Mutation profiles and outcomes of patients with acute myeloid leukemia with autoimmune disease.

6536 Background: Therapy-related acute myeloid leukemia (t-AML) defines a group of patients who develop AML after being treated with chemotherapy or radiation therapy. Most of these patients have a history of malignancy, but some have exposure to cytotoxic therapy for non-cancerous conditions such as autoimmune diseases (AD). This study looked at the characteristics, genetics, and outcomes of patients with AD and AML. Methods: This study included patients diagnosed with AML from 1999 - 2019. AML was divided into 4 groups: de novo AML, AD- AML (AML after AD diagnosis), t-AML (excluding AD-AML), secondary AML (s-AML) (AML arising out of myelodysplastic syndrome or myeloproliferative neoplasm). All molecular data were obtained via next generation sequencing (NGS) of genes frequently observed to be mutated in AML and analyzed using Chi-square test. BUM method was used to control false-discovery rate (FDR) at 8%. Overall survival (OS) and progression free survival (PFS) were estimated by Kaplan-Meier method and compared using log rank test. Results: A total of 1587 pts were included: 10.1% (n=160) with AD-AML, 71.3% (n=1131) de novo AML, 13.1% sAML (n=208) and 5.6% t-AML (n=88). The median age at diagnosis was 54.5 years; 54% were male; 67.3% received intensive induction therapy (Intensive defined as a 7+3 regimen). Cytogenetic (CG) data were available for 97.3% pts. According to MRC 2010 classification, CG were favorable in 14%, intermediate in 57%, and poor in 28% of the pts. The most prevalent AD were Crohn's/ulcerative colitis (n=30, 17%) and rheumatoid arthritis (n=23, 13%); 10.6% of pts had 2 disorders. In AD-AML group, the median age at diagnosis of first AD was 53 years; the median interval from AD diagnosis to AML was 82.5 months (mo); the majority of pts (56.25%) did not receive treatment (not including steroids) for their AD; 23.8% of pts were treated with immunomodulating agents, 23.1% with cytotoxic agents and 15% with anti-inflammatory agents; 20.1% received ≥ 2 drug classes. Poor risk CG were more common in AD-AML than in de novo AML (30% vs. 12.2%, p<.0001). Mutations in NF1 were more common in AD-AML (9.7%) than in de novo (2.7%), sAML (1.0%), and t-AML (2.7%, p=0.024). FLT3 mutations were less common in AD-AML (9.5%) than in de novo (36.1%) and t-AML (33.3%, p=0.004). Mutations in SRSF2 were more common in AD-AML (20.3%) than in de novo (12.3%) and sAML (8.8%, p=0.119). Median PFS varied among de novo (8 mo), AD-AML (6 mo) and t-AML (5 mo) groups. For the AD-AML grp, the median OS was 9 months; less than the de novo AML (11 mo), and greater than t-AML (6 mo) and sAML (6 mo, p<0.0001). Conclusions: Patients with AML and a prior history of AD have CG profile and outcomes similar to those with t-AML but with distinct molecular features (i.e. NF1 and SRSF2). These findings need to be validated in large series; however, they suggest that the MAPK signaling pathway as well as spliceosome inhibitors might be a future focus of investigation in pts with AD-AML.