Outcome Among Adolescents and Young Adults with Acute Myeloid Leukemia at Pediatric Versus Adult Centers: A Population-Based Study Using the IMPACT Cohort

Background: Survival outcomes among adolescents and young adults (AYA) with acute myeloid leukemia (AML) remain poor. In AYA with acute lymphoblastic leukemia, outcomes differ between patients treated in pediatric vs. adult centers. This has not been well evaluated in AML. We therefore compared outcomes between AYA with AML treated at pediatric vs. adult centers using a population-based clinical database. In addition, we determined other predictors of outcome within this population.

Methods: The IMPACT Cohort comprises all Ontario, Canada AYA aged 15-21 years diagnosed with one of six common cancers (including AML) between 1992-2012. Detailed demographic, disease, treatment, and outcome data were collected through chart abstraction and validated by content experts. Locus of cancer care (LOC - pediatric vs. adult center) was determined based on where the majority of therapy was delivered in the first three months after diagnosis. Linkage to population-based health administrative data identified additional cancer events (second cancers, relapse, death). Event-free (EFS) and overall survival (OS) were determined using Kaplan-Meier methods. The impact of LOC on EFS and OS was determined using multivariable Cox proportional hazard models, adjusting for demographic, disease, and treatment variables. Events included disease progression, relapse, death, and second malignancies.

Results: Among 140 AYA with AML, 89 (63.6%) received therapy at an adult center. AYA treated in pediatric centers were younger than those treated at adult centers (median 16 years vs. 19 years; p<0.001) and were more likely to live in higher-income neighborhoods [37/51 (72.5%) vs. 47/89 (52.8%); p=0.02]. Disease markers such as presenting white blood cell count and AML subtype did not differ by LOC. The 5-year EFS and OS for the whole cohort were 35.0%±4.0% and 53.6%±4.2%. Neither EFS nor OS differed by LOC (Table 1). In multivariable analyses adjusting for disease characteristics, LOC was not predictive of either EFS [adult vs. pediatric center hazard ratio (HR) 1.3, 95^thconfidence interval (CI) 0.8-2.2, p=0.27] or OS (HR 1.0, CI 0.6-1.6, p=0.97). AYA with AML living in rural areas however experienced significantly inferior outcomes as compared to their urban counterparts (EFS: HR 2.5, CI 1.3-4.7, p=0.005; OS: HR 2.0, CI 1.1-3.8, p=0.04).

Conclusions: In this population-based cohort, outcomes did not differ between AYA with AML treated at pediatric vs. adult centers, unlike what has been previously shown in AYA with acute lymphoblastic leukemia. However, rural AYA experienced substantially inferior outcomes than urban AYA, suggesting that even within a universal single payer system of healthcare, socioeconomic disparities persist in this population.