Laura Pardo

Residual disease detected by multidimensional flow cytometry signifies high relapse risk in patients with de novo acute myeloid leukemia: a report from Children's Oncology Group

Early response to induction chemotherapy is a predictor of outcome in acute myeloid leukemia (AML). We determined the prevalence and significance of postinduction residual disease (RD) by multidimensional flow cytometry (MDF) in children treated on Childrens Oncology Group AML protocol AAML03P1. Postinduction marrow specimens at the end of induction (EOI) 1 or 2 or at the end of therapy from 249 patients were prospectively evaluated by MDF for RD, and presence of RD was correlated with disease characteristics and clinical outcome. Of the 188 patients in morphologic complete remission at EOI1, 46 (24%) had MDF-detectable disease. Those with and without RD at the EOI1 had a 3-year relapse risk of 60% and 29%, respectively (P < .001); the corresponding relapse-free survival was 30% and 65% (P < .001). Presence of RD at the EOI2 and end of therapy was similarly predictive of poor outcome. RD was detected in 28% of standard-risk patients in complete remission and was highly associated with poor relapse-free survival (P = .008). In a multivariate analysis, including cytogenetic and molecular risk factors, RD was an independent predictor of relapse (P < .001). MDF identifies patients at risk of relapse and poor outcome and can be incorporated into clinical trials for risk-based therapy allocation. This study was registered at www.clinicaltrials.gov as NCT00070174.

Consistent quantitative gene product expression: #1. Automated identification of regenerating bone marrow cell populations using support vector machines.

Identification and quantification of maturing hematopoietic cell populations in flow cytometry data sets is a complex and sometimes irreproducible step in data analysis. Supervised machine learning algorithms present promise to automatically classify cells into populations, reducing subjective bias in data analysis. We describe the use of support vector machines (SVMs), a supervised algorithm, to reproducibly identify two distinctly different populations of normal hematopoietic cells, mature lymphocytes and uncommitted progenitor cells, in the challenging setting of pediatric bone marrow specimens obtained 1 month after chemotherapy. Four‐color flow cytometry data were collected on a FACS Calibur for 77 randomly selected postchemotherapy pediatric patients enrolled on the Childrens Oncology Group clinical trial AAML1031. These patients demonstrated no evidence of detectable residual disease and were divided into training (n = 27) and testing (n = 50) cohorts. SVMs were trained to identify mature lymphocytes and uncommitted progenitor cells in the training cohort before independent evaluation of prediction efficiency in the testing cohort. Both SVMs demonstrated high predictive performance (lymphocyte SVM: sensitivity >0.99, specificity >0.99; uncommitted progenitor cell SVM: sensitivity = 0.94, specificity >0.99) and closely mirrored manual cell classifications by two expert‐analysts. SVMs present an efficient, automated methodology for identifying normal cell populations even in stressed bone marrows, replicating the performance of an expert while reducing the intrinsic bias of gating procedures between multiple analysts.

A recurrent immunophenotype at diagnosis independently identifies high-risk pediatric acute myeloid leukemia: a report from Children’s Oncology Group

A recurrent immunophenotype at diagnosis independently identifies high-risk pediatric acute myeloid leukemia: a report from Children’s Oncology Group