ASXL1 Mutations in AML: Molecular Biomarker for Secondary AML?

Abstract

Background: Additional sex combs like transcription factor 1 (ASXL1) is a member of the polycomb group protein. ASXL1 mutation has been implicated in myeloid malignancy transformation. It is hypothesized that mutated ASXL1 leads to the loss of polycomb repressive complex 2 (PRC2) mediated gene repression and subsequent transforming events. Recent studies identify ASXL1 mutation as a poor prognostic marker in patients (pts) with de novo acute myeloid leukemia (AML) who present with intermediate–risk cytogenetic lesions (Patel, NEJM 2012; Schnittger, Leukemia 2013). To study the impact of ASXL1 mutations in an unselected AML population, we analyzed clinical and molecular characteristics of patients with untreated AML who express ASXL1 mutation at presentation. Methods: Using next generation sequencing, 254 adult patients with AML seen at the Hospital of the University of Pennsylvania were analyzed for mutations, including ASXL1, using a 33-gene hematologic malignancy panel. Clinical characteristics were obtained from retrospective chart review. Kaplan-Meier estimates were used to calculate overall survival (OS) from time of diagnosis. Living patients were censored at date last seen. Results: ASXL1 mutations were detected in 36/254 (14%) AML pts. There were 29 known pathologic mutations, 1 benign, 1 probable pathologic, and 9 variants of unknown clinical significance (VUS). In 6/36 (16.7%) pts, ASXL1 was the sole mutation identified. Of the 30 pts with additional mutations ([Figure 1][1]), 6/30 (20%) pts harbored 2 independent ASXL1 mutations. When the 27 patients with pathologic ASCL mutations were analyzed for co-mutations, TET2 (13/27, 48%) was the most frequent ASXL1 co-mutation. FLT3 (0/27, 0%) and NPM1 (1/27, 3.7%) were notable for their absence. Median age of pts at diagnosis was 69 years (range 23-80). Prior myelodysplastic syndrome (MDS) or myeloproliferative neoplasm (MPN) was noted in 9/36 (25%) and 11/36 (30.6%) pts, respectively. Four pts (11.1%) had received chemotherapy and/or radiation therapy for a prior non-myeloid neoplasm. Karyotype was normal in 18/36 (50%) pts, and 7 additional pts had intermediate cytogenetic lesions. There were 7 pts (19.4%) with unfavorable cytogenetics (complex karyotype (3 pts), 7q- (3 pts), and 5q- (1 pt)). Four pts (11.1%) had a favorable karyotype, with t(8;21) in 3 pts and t(15;17) in 1 pt. At presentation, median white blood cell count (WBC) was 6.4x103/uL (1.0 x -103). In pts whose AML transformed from prior MPN, median WBC was 50 X103/uL (3.3-140). Standard induction chemotherapy with an anthracycline and cytarabine was given to 17/36 (47%) pts. An additional 3/36 (8.3%) pts underwent induction therapy with clofarabine. Complete remission (CR) was documented in 14/20 (70%) evaluable pts. Of the remaining pts, 11 received a hypomethylating agent, and 5 received other therapies. Thirty-day treatment mortality for all 36 pts and for 27 pts with known ASXL1 pathologic mutation was 13.4% and 18.5% respectively. Kaplan-Meier estimate showed a median overall survival of 349 days (median follow up of 107 days (range 15-1570)). For the 27 pts with a pathologic ASXL1 mutation, the OS was 276 days ([Figure 2][2], median follow up of 145 days (range 18-1570)). Conclusion: ASXL1 mutations in de novo AML with intermediate-risk cytogenetics is associated with poor clinical outcome in cooperative group trials. Strikingly we demonstrate in a single institution, retrospective analysis that 66.7% of pts who present with ASXL1 mutations in the setting of previously untreated AML had documented MDS, MPN and/or prior chemotherapy/radiation. Further studies are necessary to evaluate if ASXL1 mutation has independent prognostic significance in AML or if it is primarily a marker for secondary leukemia. ![Figure 1:][3] Figure 1: ASXL1 and co-mutations ![Figure 2:][3] Figure 2: Overall survival for AML patients with ASXL1 pathologic mutation Disclosures No relevant conflicts of interest to declare. [1]: #F1 [2]: #F2 [3]: pending:yes