ALLNet: A Hybrid Convolutional Neural Network to Improve Diagnosis of Acute Lymphocytic Leukemia (ALL) in White Blood Cells

Sai Mattapalli,Rishi Athavale

DOI: https://doi.org/10.1109/BIBM52615.2021.9669840

2022-07-29

Abstract:Due to morphological similarity at the microscopic level, making an accurate and time-sensitive distinction between blood cells affected by Acute Lymphocytic Leukemia (ALL) and their healthy counterparts calls for the usage of machine learning architectures. However, three of the most common models, VGG, ResNet, and Inception, each come with their own set of flaws with room for improvement which demands the need for a superior model. ALLNet, the proposed hybrid convolutional neural network architecture, consists of a combination of the VGG, ResNet, and Inception models. The ALL Challenge dataset of ISBI 2019 (available here) contains 10,691 images of white blood cells which were used to train and test the models. 7,272 of the images in the dataset are of cells with ALL and 3,419 of them are of healthy cells. Of the images, 60% were used to train the model, 20% were used for the cross-validation set, and 20% were used for the test set. ALLNet outperformed the VGG, ResNet, and the Inception models across the board, achieving an accuracy of 92.6567%, a sensitivity of 95.5304%, a specificity of 85.9155%, an AUC score of 0.966347, and an F1 score of 0.94803 in the cross-validation set. In the test set, ALLNet achieved an accuracy of 92.0991%, a sensitivity of 96.5446%, a specificity of 82.8035%, an AUC score of 0.959972, and an F1 score of 0.942963. The utilization of ALLNet in the clinical workspace can better treat the thousands of people suffering from ALL across the world, many of whom are children.

Computer Vision and Pattern Recognition,Artificial Intelligence,Machine Learning,Quantitative Methods

What problem does this paper attempt to address?

The problem that this paper attempts to solve is that at the microscopic level, the morphological similarity between acute lymphoblastic leukemia (ALL) cells and healthy cells makes it difficult to quickly and accurately distinguish between the two types of cells. This challenge is crucial for the early diagnosis of ALL, because ALL is a rapidly progressing cancer and can be fatal within a few months if not treated in a timely manner. In addition, ALL is one of the most common types of childhood cancers, especially affecting children under 5 years old, which further emphasizes the importance of paying sufficient attention to it. To solve this problem, the paper proposes a hybrid convolutional neural network architecture named ALLNet, aiming to improve the diagnostic accuracy of ALL by combining the advantages of three common models: VGG, ResNet, and Inception. By using 10,691 white blood cell images in the ISBI 2019 ALL Challenge dataset for training and testing, ALLNet has shown better performance than VGG, ResNet, and Inception models on both the cross - validation set and the test set, especially in terms of accuracy, sensitivity, specificity, AUC score, and F1 score. The proposal of ALLNet not only helps to improve the diagnostic efficiency of ALL, but also makes it possible to better treat thousands of ALL patients in the clinical workspace, especially those child patients. By using an automated method to identify whether a given cell image is an ALL cell or a healthy cell, doctors can quickly obtain diagnostic results after obtaining the patient's blood cell image, which is of great significance for the fight against pediatric cancer in a race against time.

ALLNet: A Hybrid Convolutional Neural Network to Improve Diagnosis of Acute Lymphocytic Leukemia (ALL) in White Blood Cells

ALNett: A cluster layer deep convolutional neural network for acute lymphoblastic leukemia classification

An ensemble-acute lymphoblastic leukemia model for acute lymphoblastic leukemia image classification

Automatic classification of acute lymphoblastic leukemia cells and lymphocyte subtypes based on a novel convolutional neural network

An attention-based deep learning for acute lymphoblastic leukemia classification

DSCNet: Deep Skip Connections-Based Dense Network for ALL Diagnosis Using Peripheral Blood Smear Images

A fast and efficient CNN model for B‐ALL diagnosis and its subtypes classification using peripheral blood smear images

SDCT-AuxNet$^θ$: DCT Augmented Stain Deconvolutional CNN with Auxiliary Classifier for Cancer Diagnosis

Acute Lymphoblastic Leukemia Detection from Microscopic Images Using Weighted Ensemble of Convolutional Neural Networks

A study on deep feature extraction to detect and classify Acute Lymphoblastic Leukemia (ALL)

BO-ALLCNN: Bayesian-Based Optimized CNN for Acute Lymphoblastic Leukemia Detection in Microscopic Blood Smear Images

AUTOMATED ACUTE LYMPHOBLASTIC LEUKEMIA CELL CLASSIFICATION USING OPTIMIZED CONVOLUTIONAL NEURAL NETWORK

A Hybrid Feature Fusion Deep Learning Framework for Leukemia Cancer Detection in Microscopic Blood Sample Using Gated Recurrent Unit and Uncertainty Quantification

Deep Learning Algorithms for Early Diagnosis of Acute Lymphoblastic Leukemia

Refinement of ensemble strategy for acute lymphoblastic leukemia microscopic images using hybrid CNN-GRU-BiLSTM and MSVM classifier

Novel Pipeline for Diagnosing Acute Lymphoblastic Leukemia Sensitive to Related Biomarkers

A Diagnostic Model for Acute Lymphoblastic Leukemia Using Metaheuristics and Deep Learning Methods

Deep learning model for differentiating acute myeloid and lymphoblastic leukemia in peripheral blood cell images via myeloblast and lymphoblast classification

Acute lymphoblastic leukemia detection using ensemble features from multiple deep CNN models

A Framework for Early Detection of Acute Lymphoblastic Leukemia and Its Subtypes From Peripheral Blood Smear Images Using Deep Ensemble Learning Technique

A deep transfer learning based approaches for the detection and classification of acute lymphocytic leukemia using microscopic images