Automatic Detection of Acute Leukemia (ALL and AML) Utilizing Customized Deep Graph Convolutional Neural Networks

Abstract

Leukemia is a malignant disease that impacts explicitly the blood cells, leading to life-threatening infections and premature mortality. State-of-the-art machine-enabled technologies and sophisticated deep learning algorithms can assist clinicians in early-stage disease diagnosis. This study introduces an advanced end-to-end approach for the automated diagnosis of acute leukemia classes acute lymphocytic leukemia (ALL) and acute myeloid leukemia (AML). This study gathered a complete database of 44 patients, comprising 670 ALL and AML images. The proposed deep model's architecture consisted of a fusion of graph theory and convolutional neural network (CNN), with six graph Conv layers and a Softmax layer. The proposed deep model achieved a classification accuracy of 99% and a kappa coefficient of 0.85 for ALL and AML classes. The suggested model was assessed in noisy conditions and demonstrated strong resilience. Specifically, the model's accuracy remained above 90%, even at a signal-to-noise ratio (SNR) of 0 dB. The proposed approach was evaluated against contemporary methodologies and research, demonstrating encouraging outcomes. According to this, the suggested deep model can serve as a tool for clinicians to identify specific forms of acute leukemia.

Keywords: ALL; AML; deep learning networks; graph; leukemia.

Figure 1

Different types of acute leukemia, including (a) AML, (b) ALL, (c) CML, and (d) CLL.

Figure 2

The proposed primary framework for the automated diagnosis of acute leukemia involves categorizing it into two classifications: ALL and AML.

Figure 3

The data collection process for ALL and AML classes.

Figure 4

An example of the images taken in the proposed database for (a) ALL and (b) AML.

Figure 5

Graphic view of the proposed architecture.

Figure 6

Details of the layers in deeply organized architecture.

Figure 7

Five-fold cross-validation operation.

Figure 8

The graph convolutional architecture was tested with various polynomial coefficients.

Figure 9

Different polynomial coefficients were examined in the graph convolutional architecture.

Figure 10

Specific locations were chosen for the ALL and AML samples for graph embedding.

Figure 11

The suggested model’s accuracy and error for training and validation assessed across 150 iterations.

Figure 12

The classification accuracy results based on 5-fold cross-validation criteria.

Figure 13

Performance of confusion matrix and ROC curve in the proposed model.

Figure 14

Examples of two categories of veracity and falsity for unprocessed data and the fully connected network layer.

Figure 15

The performance of the proposed network compared to other networks.

Figure 16

A sample of images with varying decibels of noise applied.

Figure 17

The performance of the proposed network in relation to other networks.