Identification and Staging of B-Cell Acute Lymphoblastic Leukemia Using Quantitative Phase Imaging and Machine Learning

Abstract

Identification and classification of leukemia cells in a rapid and label-free fashion is clinically challenging and thus presents a prime arena for implementing new diagnostic tools. Quantitative phase imaging, which maps optical path length delays introduced by the specimen, has been demonstrated to discern cellular phenotypes based on differential morphological attributes. Rapid acquisition capability and the availability of label-free images with high information content have enabled researchers to use machine learning (ML) to reveal latent features. We developed a set of ML classifiers, including convolutional neural networks, to discern healthy B cells from lymphoblasts and classify stages of B cell acute lymphoblastic leukemia. Here, we show that the average dry mass and volume of normal B cells are lower than those of cancerous cells and that these morphologic parameters increase further alongside disease progression. We find that the relaxed training requirements of a ML approach are conducive to the classification of cell type, with minimal space, training time, and memory requirements. Our findings pave the way for a larger study on clinical samples of acute lymphoblastic leukemia, with the overarching goal of its broader use in hematopathology, where the prospect of objective diagnoses with minimal sample preparation remains highly desirable.

Overall Survival Prediction for Gliomas Using a Novel Compound Approach

“…CNN extracting features is usually difficult to explain the biological principles behind it ( 27 ). In order to have a deeper understanding of the process of model learning features, we generate activation maps for each activation layer (ReLU) in the network ( 28 ), as shown in Figure 14 .…”

“…Among them, "CNN (RF)" represents the use of deep features and random forest models to predict the survival period, and "CNN (DL)" represents the use of deep features and neural networks to predict the survival period. CNN extracting features is usually difficult to explain the biological principles behind it (27). In order to have a deeper understanding of the process of model learning features, we generate activation maps for each activation layer (ReLU) in the network (28), as shown in Figure 14.…”

Deepometry, a framework for applying supervised and weakly supervised deep learning to imaging cytometry

“…Label-free cytometry using machine and deep learning on images generated from a range of different measurement techniques has now been used for blood cell classification 32,33 , rare cell 34 and cancer detection 35,36 . Recently quantitative phase microscopy has been used to measure the progression of cell states including the activation of T cells 37 and progression of B-cell ALL 38 and this protocol could be used on these single-cell images to quantify and visualise progressive morphological changes.…”

Reagent-free Raman and quantitative phase imaging offer a unique morpho-molecular platform for recognition of malignancy and stages of B-cell acute lymphoblastic leukemia

“…We purchased the early-stage leukemia cell lines REH and RS4;11 from American Type Culture Collection (ATCC, USA) and procured the late-stage leukemia cell lines BALL-1 and MN60, from RIKEN BioResource Research Center (Japan) and DSMZ (Germany), respectively. All the four cell lines were cultured according to previously described protocols before imaging measurements [40].…”