Development of Machine Learning Models for Screening of Anemia and Leukemia Using Features of Complete Blood Count Reports

Abstract

Complete Blood Count (CBC) report features are routinely used to screen a wide array of hematological disorders. The complexity of disease overlap increases the probability of neglecting the underlying patterns between the features. Additionally, the expertise of healthcare professionals and heterogeneity associated with the subjective assessment of a CBC report often lead to random clinical testing. Such disease prediction analyses can be enhanced by the incorporation of Machine Learning (ML) algorithms for efficient handling of CBC features. This research presents ML-based models for the screening of two common blood disorders – anemia and leukemia, using CBC report features. A ‘fingerprint’ of 14 out of 21 features based on both statistical and clinical relevance is selected. Hybrid synthetic data are generated based on the statistical distribution of the features to overcome the constraint of small dataset size. As inferred from existing knowledge, this study is the first one to employ hybrid synthetic data for modeling hematological parameters. In this study, six ML models i.e., decision tree, random forest, support vector machine, logistic regression, gradient boosting machine, and multilayer perceptron are used. Exceptional performance has been observed by the random forest algorithm with 98% accuracy and 97%, 98%, 99%, and 2% macroaverages of precision, recall, specificity, and miss-rate respectively for the target variable. Hence, this algorithm based on CBC features appears to be an efficient support system for the screening of anemia and leukemia, which has the potential to be deployed in clinical settings for early intervention of these disorders.