Pilot Study of Ethnic Variation in Biological Features of Childhood Acute Lymphoblastic Leukemia (ALL) in Utah.

Acute lymphoblastic leukemia (ALL) is the most common childhood malignancy, and the handful of studies that investigate ethnic differences in childhood ALL consistently report an increased incidence and worse survival in African-American and Latino children. Studies examining clinical features (high white count) and biological prognostic factors (ploidy, immunophenotype, chromosomal abnormalities) by ethnicity could find no contribution that explained poor outcome among African-Americans. However, investigations exclusively focused on biological factors contributing to worse outcome in Latino children with ALL have been limited. The Latino population in Utah is approximately 10% and Latino children account for 15-20% of new ALL diagnoses treated at Primary Children's Medical Center (PCMC), the only children's hospital located in Utah. As the Latino population continues to expand in Utah and the surrounding states, it will become increasingly important to understand what biological differences, if any, exist between Latino and non-Latino children with ALL.

To investigate whether Latino children with ALL in Utah have different clinical and biological features compared to non-Latino children.

For this pilot study, we conducted a retrospective chart review of clinical and biological risk factors in 90 pediatric patients (Latino, n=40 and Caucasian, n=50) diagnosed with ALL at PCMC between 1997-2009. We included all Latino children treated during this time and randomly-selected Caucasian children for ∼ 1:1 comparison. Over 50 clinical and biological variables were recorded for each patient and compared with Fisher's Exact Test (2-tailed) and logistic regression models.

We found no statistical differences between the Latino and Caucasian cohort in our univariate analysis for the following features: age ≥ 10 years old (Latino 27.5% vs. Caucasian 26%), initial WBC > 50 × 10^-3/uL (Latino 22.5% vs. Caucasian 14%), NCI-Rome High Risk Group (Latino 37.5% vs. Caucasian 46%), initial hemoglobin < 11.5 gm/dL (Latino 85% vs. Caucasian 80%), initial platelet count < 150 × 10^-3/uL (Latino 87.5% vs. Caucasian 80%), precursor B-cell phenotype (Latino 98% vs. Caucasian 86%), slow early response (SER) marrow defined as >5% blasts (M2 or M3) for day 8 (Latino 66.6% vs. Caucasian 57.4%), day 15 (Latino 15.4% vs. Caucasian 15.8%), and day 29 (Latino 2.8% vs. Caucasian 0%), minimal residual disease (MRD) ≥ 0.1% on Day 29 (Latino, n=4 vs. Caucasian, n=7), FISH cytogenetic results (Chr. 4 hyperdiploidy: Latino, n=5 vs. Caucasian, n=9; Chr. 10 hyperdiploidy: Latino, n=2 vs. Caucasian, n=10; Chr. 17 hyperdiploidy: Latino, n=3 vs. Caucasian, n=9; Chr. 21 hyperdiploidy: Latino, n=7 vs. Caucasian, n=16), t(12;21) status (Latino, n=1 vs. Caucasian, n=7), t(9;22) status (Latino, n=2 vs. Caucasian, n=0), hepatomegaly (Latino 10% vs. Caucasian 4%), splenomegaly (Latino 18% vs. Caucasian 16%), combined hepatosplenomegaly (Latino 18% vs. Caucasian 36%), CNS disease defined as ≥ CNS 2 (Latino 12.5% vs. Caucasian 12%), DNA index <1 (Latino 10.8% vs. Caucasian 8.3%), DNA index >1 (Latino 27% vs. Caucasian 27%), patients enrolled in COG clinical trial (Latino 75% vs. Caucasian 75%), and relapse rate (Latino 15% vs. Caucasian 8%). Logistic regression found no differences between the Latino and Caucasian cohort for: age (Latino median = 5 yrs vs. Caucasian median = 4 yrs), platelet count at diagnosis (Latino median = 51 vs. Caucasian median = 56), hemoglobin at diagnosis (Latino median = 8.2 vs. Caucasian median = 8.0), WBC at diagnosis (Latino median = 10.3 vs. Caucasian median = 7.0), and initial absolute blast count (% x WBC count) (Latino median = 3024 vs. Caucasian median = 1680).

We have performed the first comprehensive evaluation of ethnic variation in clinical and biological features in pediatric ALL in Utah. Surprisingly, we found no significant differences between the Latino and Caucasian patients in the features that we examined (although several variables did show a trend). This lack of difference could be explained by: 1) lack of adequate power, 2) lack of any true difference (our Latino vs. Caucasian relapse rate in Utah was similar), or 3) ethnic variation in molecular biology beyond the level of detection for collected variables. We are now preparing to expand our patient numbers and to interrogate clinical samples from each patient using high-resolution, genome-wide technology as we continue our pilot study.

No relevant conflicts of interest to declare.