Abstract
Abstract 1592
Poster Board I-618
Background
Previous studies highlighted the importance of the cell adhesion molecule, VLA-4, for chemoresistance and minimal residual disease (MRD) in AML, suggesting promise as therapeutic target. By comparison, the prognostic role of VLA-4 in AML remains controversial with retrospective studies implying either adverse or favorable prognosis. Therefore, we prospectively evaluated VLA-4 expression in participants of a recent Children's Oncology Group (COG) AML pilot protocol.
Methods
COG-AAML03P1 enrolled 340 newly diagnosed children (aged 1 month - 21 years) with de novo non-acute promyelocytic AML, excluding those with Down syndrome, and tested the feasibility of combining gemtuzumab ozogamicin (GO) with intensive induction chemotherapy followed by GO-containing intensification therapy or matched related donor stem cell transplantation; 216 patients submitted diagnostic marrow specimens for flow cytometric determination of VLA-4 expression that was then correlated with patient demographics, laboratory characteristics, and clinical outcome. Cytogenetics and molecular prognostic markers were used for risk classification as follows: low risk (mutation in core-binding factor, NPM1, or CEBPa; n=73), high risk (-5/5q-, monosomy 7, or FLT3/ITD with high allelic ratio; n=25), or standard risk (all other patients with cytogenetic/molecular data; n=101); 17 patients had insufficient data for risk classification.
Results
Among the 216 diagnostic specimens, the mean fluorescence intensity (MFI) of VLA-4 expression varied over 35-fold from a baseline of 30 to 1110 (median, 219.5). Patients with high VLA-expression (>median MFI) were younger (p<0.001), had a lower prevalence of FLT3/ITD (p=0.002). Presence of extramedullary disease (EMD, chloroma or CNS involvement) was significantly higher in patients with high VLA-4 expression (16% vs. 5%, p=0.013), where 17/22 (77%) of patients with EMD had high VLA-4 expression. We initially inquired whether VLA-4 expression as a continuous variable correlated with disease outcome. We demonstrated that over the range of VLA-4 expression levels, every 10-unit increase in VLA-4 MFI corresponded to a 2% decrease in relapse risk (RR; p=0.023) and 2% increase in disease-free survival (DFS) from end of induction I (p=0.015). We subsequently divided the study patients into two cohorts based on median MFI and determined the clinical outcome per median VLA-4 MFI for the entire cohort as well as in specific risk groups (i.e., high risk, low risk, and standard risk). In the evaluation of the entire cohort, patients with high VLA-4 expression had a significantly superior DFS (67% vs 48%, p=0.023) and lower RR (24% vs 44%, p=0.011) compared to those with lower VLA-4 expression. In subgroup analyses, high VLA-4 expression was associated with significantly superior DFS (69% vs 34%, p=0.011) and lower RR (26% vs 61%, p=0.009) in patients with standard-risk AML but not in patients with high or low risk disease. In multivariate Cox regression analyses that included age, cytogenetics, and molecular prognostic factors, low VLA-4 expression remained significantly associated with elevated RR (hazard ratio for low VLA-4: 2.25, p=0.011). Finally, we determined the impact of MRD in the context of VLA-4 expression. The prevalence of MRD was similar for patients with low or high VLA-4 expression (29% vs. 25%, p=0.70). In patients with low VLA-4 expression, those with MRD had a RR of 75% compared to that of 35% for the MRD negative cohort (p=0.005). Corresponding DFS was 19% and 54% for those with and without MRD (p=0.018), a difference mainly attributable to increased RR. In patients with high VLA-4 expression, those with MRD had a RR of 31% vs. 23% for the MRD-negative patients (p=0.28) with a corresponding DFS of 46% and 75% for the MRD positive and negative patients (p=0.014), a difference mainly explained by higher treatment-related mortality in MRD positive patients.
Conclusion
This study demonstrates significant heterogeneity of VLA-4 expression in pediatric AML and its association with EMD and clinical outcome. This study further demonstrates that VLA-4 expression is an independent prognostic factor for clinical outcome and can identify the risk status in patients with no identifiable cytogenetic or molecular risk factors.
Disclosures
No relevant conflicts of interest to declare.
Faculty Opinions recommendation of Refinement of risk stratification for childhood rhabdomyosarcoma using FOXO1 fusion status in addition to established clinical outcome predictors: A report from the Children's Oncology Group.
