PS1434 GENOMIC LANDSCAPE AND MOLECULAR RISK IN PATIENTS WITH DIPSS LOW‐ AND INTERMEDIATE‐1‐RISK PRIMARY MYELOFIBROSIS: A STUDY OF THE GERMAN STUDY GROUP FOR MYELOPROLIFERATIVE NEOPLASMS (GSG‐MPN)

Background: The ́Dynamic International Prognostic Scoring Systeḿ (DIPSS) incorporates 5 clinical variables to estimate prognosis in primary myelofibrosis (PMF) by stratification of patients into the low‐, intermediate‐1‐ (int1), intermediate‐2‐ (int2), or high‐risk category (Passamonti et al.; Blood 2010). Independent of the DIPSS, PMF patients carrying mutations (mut) in the ́High Molecular Risḱ (HMR) genes ASXL1, SRSF2, EZH2, IDH1  and/or IDH2 have a higher risk of death or leukemic transformation (Vannucchi et al., Leukemia 2013). Therefore, analysis of the genomic landscape in low‐ and int1‐risk PMF patients may refine prognostic scoring and identify those who benefit from earlier therapeutic intervention. Aims: To assess the genomic landscape in DIPSS low‐ and int1‐ risk PMF and to explore the risk of disease progression in patients with an adverse molecular profile. Methods: Using a next generation sequencing approach (NGS) [HaloPlex HS (Agilent); Miseq (Illumina)], we performed a retrospective analysis of the complete coding regions of 42 genes associated with myeloid malignancies, including the HMR genes. 78 patients with low‐ and int1‐ risk PMF enrolled in our German Study Group‐ (GSG‐) MPN Bioregistry (NCT 03125707) were analyzed. Results: Overall, 182 mut were identified in 74/78 patients (95%). Of the three classical MPN driver mut, 50/78 (64%) patients carried JAK2 V617F, 4/78 (5%) had MPL W515 (n = 3) and Y591 (n = 1), and 20/78 (26%) were CALR mutated; co‐existing mut in JAK2 and MPL occurred in two cases; 6/78 (8%) patients were triple‐negative. Of note, 59% of the patients carried non‐driver mut, most commonly TET2 (13%), or DNMT3A (6%) as well as FLT3, KMT2A, KRAS, CBL, CEBPA , and ZRSR2 (2% each). HMR mut (HMR mut ) were identified in 20/78 of the patients (26%) and were more frequently present in the int1‐ (39%) than low‐risk (14%) category (p = .019); mut in ASXL1 occurred in 19% (15/78), SRSF2 in 9% (7/78), EZH2 and IDH2 in 4% each (3/78), and IDH1 in 1% (1/78). 11/78 patients (14%) harbored two HMR mut , of these 10/11 co‐occurred with ASXL1  (5/10 in combination with SRSF2 mut); no patient had more than 2 HMR mut . HMR mut patients harbored more co‐occurring mut than HMR wildtype (HMR wt ) patients (median 4 vs 2; p  < .0001). In addition, HMR mut were associated with the presence of JAK2 V617F (p = .03). However, no unique mutational pattern beside JAK2 V617F was identified for HMR mut and HMR wt patients. We were able to perform an explorative analysis of clinical outcomes in 47/78 patients. In this subcohort, the median follow‐up (FU)‐time was 3.5 years for HMR mut vs 6.5 years for HMR wt patients. Despite the shorter FU more HMR mut patients 5/13 (38%) than HMR wt patients 4/34 (12%) shifted to a higher DIPSS category (n = 8) or experienced leukemic transformation (n = 1) [p = .09]. All affected HMR mut patients carried either ASXL1 (n = 2) or SRSF2 (n = 2) or both mut ASXL1/SRSF2 (n = 1). Summary/Conclusion: Using a targeted sequencing approach we identified mutations in 95% of 78 DIPSS low‐ and int‐1‐ risk PMF patients. 26% of these harbored HMR mutations, underlining the clinical importance of HMR marker screening for the identification of patients potentially at higher risk. However, prospective studies with longer observation periods are needed to demonstrate the clinical impact of prognostically adverse HMR mutations. Such studies can provide the rationale for evaluating the impact of earlier therapeutic intervention. NGS in larger patient cohorts will allow the identification of distinct mutational patterns as well as gene‐gene interactions to further refine risk categorization in low/int‐1 PMF.