S136 Poster Presentations – 14th International Symposium on Myelodysplastic Syndromes / Leukemia Research 55 S1 (2017) S45–S167 <5% blasts from TNCs and ≥15% ring sideroblasts, fulfilling current definition for RARS (WHO 2001 and 2008) and RCMD-RS (WHO 2001). Moreover 1,233 patients with ≥5% ring sideroblasts and <5% blasts were analyzed in order to explore the future definition of WHO 2016, that considered as MDS-RS those patients with 5–<15% ring sideroblasts if SF3B1 mutation is present. This was a tentative analysis since SF3B1 mutation information was not available in our series. Percentage of BM blasts from NECs was calculated as follows: [%blasts from TNCs/(100-%erythroblasts) × 100]. Survival curves and univariable and multivariable Cox regression analysis were implemented. Results: Median age at diagnosis was 76 y (25–101 y). Median follow-up was 50.1 months and median OS was 96.5 months. Enumerating blasts from NECs, 10% of MDS-RS were reclassified into categories with ≥5% blasts and showed a poorer OS than did those who remained in initial categories (median OS, 68.1 vs 97.6 months, P = 0.025; Hazard ratio (HR): 1.41; 95%CI: 1.04–1.91; P = 0.026). After adjusting the analysis by IPSS cytogenetics, the prognostic impact of blasts considered from NECs maintained its significance (HR: 1.37; 95%CI: 1.01–1.85; P = 0.045). Similar results were observed applying this method to the group of MDS patients with ≥5% ring sideroblasts and <5% blasts. By considering BM blasts from NECs, 10% of these patients were reclassified into categories with ≥5% BM blasts and showed a poorer OS than did those who remained in initial categories (median OS, 60.2 vs 85.8 months, P = 0.003; HR: 1.51; 95%CI: 1.15–1.97; P = 0.003; HR adjusted for IPSS cytogenetics: 1.46; 95%CI: 1.12–1.92; P = 0.006). Conclusion: considering bone marrow blasts from nonerythroid cells improves the prognostic evaluation of MDS with ring sideroblasts. 236 IMPROVING RISK STRATIFICATION OF MDS PATIENTS BY THE INTEGRATION OF RBC-TRANSFUSION DEPENDENCY INTO THE REVISED INTERNATIONAL PROGNOSTIC SCORING SYSTEM M. Arnan1, H. Pomares1, I. Sánchez-Ortega1, E. Alonso2, M. Encuentra3, X. Grau4, A. Sureda1 1 Hematology Department, Hospital Duran i Reynals. Intitut Català d’Oncologia, Barcelona, Spain; 2Patology Department, Hospital de Bellvitge, Barcelona, Spain; 3Hospital Duran i Reynals. Intitut Català d’Oncologia, Clinical Research Unit, Barcelona, Spain; 4Department of Laboratory Hematology, Hospital Universitari Germans Trias i PujolICO- Josep Carreras Leukemia Research Institute, Badalona, Spain Background: IPSS was revised (R-IPSS) to better define the prognostic impact on OS and PFS for patients with MDS. Although RBC-transfusion dependency (RBC-TD) is associated with poor prognosis, it was not included in R-IPSS because of the relatively low proportion of patients with available data. We hypothesize that the use of RBC-TD status may improve the R-IPSS prognostic assessment. Aim: To assess the impact of RBC-TD in addition to R-IPSS in predicting survival outcome. Methods: We identified 818 untreated MDS, AML (20–30%blasts) and CMML patients. RBC-TD was defined as ≥1 RBC transfusion every 8 weeks over a 4-months period (Malcovati JCO 2007). Patients were classified into two groups according to RBC-TD at time of or before the landmark time point. Landmark analyses were conducted at 6, 12, 24 and 36 m after diagnosis. Patients experiencing the event (i.e. death) before the landmark time point were excluded. Results: Median age: 72 y (29–101). 