prothrombin G20210A (FII mutation). Oral anticoagulation treatment was initiated. A thorough systemic and family history, especially in the absence of major vasculopathic disease, should alert toward investigation for thrombophilia in middle-aged patients with atypical forms of NA-AION and initiation of anticoagulant treatment should be considered.

MicroRNAs (miRNAs),are significant regulators of human hematopoietic stem cells. Their deregulation contributes to hematological malignancies.The let-7 family has been found frequently deregulated in malignancies.In MDS various alterations of miRNAs have been reported.High Mobility Group AT-Hook 2 (HMGA2) protein functions as a transcriptional regulator. In this study, we investigated the HMGA2 expression in MDS and specifically in patients with fibrosis we studied the prognostic significance of four members of the let-7 family (let-7a, let-7b, let-7c, let-7d). RNA extraction, reverse transcription, anda SYBR Green based real-time PCR were performed for the absolute quantification of HMGA2, using standard protocols. After RNA polyadenylation and reverse transcription with an oligo-dT adapter primer, miRNAs transcript levels were determined using the SYBR Green chemistry. IBM SPSS statistics, version 26 (IBM Corporation, North Castle, NY, USA) was used for the analysis. HMGA2 gene expression was investigated in 78 patients with MDS, whereas transcript levels of four members of the let-7 family (let-7a, let-7b, let-7c, let-7d) were analyzed in 11 patients with fibrosis. Let-7a transcript levels were significantly higher in MDS patients who developed acute myeloid leukemia (AML) compared to the group that did not (p=0.0141). Let-7d presented a negative correlation (p=0.0408). A moderate (p =0.0483) negative correlation of HMGA2 with let-7c, and a strong positive correlation (p =0.0481) with let-7d, were observed. In literature, the let-7/HMGA2 linkage could be a signature in MDS pathogenesis. Let-7a level was found higher in transformation to AML,defining it as a poor prognostic factor, in contrast with the protective role of high let-7d.

Significantly higher basal levels of Stat3 in patients with poor-risk cytogenetics.

A high relapse rate for patients with acute myeloid leukemia (AML) is still a major barrier to the long-term survival of these patients. Nevertheless, considerable progress has been made both in the biology and therapy of the disease. Specifically, progress has been made in the areas of integrated genomic analysis for prognosis, the widening application of minimal residual disease (MRD) monitoring in clinical practice, the development of new agents, and the increasing use of drugs, such as IDH and FLT3 inhibitors, as a bridge to transplant. Continued progress and inquiry into these and other areas are essential to improve the survival outcome for adult patients with AML.

Topic: 32. Platelet disorders Background: ITP can be classified as primary or secondary based on the absence or presence of other causes or disorders associated with thrombocytopenia. Although these forms of ITP have been traditionally considered as distinct disease entities, comparative studies are rare and the differences and similarities between primary and secondary ITP have not been adequately elucidated thus far. Aims: To comparatively describe real world characteristics at diagnosis and evaluate disease outcome in patients with primary (Group-1) and secondary ITP (Group-2), using data from the national database (ITP registry) operated under the auspices of the Hellenic Society of Hematology. Methods: The Greek ITP registry recruits patients (n=1573) nationally through a network of 25 sites. In the present study, we retrospectively analyzed data from patients with primary and secondary ITP, aged over 18 years who were diagnosed from 1979 to 2022. Results: The number of evaluable patients was 580. Group-1 consisted of 464 (80%) and Group-2 of 116 (20%) patients. There were no differences in the median age, gender, the median platelet count or other hematological parameters between the 2 groups at diagnosis. Significantly fewer patients in Group-1 had comorbidities as compared to Group-2 (P=0.0206). Relevant medications were concurrently used at similar rates by the 2 patient groups. Bleeding manifestations were observed at a similar frequency between the 2 groups. Of those who were diagnosed as secondary ITP, 26.8% had infections, 24% had collagen disorders, 8.3% had Evans Syndrome, 5.5% had H.Pylori infection, 4.6% had Hashimoto thyroiditis and 3.6% antiphospholipid syndrome. In 3.7% of Group-2 patients ITP was drug-induced. Antiphospholipid antibody (Ab), platelet associated Ab, antinuclear Ab, HIV Ab and Rheumatoid Factor testing were performed more frequently in Group-2 patients. A significant higher proportion of Group-2 patients had a positive test for antinuclear antibodies, HCV Abs, antiphospholipid Abs and H.Pylori (P=0.0185, P=0.0004, P=0.0002 and P<0.0001, respectively). Fewer Group-1 patients received treatment at diagnosis (P=0.023). The choice of treatment at diagnosis did not differ between Group-1 and Group-2 patients with the exception of anti-CD20 which was administered more frequently in Group-1 patients (P=0.036). Overall response rates did not differ between the two groups. At 6 months post diagnosis significantly fewer Group-1 patients had bleeding manifestations and fewer Group-1 patients were treated with corticosteroids, IVIG, corticosteroids and IVIG and anti CD20 (P<0.0001, P=0.0144, P=0.0005 and P=0.039, respectively). At 6 months post diagnosis significantly fewer Group-1 patients had developed persistent ITP (p<0.005). Summary/Conclusion: The majority of adult Greek ITP patients suffered from primary ITP (Group-1). Secondary ITP (Group-2) was predominantly associated with collagen disorders and infections. Primary and secondary ITP patients did not differ in demographics, bleeding manifestations, and hematological parameters. However, patients with secondary ITP had a higher frequency of auto Abs. The form of ITP did not impact on the choice of initial treatment or the response rate but influenced the diagnostic work-up and the outcome of the disease, as suggested by the increased proportion of secondary ITP patients requiring treatment and developing persistent ITP at 6 months post diagnosis. Further investigation is warranted to probe more deeply into the differences and similarities of primary and secondary ITP in order to optimize management. Keywords: ITP, Immune thrombocytopenia (ITP)

