Somatic mutations predict risk of invasive fungal disease in patients with acute myeloid leukemia treated with intensive induction chemotherapy

Abstract Background: Invasive fungal disease (IFD) is a complication of intensive induction chemotherapy (IC) in patients with acute myeloid leukemia (AML) that contributes to treatment-associated morbidity and mortality. Universal antifungal prophylaxis is effective in reducing IFD rates, but over-treats many patients, selects for resistant organisms, and can complicate care due to drug-drug interactions. Somatic AML mutations are reported to alter the functional properties of differentiated immune cells that may modify infection risk. Current pre-treatment risk models do not integrate host or disease-specific factors, including AML genetics, to identify patients at highest risk for IFD that may benefit most from antifungal prophylaxis. We evaluated baseline host and disease characteristics, in a cohort of AML patients receiving IC to predict treatment-related IFD risk. Methods: We identified 538 newly diagnosed AML patients who received IC between 2014-2023 at Dana-Farber Cancer Institute where universal antifungal prophylaxis is not institutional practice. Patients receiving prophylaxis and those with IFD diagnosed before IC were excluded. The primary outcome was the cumulative incidence of proven, probable, and possible IFD per EORTC consensus definitions within 100 days of IC, adjudicated by 2 infectious diseases specialists. The association of baseline host, disease, and treatment characteristics with the primary outcome was evaluated using Fine and Gray competing risk regression, with death and allogeneic transplant (HCT) as competing events. Mutations present in ≥5% of the cohort were included as candidate predictors. All covariates were first assessed in univariable models and those with p0.1 were included in the multivariable analysis. Results: The median age of the patients was 62 years (range: 19-79), 16.7% of patients had prior MDS or MPN, 5.4% had a TP53 mutation, 35.7% had AML-MR mutations, 29.9% had an NPM1 mutation, and 10.6% had a CBF rearrangement. The IC regimens consisted of 7+3 (87%) and CPX-351 (13%). The 100-day mortality was 11% and 16.4% underwent HCT within 100 days of IC. The cumulative incidence of proven/probable/possible IFD by 100 days was 16.2% (n=87). Among 39 cases (7.2%) with proven/probable IFD, 18 had mold (Aspergillus sp. most common, n=9) and 21 had yeast infections (Candida sp. most common, n=16). The lungs were the most common IFD site (63.2%, n=55) followed by disseminated infection (21.8%, n=19). The median time to IFD diagnosis was 21 days from IC. Diagnostic mutations in SRSF2 HR:2.28; 95%CI:1.36-3.82; p=0.002, TP53 HR:2.45; 95%CI: 1.16-5.20; p=0.019, and IDH1 HR:1.92; 95%CI:1.16-3.18; p=0.011, as well as pre-IC absolute neutrophil count (ANC) 1000/mm3 HR:1.64; 95%CI:1.07-2.53; p=0.024 and pre-existing chronic lung disease HR:1.84; 95%CI:1.05-3.23; p=0.034 were independently associated with IFD. Patients with IDH1 mutations had lower pre-IC ANC (500/mm3 vs 1360/mm3; p0.001). Pre-IC ANC was similar in SRSF2 and TP53 mutated vs wild-type cases. There was no difference in IFD incidence between patients receiving 7+3 and CPX-351 and no impact of antecedent myeloid malignancy or prior hypomethylating agent exposure. Post-treatment variables including prolonged severe neutropenia (ANC 500/mm3) and persistent disease have been linked to increased IFD risk. Exploratory analysis revealed that the duration of severe neutropenia was longer (29 vs 26 days; p=0.006) and reinduction rates were higher (27% vs 14.1%; p=0.046) in those with IFD. Reinduction frequency was higher in those with SRSF2 mutations vs. those without. Among single induction patients, neutropenia duration was longer in patients with pre-IC ANC 1000/m3 compared to those with ANC ≥ 1000/m3 (27 vs 22 days; p0.001). These findings suggest that although patients with IFD tend to have prolonged IC-induced neutropenia, baseline factors may be key determinants.Conclusions: In a cohort of patients not receiving antifungal prophylaxis, we show that integrating pre-treatment host characteristics, blood counts, and AML molecular data identifies patients at high risk of IFD following IC. Mutations in TP53, SRSF2, and IDH1 may lead to qualitative or quantitative immune alterations that increase susceptibility to IFD upon IC induced myelosuppression. Our findings indicate that these baseline risk factors identify vulnerable AML subgroups for whom a targeted approach to antifungal prophylaxis may be used.