High Early Death Rates, Treatment Resistance and Short Survival of Black Adolescent and Young Adults (AYAs) with Acute Myeloid Leukemia (AML) (Alliance)

Background: AML is a highly aggressive hematologic malignancy. Patient (pt) outcomes are affected by disease-related factors including cytogenetic findings and gene mutations, as well as pt-related factors, such as age and race. Younger pts have superior survival: ~50% of pts diagnosed as AYAs (18-39 years) may be cured of their disease. However, the impact of race on the outcome and associated disease profiles in this pt population are unknown.

Methods: We compared survival and molecular profiles of 655 Non-Hispanic Black and Non-Hispanic White (hereafter referred to as Black, n=89 and White, n=566) AYA AML pts treated on frontline Cancer and Leukemia Group B/Alliance for Clinical Trials in Oncology protocols based on standard intensity cytarabine/anthracycline induction therapy between 1986 and 2016. Three hundred ten pts were analyzed molecularly via targeted sequencing of 81 genes. Additionally, we performed integrated genomic profiling (whole-exome sequencing and transcriptome sequencing) and measured residual disease (MRD) in serial samples of 4 Black pts who relapsed with their disease.

Results: A comparison of clinical characteristics of AYA AML pts by race revealed almost identical age and sex distribution, and we found no significant differences between clinical features at diagnosis. With regard to genetic profiles, 42% of White pts were cytogenetically normal, whereas only 18% of Black pts had cytogenetically normal AML (CN-AML; p<0.001). The abnormal karyotypes in Black pts more often contained abnormalities associated with core-binding factor (CBF) AML (39% v 25%, p=0.01; Fig. 1A). White pts had more known pathogenic NPM1 variants (29% v 9%, p=0.01), whereas Black pts had a higher incidence of ZRSR2 pathogenic variants (9% v 0.4%, p=0.004) and tended to have pathogenic KRAS variants more often (12% v 5%, p=0.11; Fig. 1B).

Black AYA AML pts had worse outcomes including a higher early death rate (ED, defined as death within 30 days of diagnosis; 11% v 2%, p<0.001), a trend towards lower complete remission (CR) rate (73% v 82%, p=0.06) and a shorter overall survival (OS; median, 1.5 v 3.1 years [y], p=0.002). Notably, this survival disparity was almost exclusively driven by pts aged 18-29 y: Black pts had a higher ED rate (16% v 3%, p=0.002), a lower CR rate (66% v 83%, p=0.01) and shorter OS (median, 1.3 v 10.2 y, p<0.001) but not disease-free survival (DFS; p=0.16) than White pts aged 18-29 y. In contrast, there were no significant differences in these outcome metrics between Black and White pts aged 30-39 y (Fig. 2). Among all pts consolidated with intensive chemotherapy (n=566), multivariable analysis revealed Black race as an independent prognosticator of shorter DFS (p=0.04) and OS (p<0.001). These differences in OS were also significant when we included pts who received allogeneic transplantation in 1st CR (n=655; p<0.001). 18-29 y old Black pts with any non-CBF AML had very poor OS compared to White pts (5-y rates, 12% v 45%, p<0.001). CBF-AML pts aged 18-29 y tended to have an inferior OS compared with White pts (5-y rates, 41% v 44%, p=0.10).

To gain insights into the genetic features of Black AYA AML pts at different stages of the disease, we performed integrated genomic profiling on paired leukemic samples from diagnosis and relapse of 4 Black AYA pts. In all pts, the original dominant leukemic clone persisted and was dominant at relapse (Fig. 3). This suggests that the leukemic clone persists during treatment with conventional cytotoxic chemotherapy. This observation was further supported by MRD detection of NPM1 mutations in NPM1-mutated pts at time of morphologic CR.

Conclusion: Black AYA AML pts present with distinct molecular features, including very high frequencies of CBF AML, and low frequency of NPM1. Pts aged 18-29y account for the race-associated survival disparity, especially non-CBF pts who have dramatically poor survival. On the one hand, the lower CR rates combined with persistence of dominant clones at relapse suggest reduced response to induction chemotherapy, and suggests the need for different treatment intensities and/or modalities in this pt cohort. On the other hand, high early death rates are indicative of delay in diagnosis and care, including health inequities, calling for systematic changes particularly for this population.

Figure 1

View largeDownload slide

Figure 1

View largeDownload slide

Close modal