Patient-Associated Survival Modulators in Adolescent and Young Adult (AYA) Patients (Pts) with Acute Myeloid Leukemia (AML)

Background: AML is a highly aggressive hematologic malignancy, but the AYA pts [aged 18-39 years (y)] have improved survival, with ~50% of pts achieving long-term cure. Pt outcomes are known to be affected by disease-related factors including cytogenetic findings and gene mutations, which were established in large genomic studies. In addition, pt-related factors, such as self-reported race and socioeconomic contributors have been found to affect survival in population-based studies. Whereas population-based studies are performed on large pt cohorts to establish meaningful conclusions, they usually lack details with respect to additional pt-related factors, including complete comorbidities, insurance status as well as information on disease presentation, such as time from the first provider contact to confirmation of AML diagnosis (time to diagnosis) and treatment, which might also influence pt survival.

Rationale and Methods: Thus, we set out to establish recent survival rates of AML AYA pts at the population level using pts reported to one of 12 Surveillance Epidemiology End Results (SEER) Program registries of the National Cancer Institute [n=1119; median age, 30 y,1992-2014; n=239; median age, 29 y, 2015-2020], and analyze possible factors influencing survival. In parallel, we performed an in-depth analysis of treatment response and survival, including time to diagnosis, and additional pt-associated variables pertinent to AYA pts (employment status, housing, insurance, comorbidities, distance to treatment center, research study enrollment, biobanking) based on retrospectively collected data of AYA AML pts cared for at a major US Comprehensive Cancer Center (CCC) in 2015-2020 (n=96, median age, 31 y), for whom detailed clinical information was available.

Results: Generally, overall survival (OS) of AYA AML pts significantly improved over time, with a marked improvement of 3-y OS rates from 47% in pts diagnosed prior to 2009, to 63% in pts diagnosed in 2010-2014 and 69% in pts diagnosed in 2015-2020. Analysis of possible associations between demographic parameters and risk of death among AYA AML pts reported to SEER Program registries did not show any statistically significant impact of sex, age, county-level median household income or metro/rural status. The only factor affecting OS of AYA AML pts was race. Specifically, Black AML pts had a higher risk of death compared with that of White AML pts [HR 2.13 (95% CI 1.03 - 4.39)], with 3-y OS rates of 47% and 80% ( P=.04), respectively. Survival disparities did not improve over time. CCC AYA AML pts had similar OS rates to those seen in the population-based analyses, with a 3-y OS rate of 70%. Ninety-three percent of pts achieved a complete remission, and 31% of pts relapsed. Analyses of CCC AML pts allowed for preliminary assessment of additional parameters. Ninety-one percent of pts (77/85) received a diagnosis in ≤2 weeks after first provider contact. Although pt numbers were small, there was a trend for a longer diagnosis time in Black pts, 27% (3/11) of whom were diagnosed >2 weeks after first presentation, compared with only 8% (5/61; P=.10) of White pts. Interestingly, analysis of possible additional outcome modulators in CCC AYA pts suggested that presence of other medical conditions at diagnosis associated with lower relapse rates (16% vs 34%, P=.04), and unemployment tended to associate with higher relapse rates (34% vs 17%, P=.06). In addition, of pts approached for biobanking, fewer Black pts consented to sample collection than White pts (60% vs 87%, P=.06).

Conclusion: Survival of AYA pts with AML significantly improved over time, likely driven by availability of allogeneic stem cell transplantation and improved supportive care. Black race still represents the most important negative survival prognosticator, without marked signs of improvement. An exploratory analyses of additional pt-associated variables suggested less recognized factors that may impact survival, such as trends for longer time to diagnosis and reduced participation in biobanking protocols in Black pts compared with White pts, and a higher relapse rate in pts who are unemployed. Although these differences were not yet statistically significant , we are currently in the process of increasing the sample size to further evaluate these potentially important trends.

Support: American Cancer Society, Leukemia Lymphoma Society