Discrimination of malignant neutrophils of chronic myelogenous leukemia from normal neutrophils by support vector machine

Malignant neutrophils of chronic myelogenous leukemia (CML) have similar antigen expression patterns compared to their normal counterparts, thus making the cells difficult to distinguish by clinical flow cytometry. In this study, we applied the support vector machine method to build a malignant neutrophil prediction model based on nine CML patients and nine healthy donors. This approach effectively differentiated between malignant and normal neutrophils with high specificity and sensitivity (<= 95.80% and <= 95.30%, respectively). This approach may broaden the application of flow cytometry for differentiation between CML and normal neutrophils and become an important diagnostic tool in CML. (C) 2013 Elsevier Ltd. All rights reserved.