Session: 904. Outcomes Research—Malignant Conditions: Poster II
Hematology Disease Topics & Pathways:
Diseases, AML, Clinically relevant, Myeloid Malignancies
A 2012 retrospective review of acute myeloid leukemia (AML) patient outcomes in the SEER national database showed higher mortality rates for African Americans and Hispanics despite lower rates of incidence and more favorable genetics at presentation compared to non-Hispanic whites (Patel. Cancer Causes & Control, 2011). While age is a known negative prognostic factor, it has been difficult to analyze the contribution of additional demographic factors predicting outcomes in AML. The purpose of this study was to investigate the impact of race on disease outcome in a diverse patient cohort of intermediate risk AML treated at two urban academic centers in Chicago.
We retrospectively analyzed patients with intermediate risk AML cytogenetics treated with intensive chemotherapy with or without allogeneic stem cell transplant from 2003-2018 at Northwestern Memorial and the University of Illinois Hospital. Baseline clinical and molecular characteristics and treatment response were compared across racial groups. Categorical variables were compared using Pearson’s chi square or the Fisher’s exact test while overall survival (OS) was estimated by the Kaplan-Meier analysis and compared by the log-rank test. Risk factors for OS were assessed by the Cox regression model.
A total of 111 intermediate risk AML patients were included with the 3 major groups being Caucasian (non-Hispanic white), African American (AA) and Hispanic patients (Figure 1). Nearly half of each group was> 60 years of age at presentation. At the time of diagnosis, AA had a significantly higher body mass index (BMI) with a mean of 33.7 kg/m2 compared to 28.1 kg/m2 in the Caucasian group and 27.7 kg/m2 of Hispanic/others (p = 0.0047). There was no statistically significant difference in presenting labs including LDH, WBC count, blast percentage or molecular characteristics including FLT3-ITD or NPM1 mutation status (Figure 2).
In log-rank analysis, Caucasians had significantly improved median OS compared to AA and Hispanics/others (31.6 vs. 16.7 vs.18.1 mos; p=0.026) (Figure 3). Multivariate analysis indicated that non-Caucasian patients had a 1.8 times higher risk of death compared to Caucasians (p = 0.059) even when adjusted for age and NPM1 and FLT3-ITD mutational status.
We sought to examine whether this outcome disparity could be explained by chemotherapy resistance. The predominantly Hispanic group had a decreased response to induction chemotherapy compared to other groups (52% complete response (CR) vs. 70.6% of Caucasians and 68.8% of AA ; p=NS). In multivariate analysis, age (p = 0.046) and NPM1 status (p = 0.0006) predicted response to therapy while race did not. The favorable risk NPM1 mutation was enriched in Caucasians while the unfavorable FLT3 mutation was more commonly found in the primarily Hispanic group of patients (p = NS).
Transplant emerged as a significant predictor of overall survival (p = 0.03). There was a difference approaching statistical significance in transplant utilization between the groups with only 36.8% of AA transplanted compared to 58.1% of non-African Americans (p = 0.09). Underrepresentation of minorities in the transplant registry was reflected in the significantly higher utilization of alternate donor sources in the AA and Hispanic patients. The proportion of patients receiving haploidentical and cord blood transplants was 5.7% of Caucasians vs. 57.1% of AA and 40% of Hispanic/others (p = 0.0006).
In a cytogenetically homogenous cohort of AML patients we show that Caucasian patients had significantly superior overall survival compared to other racial groups. We found a statistically significant difference in BMI between racial groups, which may be reflective of comorbidities and chronic illnesses that affect clinical outcome. Our data also showed there was increased utilization of transplant in Caucasian patients likely related to donor availability and insurance coverage. We will next use Geographic Information Systems and payer source data to dissect the contribution of socioeconomic status and health care access to these disparities in disease outcome. Identifying health service gaps in these populations will enable healthcare providers to provide personalized and equitable care to an increasingly diverse patient population.
Disclosures: Frankfurt: AbbVie: Membership on an entity's Board of Directors or advisory committees; Celgene, Jazz, Agios: Membership on an entity's Board of Directors or advisory committees, Speakers Bureau. Khan: Teva: Speakers Bureau.
See more of: Oral and Poster Abstracts