<i>IDH M</i>utated AML Is Associated with Increased Risk of Pulmonary Morbidity and Mortality during Induction Chemotherapy

Introduction: IDH1/2 mutations are common in Acute Myeloid Leukemia (AML) and shape prognostic and therapeutic decision making. Induction chemotherapy with intensive or reduced intensity regimens remains the standard of care for IDH1/2mut AML but is associated with high rates of morbidity and mortality with upwards of 15% of patients suffering acute hypoxic respiratory failure. Appropriate patient selection and management of adverse events during induction chemotherapy is of utmost importance. Thus, we conducted a retrospective analysis of patients with newly diagnosed AML who were found to have IDH1/2 mutations to assess whether there was increased incidence of pulmonary morbidity and mortality during induction. Furthermore, we reviewed causes of respiratory failure in this population in order to suggest possible changes in preferred induction method. Methods: Electronic health records for 178 patients with newly diagnosed AML undergoing induction chemotherapy at the Hospital of the University of Pennsylvania from 2017 to 2022 were reviewed. Demographic data, comorbidities, clinical presentation, laboratory values, imaging findings, and molecular genetics were collected for IDH wild type (N=125) and IDH1/2 mutated (N=53) patients. Data was collected from time of diagnosis until completion of induction or death. Acute Hypoxic Respiratory Failure (AHRF) was defined as requiring 4L nasal cannula or SpO2 &amp;lt; 90%. Standard consensus definitions for fungal, bacterial, and viral infections were used. Results: In the IDHmut cohort, 44% of patients developed acute hypoxic respiratory failure compared to 30% IDHwt (p=0.075). Both clinical pneumonia (33% vs 49%, P=0.031) and possible fungal pneumonia (11% vs 25%, P=0.02) were more common in the IDHmut group. Further evaluation of IDH mutated patients revealed an increased likelihood of abnormal first chest X ray upon admission when excluding bilateral pleural effusions (13.0% vs 23.6%, P=0.03). Notably, IDHmut patients were significantly more likely to die from pulmonary causes during induction chemotherapy than their wild type counterparts (6.1% vs 15.3%, P=0.02). Subgroup analysis of IDHmut patients with AHRF revealed only 11.7% to have an infectious source identified. Of patients with possible invasive fungal infection (IFI), only 5.8% (1/17) met criteria for probable IFI and none had confirmed IFI. Conclusion: In total, IDHmut AML patients are more likely to develop AHRF during induction chemotherapy and die from complications of this than IDHwt patients. Pulmonary complications are often infectious in nature during induction chemotherapy, but this population lacked significant serologic evidence of an infectious source. Particularly, there was an increase in patients with peri-lymphatic nodular thickening on imaging consistent with fungal pneumonia but without serologic evidence of fungal infections. In considering other possible causes, pulmonary leukemic infiltration (PLI) could explain the increased incidence of AHRF with bilateral multifocal opacities and peri-lymphatic thickening. This raises questions regarding the preferred induction strategy in these patient's as well as the utility of PLI directed therapy. Further studies aimed at characterizing this population are required to determine optimal management strategy.