Multivariate analysis of prognostic factors for overall survival in the patient cohort (n=74)

Flow cytometric analyses of cellular hierarchy in Lin-CD34+ selected cells.

Pretreatment signaling profiles in clonal and non-clonal CD34+ progenitors.

Multivariate analysis of prognostic factors for event-free survival in the patient cohort (n=74)

Mutations analysis of the TET2 and TP53 coding regions using Next-Generation Sequencing (NGS)

Accurate risk stratification of myelodysplastic neoplasms (MDS) is a key feature of treatment selection and predicts overall survival (OS). Taking into consideration several disease and patient-related characteristics, various prognostic scoring systems have been formulated. On the other hand, the elevated mean corpuscular volume (MCV) of the erythrocytes [macrocytosis, defined as an MCV over 100 femtolitres (fL)] is a classical feature of MDS and is considered a marker of dysplasia in patients with anemia and no other identifiable causes of macrocytosis, while it has been even used in a scoring system to establish the probability of MDS.1 Nonetheless, macrocytes can be classically found in patients with megaloblastic anemia, hypothyroidism, alcohol abuse, liver disease, and other bone marrow failure syndromes. The prevalence of macrocytosis in patients with MDS is considered high but has never been accurately estimated in large cohorts and its pathophysiological and prognostic roles have not been thoroughly evaluated. In the present study we investigated the prognostic role of macrocytosis in patients with MDS registered in the Hellenic 5-azacytidine Registry which is part of the Hellenic (Greek) National Registry of Myelodysplastic and Hypoplastic Syndromes comprising patients from 28 centers meeting the 2008/2016 World Health Organization criteria for MDS. The study was approved by the Hellenic Society of Hematology. The patients were stratified per the International Prognostic Scoring System (IPSS)2 and the revised IPSS (IPSS-R).3 The main hematologic variables, including MCV, were recorded at diagnosis and before any therapeutic interventions (such as red blood cell transfusions or erythropoiesis stimulating agent administration). Macrocytosis was defined as an MCV over 100 fL. Patients with an MCV lower than 80 fL as well as patients with a serum ferritin level suggestive of iron deficiency (<30 ng/mL) were excluded from the analysis, as well as patients with beta thalassemia trait which is a common disorder in Greece4 and megaloblastic anemia due to B12 vitamin or folate deficiency. The Pearson chi-square test, the independent-samples Mann–Whitney U test, and Kaplan–Meier analysis were used appropriately. Multivariate Cox regression models were built to assess association between categorical variables and the relative risk for death. The level of significance was set at a probability value lower than 5% (2-sided p < .05). The study comprised 833 patients, 73 of whom were excluded from the analysis due to a low MCV (N = 67, corresponding to 8.1% of the whole cohort, a rate compatible with the prevalence of beta thalassemia trait in Greece)4 or low ferritin levels (N = 6). The main characteristics of the final population of 760 patients are shown in Table S1. Macrocytosis was identified in 257 (33.8%) patients at diagnosis. As shown in Table S2, among the baseline characteristics of the patients, macrocytosis correlated only with higher platelet counts (123 × 109/L vs. 100 × 109/L for those without macrocytosis, p = .007). The median time from diagnosis to treatment initiation with 5-azacytidine was significantly shorter in patients without macrocytosis (3.2 months vs. 5.7 months for patients with macrocytosis; p = .015). Macrocytosis was correlated with an increased OS (median OS of 44.1 months vs. 33.5 months for those without macrocytosis, p = .002, Figure 1A), while there was a non-significant trend for lower acute myeloid leukemia (AML) transformation rate in patients with macrocytosis (22.7 months vs. 17.5 months for those without macrocytosis; p = .126).Multivariate analysis showed that macrocytosis was an independent predictor of a longer OS when tested against IPSS-risk or thrombocytopenia (Table S3). After stratification of the severity of macrocytosis (MCV, 100–110 fL vs. >110 fL) it was found that it did not have an impact on OS, or AML transformation rate. In