Open AccessPS1458 RISK FACTORS AND OUTCOME OF ACUTE MYELOID LEUKEMIA SECONDARY TO POST‐POLYCYTHEMIA VERA AND POST‐ESSENTIAL THROMBOCYTHEMIA MYELOFIBROSIS: AN ANALYSIS OF THE MYSEC COHORT
Author(s) -
Mora B.,
Guglielmelli P.,
Rumi E.,
Maffioli M.,
Barraco D.,
Rambaldi A.,
Caramella M.,
Komrokji R.,
Kiladjian J.J.,
Gotlib J.,
Iurlo A.,
Cervantes F.,
Devos T.,
Palandri F.,
De Stefano V.,
Ruggeri M.,
Silver R.T.,
Benevolo G.,
Albano F.,
Cavalloni C.,
Pietra D.,
Barbui T.,
Rotunno G.,
Bertù L.,
Cazzola M.,
Vannucchi A.M.,
Passamonti F.
Publication year - 2019
Publication title -
hemasphere
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.677
H-Index - 11
ISSN - 2572-9241
DOI - 10.1097/01.hs9.0000564096.75769.8f
Subject(s) - essential thrombocythemia , myelofibrosis , medicine , polycythemia vera , cumulative incidence , myeloid leukemia , cohort , univariate analysis , international prognostic scoring system , incidence (geometry) , oncology , multivariate analysis , bone marrow , myelodysplastic syndromes , physics , optics
Background: Information on the molecular and phenotypic correlates of acute myeloid leukemia secondary (S‐AML) to post‐polycythemia vera (post‐PV) and post‐essential thrombocythemia (post‐ET) myelofibrosis (known as secondary myelofibrosis, SMF) is scant. Aims: The primary objective of this study is to assess the incidence and outcome of S‐AML in a large cohort of SMF. To find out predictors of S‐AML occurrence, we performed a whole investigation of risk factors at the time of SMF, also applying established prognostic models. Methods: The MYSEC ( Myelofibrosis Secondary to PV and ET Collaboration ) project was recently updated collecting 805 SMF cases. We performed time‐to‐event (S‐AML) analysis with Fine & Gray competing risks model using either death or the time of stem cell transplantation as competing risks. To explore whether IPSS ( International Prognostic Scoring System ), DIPSS ( Dynamic IPSS ) and MYSEC‐PM ( MYSEC ‐ Prognostic Model ) predicted S‐AML occurrence, we calculated the difference in Cumulative Incidence Function (CIF) among risk categories for each score and applied Gray's model for testing the homogeneity of CIFs. Results: Within 805 SMF patients, 71 (8.8%) developed S‐AML. Median time from SMF to S‐AML evolution was 1.6 years (range: 0.6 ‐ 14.6). The S‐AML incidence rate was 2.3 x100 person‐year of follow up (CI 95%: 1.8 ‐ 2.9 x100). In univariate analysis, we found a direct association between increased S‐AML risk and platelets <150 x10^9/l, hemoglobin <10 g/dl, PV/ET duration > 14 years, blasts ≥ 3% and non‐ CALR genotype, and an inverse association with the use of JAK2 inhibitors. In a multivariate analysis, only blasts ≥ 3%, PV/ET duration > 14 years and non‐ CALR genotype maintained their association with an increased risk of S‐AML (Table 1) . No impact on S‐AML risk was found as for leukocyte count, spleen size, symptoms, karyotype and age at SMF; neither imbalance was evident for gender, previous PV versus ET, and cytoreduction. Among IPSS, DIPSS and MYSEC‐PM, only the latter score stratified SMF patients for S‐AML risk ( P = 0.004) (Figure 1a) . The estimated hazard ratios (HRs) when changing score category were: 1.55 (CI 95%: 0.69 ‐ 3.50) from low to intermediate‐1, 1.09 (CI 95%: 0.51 ‐ 2.34) from intermediate‐1 to intermediate‐2, and 2.42 (CI 95%: 1.01 ‐ 5.77) from intermediate‐2 to high risk. Comparing to lower risk groups, HR for high risk cases was 4.07 (CI 95%: 1.66 ‐ 10.03). Among S‐AML, 65 (91.5%) patients died and the median survival was 3.9 months (CI 95%: 2.4 ‐ 6.8) (Figure 1b) .Summary/Conclusion: In our dataset of 805 SMF patients, ∼ 9% evolved into S‐AML. The S‐AML incidence rate was 2.3 x100 person‐year and the median survival dismal. In a multivariate analysis, blasts ≥ 3%, non‐ CALR genotype and longer duration of PV/ET correlated with an increased risk of S‐AML. Besides, our data supports the relevance of MYSEC‐PM in predicting S‐AML evolution, especially for the opposite categories of low and high risks.