PF655 PROGNOSTIC RELEVANCE OF THE GENE EXPRESSION SIGNATURE IN PRIMARY AND SECONDARY MYELOFIBROSIS

Abstract

Background: Primary myelofibrosis (PMF) together with polycythemia vera (PV) and essential thrombocythemia (ET) belongs to the classic Philadelphia-negative myeloproliferative neoplasms (MPNs). PV and ET can evolve to secondary myelofibrosis (SMF) giving rise to post-PV (PPV) and post-ET (PET) myelofibrosis (MF). PMF and SMF patients are currently managed in the same way and prediction of survival is based on the same prognostic models, even if it has been demonstrated that they can’t accurately distinguish different risk categories in SMF. In the last few years interest grew concerning the ability of gene expression profiling (GEP) to provide valuable prognostic information for clinical decision making. Several studies demonstrated that GEP can improve risk classification in hematologic malignancies such as AML, MDS and lymphoma. Aims: As a preliminary result we performed GEP in PMF and SMF CD34+ cells and identified a gene signature that was able to distinguish patients with a better prognosis from that with a worse one. To confirm these results and to ease the possible clinical use of a molecular signature based on gene expression we decided to move to a more accessible cell population. Methods: Granulocytes were isolated from 114 MF patients (35 prefibrotic/early PMF, 37 overt PMF, 26 PET and 16 PPV) and total cellular RNA was extracted. Gene expression profiling (GEP) was performed using Affymetrix platform. In order to identify genes whose expression is related to survival in MF patients, we performed a Cox regression analysis by means of Partek GS Software. Results: Cox regression analysis led to the identification of a list of 650 transcripts which discriminate high-risk (HR) MF patients from low-risk (LR) ones. According to our results, the frequency of deceased patients was increased in the HR group and survival curves demonstrated that the median overall survival was lower in the HR group compared to the LR one (3.25 y vs 6.88 y, Log-rank p-value <0.01). Moreover, we observed that 11 out of 13 AML transformed patients clustered within the HR group. Interestingly, we found that HR group was enriched in patients carrying at least one high molecular risk mutation (EZH2, ASXL1, IDH1/2, SRSF2), while there was no significant difference between the two groups regarding the disease and the driver mutation. Concerning clinical variables, we observed that median age at diagnosis was higher in the HR group (65.6 y vs 62.1 y, p-value <0.05). WBC count was increased in the HR group (12.9 x109/L vs 9.3 x109/L, p-value <0.05), while platelet count was decreased in the same group compared to LR ones (218 x109/L vs 352 x109/L, p-value <0.01). Moreover, the percentage of patients with more than 1% circulating blasts is greater in the HR group. Finally, we studied the distribution of samples classified according to DIPSS prognostic model. Strikingly, our gene signature classifies as HR several patients belonging to the DIPSS low and intermediate-1 categories. These patients are deceased or leukemia transformed into a shorter time frame than the median survival reported for the reference prognostic class. Summary/Conclusion: Our results demonstrate that GEP of primary MF cells colud be a useful tool for risk prediction in PMF and SMF since it can improve the identification of patients’ subgroups characterized by a poor prognosis. These results should be validated in independent patient cohorts in order to confirm their predictive power.