65% males. R-IPSS categories: 29.9% Very Low, 44% Low, 15.8% Intermediate, 4.7% High and 5.7% Very High risk. Median OS was lower in males (4.4 vs. 6.6y; p < 0.001) and in elderly patients (10, 5.7 and 2.9 y in <65, 65–80, >80 y respectively [ p < 0.001]). According to IPSS-R, median OS was 8, 6.8, 3.9, 1.8 and 0.6 y in very low, low, intermediate, high and very high risk groups respectively ( p < 0.001). In multivariate analysis, RBC-TD was associated with poor survival independent of IPSS-R, gender and age (Table 1). Landmark analysis demonstrated that median survival of RBC-TD patients at 6 m was significantly lower compared to RBC-transfusion independent ones (3.2 vs. 8.5 y; p < 0.001). Similarly, RBC-TD patients had lower OS at 12 m (3.6 vs. 9.2 y; p < 0.001), 24 m (4.8 vs. 11 y; p < 0.001) and 36 m (5.8 vs. 11.9 y; p < 0.001). Analyzing the impact of RBC-TD into three different R-IPSS categories (very low and low vs intermediate vs high and very high), significantly lower median survival was observed in lower and intermediate R-IPSS categories at diagnosis and at each landmark analysis. In contrast, RBC-TD did not impact in survival of high and very high R-IPSS patients (Table 2). Conclusions: Our single-centre results confirm the negative impact on survival of TD at diagnosis and at any time during MDS disease, especially in those classified as Very Low, Low and Intermediate R-IPSS patients. As therapeutical decisions are based on the initial prognostic risk assessment, the inclusion of transfusion dependency on the risk stratification may provide more precise prognostic information with impact on the therapeutic approach. 237 PREDICTION OF DISEASE PROGRESSION USING MUTATIONAL SCREENING IN PATIENTS WITH MYELODYSPLASTIC SYNDROME M. Belickova1, J. Vesela1, H. Votavova1, M. Vostry1, A. Jonasova2, Z. Zemanova3, J. Brezinova4, J. Cermak5 1 Genomic, Institute of Hematology and Blood Transfusion, Prague, Czech Republic; 2Ist Dept Medicine, General Hospital, Prague, Czech Republic; 3General Hospital, Center of Oncocytogenetic, Prague, Czech Republic; 4Cytogenetic, Institute of Hematology and Blood Transfusion, Prague, Czech Republic; 5Clinical, Institute of Hematology and Blood Transfusion, Prague, Czech Republic About 90% of MDS patients carry at least one somatic mutation and their outcome can vary widely; therefore, we focused on the identification of specific mutations predicting disease progression. We examined paired samples from 34 patients by TruSight Myeloid Sequencing Panel (Illumina) containing 54 genes for determination of mutational profile. Patients were evaluated at the time of Poster Presentations – 14th International Symposium on Myelodysplastic Syndromes / Leukemia Research 55 S1 (2017) S45–S167 diagnosis and during progression to more advanced MDS subtype or to AML. A total of 53 mutations at diagnosis and 78 mutations in progression across 54 genes were identified in the patient cohort. Patients in progression carried at average 2.3 mutations, three patients harboured five mutations and only two patients had no detectable mutation. The variant allele frequency (VAF) of mutations detected at the time of diagnosis increased during disease progression, except of five mutations which were suppressed by the expansion of clones carrying other mutations. Fifty-six percent of patients carried mutations in RUNX1, TP53, PTPN11, SETB1 or NRAS gene at the time of diagnosis and another 20% of patients evolved these mutations in the course of disease. The median overall survival (OS) was 30.6 months in patients with this combination of mutations compared to 84.5 months (HR 3.1; p = 0.01) in patients without mutations. The most common acquired mutations at the time of progression were found in PTPN11, SETB1 and NRAS genes. Overall survival did not depend on the number of mutations detected at the time of diagnosis, but on the specific mutated gene. Comparison of mutational profiles of MDS patients at diagnosis and during progression suggested candidate genes involved in AML transformation. Our study demonstrated that mutated RUNX1, TP53, PTPN11, SETB1 and NRAS genes were most associated with the disease progression. Mutations in these genes were often detectable with low VAF already at the time of diagnosis and in some patients were detectable until the time of progression. The results indicate that routine monitoring of mutations in MDS should be performed to refine the risk prediction of disease progression. Acknowledgement: Supported by AZV grants (16-33485A and 1631689A) and the project for conceptual development