Lymphoblastic leukaemia (blood cancer) detection: AI-Assisted Diagnosis Through Blood Sample Images
DOI:
https://doi.org/10.63503/acset.108Abstract
Leukaemia is a serious disease that affects the blood. It happens when bad white blood cells, called blast cells, grow out of control in the blood and bone marrow. Finding these cells early and accurately is extremely important for figuring out what is wrong, how sick someone is, and which treatment will work best. Doctors usually look at blood slides under a microscope to find these cells. This method works. It takes a long time and the doctors have to be very careful. Sometimes doctors can make mistakes because they get tired or disagree with what they see. This can be a problem, especially when they have to review many blood slides or lack all the resources they need. Leukaemia diagnosis is a process, and leukaemia cells can be hard to find. That is why leukaemia diagnosis needs to be accurate and fast. Leukaemia treatment needs to be planned, and leukaemia patients need the best care possible. Leukaemia is a disease, and leukaemia patients deserve the best chance to get better. In this paper, we describe an AI-powered automatic detection system of leukaemia blast cells using deep Convolutional Neural Networks (CNNs) that can be used by haematologists in the clinical diagnostic process. The system uses a VGG16-based CNN trained via transfer learning on the C-NMC: B-lineage acute lymphoblastic leukaemia dataset, which consists of more than 10,000 blood cell images annotated by experts across various leukocyte types. Several image preprocessing and data augmentation techniques, as well as a method to address class imbalance, were applied to improve the system's robustness and generalisation. Experimental results show that the proposed system performs on par with human experts, achieving greater than 95% sensitivity for blast cells and an ROC-AUC higher than 0.98. Grad-CAM explainable AI visualisations are included to make these diagnostic inferences more transparent and to foster clinical acceptance, illustrating the cellular areas that contribute most to the prediction. Overall, the framework demonstrates the ability to substantially improve diagnostic consistency and throughput while enabling fast, accurate, and understandable leukaemia screening.
References
[1] World Health Organization, Global Tuberculosis Report 2024. Geneva, Switzerland: WHO Press, 2024.
[2] R. Rana and A. Gaba, “Tuberculosis detection using deep learning,” Int.
J. Adv. Eng. Manag., vol. 4, no. 8, pp. 317–323, Aug. 2022.
[3] S. Vijayan et al., “Implementing chest X-ray AI to enhance TB screening in India,” PLOS Digit. Health, vol. 2, no. 12, Dec. 2023.
[4] M. Nijiati et al., “Deep learning assistance for TB diagnosis in low-resource settings,” J. X-Ray Sci. Technol., Jun. 2021.
[5] U. K. Lopes and J. F. Valiati, “Pre-trained CNNs as feature extractors for tuberculosis diagnosis,” Comput. Biol. Med., vol. 89, pp. 135–143, 2017.
[6] K. He, X. Zhang, S. Ren, and J. Sun, “Deep residual learning for image recognition,” in Proc. IEEE CVPR, 2016, pp. 770–778.
[7] A. Krizhevsky, I. Sutskever, and G. E. Hinton, “ImageNet classification with deep convolutional neural networks,” in Adv. Neural Inf. Process. Syst., vol. 25, pp. 1097–1105, 2012.
[8] S. Hwang et al., “Multi-stage convolutional neural network architecture for medical imaging,” IEEE Trans. Med. Imaging, 2019.
[9] S. Ioffe and C. Szegedy, “Batch normalization: Accelerating deep network training by reducing internal covariate shift,” in Proc. ICML, 2015, pp. 448–456.
[10] D. P. Kingma and J. Ba, “Adam: A method for stochastic optimization,” in Proc. ICLR, 2015.
[11] J. Smith et al., “Inter-radiologist variability in chest X-ray interpreta-tion,” Radiology, vol. 45, no. 3, pp. 112–120, 2023.
[12] G. Litjens et al., “A survey on deep learning in medical image analysis,” Med. Image Anal., vol. 42, pp. 60–88, 2017.
[13] R. R. Selvaraju et al., “Grad-CAM: Visual explanations from deep networks via gradient-based localization,” in Proc. IEEE ICCV, 2017,
pp. 618–626.
[14] Y. LeCun, Y. Bengio, and G. Hinton, “Deep learning,” Nature, vol. 521, no. 7553, pp. 436–444, 2015.
[15] S. Rajpurkar et al., “CheXNet: Radiologist-level pneumonia detection on chest X-rays with deep learning,” arXiv:1711.05225, 2017.
[16] T. Pande et al., “Sensitivity–specificity tradeoffs in AI diagnostics,” Lancet Digit. Health, vol. 6, no. 4, 2024.
[17] M. Brouwer et al., “Workflow integration of AI tools in resource-limited clinics,” Front. Public Health, vol. 11, 2024.
[18] S. Jaeger et al., “Two public chest X-ray datasets for computer-aided screening of pulmonary diseases,” Quant. Imaging Med. Surg., vol. 4, no. 6, pp. 475–477, 2014.
[19] T. Rahman et al., “Reliable tuberculosis detection using chest X-ray with deep learning, segmentation and visualization,” IEEE Access, vol. 8, pp. 191586–191601, 2020.
[20] R. Wajgi et al., “Optimized tuberculosis classification system for chest X-ray images: Fusing hyperparameter tuning with transfer learning approaches,” Eng. Rep., vol. 6, no. 11, 2024.
Downloads
Published
Conference Proceedings Volume
Section
License
Copyright (c) 2026 Adroid Conference Series: Engineering and Technology
This work is licensed under a Creative Commons Attribution 4.0 International License.
