Detection of the KITD816V mutation in myelodysplastic and/or myeloproliferative neoplasms and acute myeloid leukemia with myelodysplasia-related changes predicts concurrent systemic mastocytosis

Greater than 90% of cases of systemic mastocytosis (SM) harbor pathogenic KIT mutations, particularly KITD816V. Prognostically-significant pathogenic KIT mutations also occur in 30-40% of core binding factor-associated acute myeloid leukemia (CBF-AML), but are uncommonly associated with concurrent SM. By comparison, the occurrence of SM in other myeloid neoplasms bearing pathogenic KIT mutations, particularly those with a chronic course, is poorly understood. Review of clinical next-generation sequencing (NGS) performed at our institutions in patients with known or suspected hematologic malignancies over an 8-year period revealed 64 patients with both a pathogenic KIT mutation detected at one or more timepoints and available bone marrow biopsy materials. Patients with KITD816V-mutated myelodysplastic syndromes (MDS), myeloproliferative neoplasms (MPN), or overlap MDS/MPN (n = 22) accounted for approximately one-third of our cohort (34%). Comprehensive morphologic and immunophenotypic characterization revealed that nearly all cases (n = 20, 91%) exhibited concurrent SM. In contrast, of the 18 patients (28%) with AML and KITD816V, only eight (44%) showed evidence of SM at any point in their disease course (p = 0.0021); of these eight, the AML component was characterized as AML with myelodysplasia-related changes (AML-MRC) in all but one instance (n = 7, 87%). Twelve patients (19%) had pathogenic KIT mutations other than p.D816V, all in the setting of AML (CFB-AML, n = 7; AML, not otherwise specified, n = 2; AML-MRC, n = 1; acute promyelocytic leukemia, n = 1); only two of these patients (17%), both with CBF-AML, exhibited concurrent SM. The remaining 12 patients (19%) had SM without evidence of an associated hematological neoplasm (AHN). For nearly one-third of the 30 SM-AHN patients in our cohort (n = 9, 30%), the SM component of their disease was not initially clinicopathologically recognized. We propose that identification of the KITD816V mutation in patients diagnosed with MDS, MPN, MDS/MPN, or AML-MRC should trigger reflex testing for SM.