Socioeconomic Status Is Associated with Prognostic Factors in Childhood Acute Lymphoblastic Leukemia - Implications for Outcomes: A Report from the Children's Oncology Group

Abstract

Introduction: Acute lymphoblastic leukemia (ALL) is the most common form of childhood cancer. Historical studies have found an association between ALL incidence with higher socioeconomic status (SES), though newer findings are contradictory for this association. Nevertheless, it has been well established that inferior outcomes in pediatric ALL are associated with lower SES. We sought to identify if specific ALL disease characteristics at diagnosis are associated with SES to determine if underlying biology plays a role in the known association between inferior outcomes and lower SES. Methods: Demographic and disease characteristics from children and young adults (aged 2-30 years) diagnosed with ALL 2004-2017 and treated on Children's Oncology Group (COG) frontline treatment protocols (N=4,726, AALL17D2) were matched 1:1 (by age, sex, race, and sampling year) and compared to National Health and Nutrition Examination Survey (NHANES) controls. SES for cases was defined using census data to determine the percent of population in a ZIP code living at or below the federal poverty threshold. SES for controls was defined by the individual's poverty income ratio. Chi-square testing and multivariate logistic regressions were performed, adjusting for sex, race/ethnicity, and age, to assess associations between SES and ALL, as well as specific risk-stratifying ALL disease characteristics, including cytogenetics and central nervous disease (CNS) involvement. Results: ALL patients (72% B-ALL, 28% T-ALL) were more likely to be male (62%), 58% were non-Hispanic white, 9% non-Hispanic black and 24% identified as Hispanic. By chi-square analysis, a difference in SES was identified between cases and controls (p<0.0001). This was further characterized by multivariate logistic regression, which found a positive significant association between ALL incidence and high SES, when using low SES as the reference group (OR 2.54, 95% CI 2.23-2.89, p < 0.0001). When disease characteristics were evaluated, there was an association between SES and CNS status, with presence of ALL involvement in the CNS (CNS 2 or 3 status) more strongly associated with low SES (p=0.0009). When stratified by sex and race, an association between low SES and CNS status at diagnosis remained among non-Hispanic white males and Hispanic males (p=0.008 and p=0.01, respectively). There was also an association between high SES and the presence of the favorable cytogenetic characteristic trisomy of chromosomes 4 and 10 (p=0.01), though this was not seen in a specific group when stratified by sex and race. Among non-Hispanic white males, an association was found between low SES and presence of rearrangement of the MLL gene, a poor prognostic cytogenetic characteristic in ALL (p=0.04). Among Hispanic males, an association was found between high SES and the presence of an ETV6-RUNX1 translocation status, a favorable prognostic cytogenetic characteristic in ALL (p=0.05). These associations were not seen when all cases were assessed together. We did not find an association between SES and immunophenotype, hypodiploidy, initial white blood cell count, or BCR-ABL1 fusion status Conclusions: To our knowledge, this is the first study to assess associations between SES and ALL disease characteristics that have known prognostic implications in disease outcome. We found that poor prognostic features of ALL including CNS involvement at diagnosis (regardless of sex and race) and presence of MLL rearrangement specifically in non-Hispanic white males, were both associated with low SES. In contrast, favorable prognostic features of ALL were associated with high SES, such as the presence of trisomy of chromosomes 4 and 10, and among Hispanic males the presence of an ETV6-RUNX1 translocation. Our study also replicated the association between ALL diagnosis and high SES, which has been a controversial finding. Nonetheless, our findings suggest that there may be biological factors that play a role in the known association between inferior ALL outcomes and low SES. Further studies to better elucidate this potential biologic role are needed as there may be important implications for future treatment options that could improve outcomes for children with ALL who have low SES, as well as improve disparate outcomes currently seen in ALL treatment. Disclosures No relevant conflicts of interest to declare.