Absolute T Cell Counts At Diagnosis Predict Absolute Lymphocyte Counts Throughout Therapy For Pediatric Acute Lymphoblastic Leukemia: Evidence For Age-Related Loss Of T Cells and Poor Outcome

Abstract

BackgroundAbsolute lymphocyte count (ALC) during early therapy has been established as an independent prognostic factor in pediatric acute lymphoblastic leukemia (ALL). Importantly, ALC can refine minimal residual disease (MRD)-based risk stratification, particularly for MRD-positive patients. However age-dependent effects on ALC, the lymphocyte subtype (T, NK, B) which underlies this finding, and the role of ALC at relapse are not yet known. MethodsPatients with de novo pediatric B-cell ALL had blood samples drawn biweekly during induction chemotherapy and lymphocyte subsets analyzed at different time points by flow cytometry. In addition, ALC was recorded at time points throughout induction and early consolidation chemotherapy, in maintenance therapy, and at time of relapse. Demographic, NCI risk, and treatment characteristics were also collected. Continuous variables were analyzed with Pearson's coefficient, and categorical variables were analyzed with two-sided Barnard's test. Survival statistics were measured with Kaplan-Meier curves and accelerated failure time analysis. ResultsA total of 21 children and young adults with B-cell ALL were identified with a median age of 17 years (range 0.5-35 years). Although not unexpectedly, T cells comprise the majority of lymphocytes that make up the ALC in nearly all cases. Interestingly, absolute CD3+ T cell count at diagnosis (ATC-0) correlates with ALC-15 (p=0.0008) and ALC during maintenance chemotherapy (p=0.01), revealing that T cell numbers at diagnosis predict total lymphocyte count recovery during early treatment as well as 8+ months later. The averages of ATC-0= 828 cells/ul and ALC-maintenance= 940 cells/ul were significant cutoffs as 89% of patients with low ATC-0 had low ALC during maintenance, and 75% of patients with high ATC-0 at diagnosis had high ALC during maintenance (p=0.03). In addition, age was inversely correlated with both ATC-0 and ALC-15 (p=0.035 and 0.006 respectively), linking that older patients have lower ALCs, which correlate to poor outcome. Indeed, 80% of patients with ALC-15 over 500 were < 10 years of age, while 87% of patients with ALC-15 under 500 cells/ul were >/= 10 years at diagnosis (p=0.007). Interestingly, patients with high ATC-0 had 100% 5 year OS while patients with low ATC-0 had 52% 5 year OS (p=0.1), a trend toward significance on Kaplan-Meier curves due to the low number of ATC-0 high patients (n=4) we had in this high risk patient population. However, ALC-15, ALC in maintenance, and age at diagnosis all independently had a significant correlation with RFS in accelerated failure time analysis (p=0.01, 0.0001, and 0.03 respectively). And importantly, ALC at relapse significantly correlated with OS in accelerated failure time analysis (p=0.0001). DiscussionIn this study, we demonstrate that ALC in ALL patients is inversely correlated with age, reflecting the normal decrease in ALC seen in healthy children, adolescents, and young adults. We also identified T cells as the predominant lymphocyte subtype comprising the ALC, which may reflect an immune reservoir in these patients. Importantly, T cell numbers at diagnosis (ATC-0) predict lymphocyte counts throughout treatment and also inversely correlate with age. This suggests that patients with low T cell numbers at diagnosis continue to have low T cell numbers late into treatment. As ALC and ATC decline with age, this could provide one potential mechanism for why older patients have worse outcomes from ALL. Clinically, ATC at diagnosis may identify patients at risk for relapse and help determine appropriate induction and post-induction therapy. In the future, patients with low ATC/ALC may benefit from immune-modulating therapy, which may overcome a limited immune reservoir. AcknowledgmentsSt. Baldrick's Foundation Fellow AwardKimberly Patterson Leukemia Research Fellowship Disclosures:No relevant conflicts of interest to declare.