ALL-025: Application of Machine Learning in the Diagnosis of ALL

Abstract

Context: The ordinary morphologic diagnosis of ALL by pathologists depends on examining patient peripheral blood together with the bone marrow blood films. However, this manual aspect for diagnosis is susceptible to discrepancies. We have a newly introduced technology that allows for overcoming individual variations in the diagnosis of ALL, so-named machine learning, depending on a complex preprogrammed convolutional network matrix. Objectives: Challenging machine-aided systems, which utilize microscopic blood film images to recognize and diagnose ALL based on a preprogrammed deep convolutional neural network (CNN), i.e., machine learning algorithms. Materials and Methods: We collected a dataset of images composed of PB & BM smear images of two classes; ALL and normal control blood. We analyzed 192 samples of digital images: 96 images of patients with ALL and 96 images of healthy normal controls. For each smear sample, we collected the results of clinical data (clinical history and examination) and laboratory data (morphological, cytochemical, and immunophenotyping assessment). We challenged seven types of CNN models: Alexnet, VGG16, VGG19, GoogLeNet, ResNet50, ResNet101, and Inception-v3 to diagnose ALL. Results: Comparing the ability of 7 models to diagnose ALL revealed that the Alexnet had the lowest accuracy of 95.51%, followed by VGG16 at 92.13% and VGG19 at 93.83%, whilst the Inception-v3 had a promising highest accuracy of 99.98%.and was able to detect almost all ALL cases. Conclusion: The statistical measures of the Inception-v3 performance revealed promising results. The sensitivity, specificity, and accuracy of Inception-v3 reached 99.98% for detection of ALL.