of research organization (00023736) from the Ministry of Health of the Czech Republic. (NECs). WHO-2008 recommended considering blasts from TNCs, whereas in 2016 WHO-revision, NECs count rule was eliminated. Recent data evidenced that considering blasts from NECs improves outcome prediction in MDS when assessing IPSS and WHO classification (Arenillas et al., J Clin Oncol 2016). Aim: To evaluate whether calculating IPSS-R by enumerating blasts from NECs improves outcome prediction in MDS-E. Methods: We analyzed 498 de novo MDS-E diagnosed according to WHO 2008 from the GESMD. Percentage of BM blasts from NECs was calculated as follows: [%Blasts from TNCs/(100-% erythroblasts) × 100]. Survival analyses were performed. C-statistics for censored data (Dxy and CPE) were implemented to assess the method with the best predictive value for OS and time to AML evolution. Results: Median age was 74 y (27–94 y). Median follow-up was 54.08 months and median OS was 59.7 months. We assessed OS and time to AML predicted by IPSS-R considering blasts from TNCs and NECs (recoded IPSS-R). Five groups with significant differences in OS were observed only by applying the recoded IPSS-R (Figure 1). By using our approach, median OS of the intermediate group improved from 28.9 to 41.2 months whereas patients classified into higher-risk categories maintained the outcome expected in these categories (OS <23 months). 41/448 patients (9.2%) classified into lower-risk categories were reclassified into higher-risk ones and showed a significantly shorter OS and time to AML (median OS, 22.6 vs 74.3 months, P < 0.001;median time to AML,74.6 vs N.R. months, P < 0.001) (Figure 2). Likewise, 35/69 patients (51%) classified into the intermediate group were reclassified into higher-risk categories and showed a significantly shorter OS and time to AML (median OS, 13.9 vs 43.7 months, P = 0.029; median time to AML,74.6 vs N.R. months, P = 0.048). Finally, the recoded IPSS-R showed a better capability for predicting OS and time to AML [(survival, recoded IPSS-R vs IPSS-R: 0.374 vs 0.366, 0.654 vs 0.645; Dxy and CPE respectively) (time to AML, recoded IPSS-R vs IPSS-R: 0.592 vs 0.536, 0.700 vs 0.673; Dxy and CPE respectively)]. 238 CALCULATING IPSS-R BY ENUMERATING BONE MARROW BLASTS FROM NONERYTHROID CELLULARITY IMPROVES OUTCOME PREDICTION IN MDS WITH ERYTHROID PREDOMINANCE X. Calvo1, L. Arenillas1, E. Luño2, L. Senent3, M. Arnan4, F. Ramos5, C. Pedro6, M. Tormo7, J. Montoro8, M. Diez-Campelo9, M.L. Blanco10, B. Arrizabalaga11, B. Xicoy12, S. Bonanad3, A. Jerez13, M. Nomdedeu14, A. Ferrer1, G. Sanz3, L. Florensa1 1 Hospital del Mar, Laboratorio de Citología Hematológica, Barcelona, Spain; 2Hospital Central de Asturias, Servicio Hematología, Oviedo, Spain; 3Hospital La Fe, Servicio Hematología, Valencia, Spain; 4ICOHospitalet de Llobregat, Servicio Hematología, Hospitalet de Llobregat, Spain; 5Hospital Universitario de León, Servicio Hematología, León, Spain; 6Hospital del Mar, Servicio Hematología, Barcelona, Spain; 7 Hospital Clínico Universitario de Valencia, Servicio Hematología, Valencia, Spain; 8Hospital Universitario Vall Hebron, Servicio Hematología, Barcelona, Spain; 9Hospital Universitario Salamanca, Servicio Hematología, Salamanca, Spain; 10Hospital de la Santa Creu i Sant Pau, Servicio Hematología, Barcelona, Spain; 11Hospital Universitario Cruces, Servicio Hematología, Baracaldo, Spain; 12ICOBadalona, Servicio Hematología, Badalona, Spain; 13Hospital Morales Meseguer- IMIB-Arrixaca, Servicio Hematología, Murcia, Spain; 14 Hospital Clinic Barcelona, Servicio Hematología, Barcelona, Spain Introduction: IPSS-R is the score with the best prognostic capability in MDS. Some concerns have been generated about the real prognostic significance of the intermediate risk category, since these patients showed an outcome closer to the expected in higher-risk. In MDS with ≥50% BM erythroblasts (MDS-E) there was no consensus on the best method for enumerating BM blasts—from total nucleated cells (TNCs) or from nonerythroid cellularity S137 Fig. 1.