lower-risk patients (N = 385), macrocytosis was identified in 128 (33.2%) and correlated with the absence of thrombocytopenia (p = .013), higher neutrophil counts (2.13 × 109/L vs. 1.81 × 109/L for patients without macrocytosis, p = .015), and a lower bone marrow blast percentage (3.0% vs. 4.5% for patients without macrocytosis, p = .001). Finally, it correlated with a longer OS (60.6 months vs. 47.4 months for patients without macrocytosis, p = .008, Figure 1B) and a longer time interval between diagnosis and 5-azacytidine initiation (16.8 months vs. only 7.2 months for those without macrocytosis, p ≤ .0001). Macrocytosis retained once again its statistical significance as a prognostic factor of OS when tested against neutropenia, low bone marrow blast count (≤1 vs. >1), and the IPSS risk (Table S3). In patients with higher risk MDS (N = 375), macrocytosis was identified in 129 (34.4%) and was not correlated with any of the baseline variables. Nevertheless, higher risk MDS patients with macrocytosis had longer OS than patients without macrocytosis (32.4 months vs. 23.0 months, p = .044, Figure 1C) and a longer time to AML transformation (15.7 months vs. 14.2 months for those without macrocytosis; p = .013). In patients with complex karyotype (n = 107), macrocytosis was present in 34 (31.8%) patients and was not correlated with any of the baseline variables, but it was once again correlated with a longer OS (28.7 months vs. 16.1 months for those without macrocytosis, p = .001, Figure 1D), although it was not correlated with the diagnosis-to-treatment time interval (p = .843). This survival correlation was independent of the IPSS (Table S3). In patients with a poor karyotype per the IPSS (n = 146), no correlations with any of the tested variables were identified, except for OS (27.3 months in macrocytic vs. 16.1 months in non-macrocytic patients, p = .006). Details on the cytogenetics of the patients are presented in Table S4. Regarding response to treatment with 5-azacytidine, patients with macrocytosis had a response rate (including complete, remission, partial remission, and hematologic improvement per the International Working Group response criteria for MDS) of 56.8%, while patients without macrocytosis had a response rate of 46.2% (p = .026). Macrocytosis is a classically described feature of MDS reflecting dyserythropoiesis and a lower number of mitotic divisions during the maturation process of the early erythropoietic cells. Although macrocytosis has been repeatedly reported as a prominent feature of patients with MDS, no quantification of this correlation has been attempted so far in the context of a study. Here we establish for the first time the rate of macrocytosis in a large cohort of patients with MDS (33.8%). This rate, although significant, does not reflect the prevailing idea that macrocytosis is a universal feature of MDS. In this cohort, there was a clear lack of correlation of macrocytosis with other features defining prognosis, such as the IPSS score and cytopenias except for the platelet count. Platelets along with red blood cells arise from the same progenitor cell. Because macrocytosis may reflect the asynchronous cytoplasm and nucleus maturation when dysplasia is present, the increased platelet count associated with macrocytosis observed in this study could be explained by the effort of this progenitor cell to overcome dysplasia and the increased cell turnover leading to increased bone marrow cellularity. An analogous process drives iron deficiency anemia-associated thrombocytosis. Interestingly, macrocytosis was not correlated with the severity of anemia or the need for transfusions, highlighting once again its independent prognostic nature, but it was correlated with a shorter time interval from diagnosis to treatment initiation, as well as a higher OS in the whole cohort, but also in both lower and higher risk patients. This correlation with survival was found to be independent of the IPSS categorization, highlighting macrocytosis as a strong independent factor of a better prognosis in patients with MDS, especially in lower risk patients, indicating which of them will progress and will have to be treated with a hypomethylating agent. These results consort with the results of a study on 321 patients with MDS showing that normocytosis was an independent adverse prognostic factor for patients