Classification of immature white blood cells in acute lymphoblastic leukemia L1 using neural networks particle swarm optimization

Acute Lymphoblastic Leukemia (ALL) is a type of leukemia that is related to a large number of lymphoblast cells in the peripheral blood and bone marrow. The initial step in diagnosing the disease is an individual immature White Blood Cells (WBC) assessment by the hematologists. Visual interpretation and detection of immature WBC is a time-consuming and burdensome task for hematologists. The reliable and confident examination of the ALL blood specimen relies on a valid classification of lymphoblast cells. In this paper, we proposed two-stages Artificial Neural Networks integrated with the Particle Swarm Optimization method to classify the immature WBC in ALL patients. The proposed method includes binary classification of lymphoid cells in the first stages and binary classification of lymphoblast cells in the second stages. In this study, we have used five peripheral blood specimen samples obtained from Sardjito Hospital ALL dataset to develop the proposed model. The proposed model consists of data preprocessing, features selection, features extraction, and two-stage classification. The performance of our approach is compared with the common backpropagation neural networks classification (multiclass NN-BP) and multiclass neural networks particle swarm optimization (multiclass NN-PSO). The results show that the proposed method has better accuracy than other compared models with 86.92% accuracy and should be developed and applied in other cases.