Occurrence and Resolution of Anthracycline Cardiotoxicity and Impact on Treatment Outcomes Among Children Treated on the AAML1031 Clinical Trial: A Report from the Children's Oncology Group

Background: Children treated for acute myeloid leukemia (AML) receive high doses of anthracyclines. Anthracycline cardiotoxicity causes significant short and long term morbidity and mortality, thus it is crucial to understand its risk factors, natural history, and impact on treatment outcome. Methods: Between June 2011 and September 2018, AAML1031 enrolled patients aged <30 years for treatment of de novo AML. Echocardiographic evaluations (echo) were required before each treatment course, at the end of protocol therapy, and at yearly intervals during off-protocol follow-up. The lowest shortening fraction (SF) and ejection fraction (EF) in each reporting period were prospectively collected. Cardiotoxicity was defined as SF <28% or EF <55% which represent the protocol-mandated thresholds at which anthracyclines were to be withheld. Higher-grade toxicity (SF <24%, EF <50%) consistent with grade 2 or higher left ventricular systolic dysfunction (grade 2+ LVSD) per CTCAE v3 definitions, and its resolution were also assessed. Resolution was defined as SF returning to baseline or >30% and EF returning to baseline or >60%. Resource utilization (RU) data for patients enrolled at hospitals contributing to the Pediatric Health Information Systems database (PHIS) were merged with AAML1031 study data. Cox regressions were used to assess risk factors for incident cardiotoxicity and recovery among cases, as well as to compare event-free (EFS) and overall survival (OS) in patients with and without cardiotoxicity. Poisson regressions were used to compare on-protocol RU. Cardiotoxicity was treated as a time-varying exposure introduced at time of documented onset in comparisons of EFS, OS and RU. Patients with high allelic ratio FLT3/ITD+ AML were excluded from analyses. Results: 1092 patients were treated on AAML1031 and included in the primary analyses. Echo compliance was high overall (mean ±sd: 74.5% ±19.6), but lower and more variable during off-protocol (mean ±sd: 58.1%, ±31.0) than on-protocol (mean ±sd: 90.6% ±19.5) reporting periods. Over a median follow-up of 3.5 years, 39% (n=428) of patients experienced protocol-defined cardiotoxicity, of whom 54% (n=232) suffered grade 2+ LVSD. Cumulative incidence was higher among patients randomized to receive bortezomib, non-infants, females, and patients with low risk disease who receive more cumulative anthracycline during frontline therapy (Table 1). Patients treated with dexrazoxane were less likely to develop cardiotoxicity. Among patients with grade 2+ LVSD, 49% experienced recovery of cardiac function within a median of 14 months (IQR: 9-20) from initial detection. High risk disease classification and dexrazoxane exposure were associated with LVSD recovery (Table 1). Both EFS (HR 1.23, 95% CI: 1.01-1.50) and OS (HR 1.45, 95% CI 1.15-1.84) were significantly worse in patients with cardiotoxicity meeting per-protocol thresholds for treatment modification. The magnitude of the association with EFS (HR 1.24, 95% CI: 0.98, 1.58) was similar when restricting to grade 2+ LVSD, but more pronounced for OS (HR 1.80, 95% CI 1.38, 2.34). A subset of 389 patients (23% suffered grade 2+ LVSD) were treated at PHIS hospitals and contributed to RU analyses. Cardiotoxicity was associated with greater cardiac ICU requirements, increased use of cardiac-directed interventions including antihypertensives, vasopressors, and heart failure medications, and reduced cardioprotection (Table 2). Conclusions: Echo compliance and cardiotoxicity rates were higher than previously described due in part to more frequent echoes and required reporting of raw echo results. Approximately 20% of pediatric AML patients on AAML1031 developed grade 2+ LVSD, only half of whom had documented resolution of LVSD. Dexrazoxane was strongly associated with reduced LVSD risk and potentially superior recovery. Our results offer further evidence that early cardiotoxicity decreases EFS/OS and that cumulative anthracycline dose is the principal predictor of LVSD. These data argue for comparative effectiveness studies of cardiac-directed interventions after LVSD, and identification of clinical and genetic risk factors for anthracycline cardiotoxicity and its recovery. Such studies may allow for anthracycline de-escalation in select patient populations. Work is ongoing to inform optimal supportive care practices for patients with LVSD using clinical data merged with PHIS. Disclosures Leger: Abbott: Research Funding; Jazz Pharmaceuticals: Membership on an entity's Board of Directors or advisory committees; BTG Pharmaceuticals: Membership on an entity's Board of Directors or advisory committees.

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.