with low blast (<5%) MDS, even when tested against clinical, laboratory, and mutation variables.5 The strong correlation of macrocytosis with higher OS probably reflects a more benign phenotype of patients with macrocytosis. The correlation of macrocytosis with erythroid dysplasia may offer a lead to explain this association, probably suggesting that macrocytes derive from erythroblasts surviving the apoptotic processes, implying that less apoptosis, hence more effective erythropoiesis, may lead to more pronounced macrocytosis, while the maturation arrest that accompanies the leukemic phenotype is correlated with less macrocytosis. Understanding the modulators of enuclation and organelle clearance during erythropoiesis may elucidate the pathogenesis and prognostic significance of dyserythropoiesis. In patients with poor cytogenetic features, macrocytosis retained its significance as an independent prognostic factor of a longer OS, and this is in accordance with the results of a study on 164 patients with MDS and abnormal karyotypes, where macrocytosis was independently correlated with a longer OS.6 No further reports on the prognostic role of MCV/macrocytosis in MDS were identified in the literature. The strength of the present study relies on the large number of well-characterized patients with MDS, permitting the identification of the prevalence of macrocytosis in MDS as well as its prognostic significance. Moreover, an effort to exclude patients with known causes of microcytosis/macrocytosis (such as iron deficiency, beta thalassemia trait, and megaloblastic anemia) was made to avoid any confounding factors, but factors such as alcohol consumption and medications, as well as acquired alpha-thalassemia in MDS were not addressed. Moreover, the lack of molecular data in the registry is another limitation of the analysis. A potential limitation of the study is that it comprises a preselected population of patients eventually treated with 5-azacytidine, although these patients were not selected based on their prognostic or any other characteristics. In conclusion, macrocytosis, evident in about a third of patients with MDS, appears to be a significant independent prognostic factor for a steadier clinical course, longer time from diagnosis to treatment, and a higher OS in patients with MDS. Further validation of these results may prompt the inclusion of macrocytosis in the available risk stratification systems, to help patient selection for treatment and refine prognostic categorization. The authors declare no conflicts of interest. The data that support the findings of this study are available from the corresponding author upon reasonable request. Table S1. Baseline patient characteristics Table S2. Characteristics of patients with and without macrocytosis. Table S3. Cox Regression analysis for overall survival with several models including macrocytosis along with other variables that proved to be statistically significant in univariate analysis. (A) Whole cohort. Model comprising two variables: macrocytosis (MCV >100 fL) and IPSS category (higher vs. lower). (B) Whole cohort. Model comprising two variables: macrocytosis (MCV >100 fL) and thrombocytopenia (platelet count <100 × 109/L). (C) Lower risk patients. Model comprising three variables: macrocytosis (MCV >100 fL), absence of neutropenia (neutrophil count over 1.5 × 109/L) and low bone marrow blast count (≤1%). (D) Lower-risk patients. Model comprising two variables: macrocytosis (MCV >100 fL) and IPSS category (low vs. intermediate 1). (E) Complex karyotype group. Model comprising two variables: macrocytosis (MCV >100 fL) and IPSS category (low vs. intermediate 1). CI, confidence intervals; IPSS, international prognostic scoring system; MCV, mean corpuscular volume Table S4. Cytogenetic data of patients with and without macrocytosis. Data refer to the most common cytogenetic abnormalities, while less common abnormalities (found in <10 patients) are not shown. The columns of multivariate analysis show the results of Cox Regression analysis for overall survival using models comprising macrocytosis and each of the individual cytogenetic abnormalities. The distribution of patients in the different 2016 WHO-defined categories for MDS is also shown. The hazard ratio (HR) and the p value refer both to macrocytosis. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.