Project Summary/Abstract Acute myeloid leukemia (AML) is the second most common malignancy among children and adolescents. The prognosis of pediatric AML has improved significantly over recent decades, but still nearly half of patients suffer refractory disease or relapse following a first remission. These patients have a relatively poor prognosis, with a probability of five-year survival following relapse of only ~35%. This is achieved with intensive chemotherapy leading to potentially significant toxicity and even treatment-related mortality. While there has been substantial standardization in practice related to frontline therapy for AML, there is greater variability in treatment of relapsed disease – variability that is poorly understood. Furthermore, given the variability in treatment and the small numbers of patients at any one institution, there is a dearth of information on toxicity risks for the treatments being utilized for refractory or relapsed disease. Cardiotoxicity is a prevalent adverse consequence of AML therapy. However, given the incomplete data captured by frontline clinical trial databases once a patient experiences relapse, data on the onset, progression and potential recovery of cardiotoxicity are generally restricted to the time period during and shortly after frontline therapy and late in long-term survivorship. The current lack of data over the duration of the patient experience precludes a complete assessment of the incremental risks associated with AML therapy, including transplant and salvage. Data on how early cardiotoxicity onset during frontline therapy may impact approach to treatment for relapsed disease, the impact of salvage therapies on cardiac function, and subsequent survival outcomes is thus, also poorly understood. The aims of the proposed work are to establish a unique data resource by combining data from the Children's Oncology Group (COG) clinical trial databases, electronic medical record (EMR) abstraction for a multi-center AML cohort, longitudinal quantitative echocardiographic measures of measures of left ventricular (LV) size, systolic and diastolic function, and administrative data from the Pediatric Health Information System, that will enable assessments comparing toxicity and outcomes associated with the variety of approaches to therapy for refractory/relapsed AML. Specifically, our aims include the following: (1) determine the natural history cardiotoxicity and the incremental effects of treatment from initial diagnosis, through relapse and early post-relapse follow-up, (2) to quantify the influence of prevalent salvage therapies on longitudinal changes in LV size, systolic and diastolic function and the incidence of cardiomyopathy, and (3) to evaluate salvage chemotherapy and transplant as mediators of the documented association between early cardiotoxicity onset during frontline therapy and decreased overall survival.

Project Narrative Data on cardiac function following treatment for relapsed pediatric acute myeloid leukemia (AML) are poorly captured on frontline treatment trials and thus the natural history of cardiotoxicity in pediatric AML is incompletely understood and the impacts of early toxicity on treatment selection upon relapse and subsequent impacts on outcomes have not been evaluated. The proposed project will establish a unique data resource by merging Children's Oncology Group clinical trial data, electronic health record data, longitudinal standard of care echoes, and administrative data from the Pediatric Health Information System that will allow us to address these significant knowledge gaps. Given that nearly 50% of AML patients experience relapse, and relatively poor prognosis of recurrent disease, a better understanding of the toxicity risks associated with salvage therapies and the impact on effectiveness is critical for improving AML outcomes.