ALL-ViT: A Novel Approach for Detection of Acute Lymphoblastic Leukemia
Praveen Pandey, Rohit Tarun RG, Rudra Prasad Pati, Sofia Singh · 2025
Acute Lymphoblastic Leukemia (ALL) is a life menatorial hematologic malignancy, responsible for 3.1% of lukemia-related deaths globally in 2020[6]. The high mortality rate associated with ALL is primarily attributed to the lack of early detection methods, which results in delayed diagnosis and limited treatment outcomes[1]. This research presents a cutting-edge transformer-based artificial intelligence (AI) model designed to address this critical challenge by enabling early detection of ALL with unparalleled accuracy and reliability. Our proposed model leverages advanced deep learning techniques to identify various stages of ALL, including the early, intermediate, and advanced stages, using over 6512 cell images from the ALL-Image dataset [15]. Rigorous training and evaluation of the model demonstrate a remarkable predictive accuracy of 99.9%, underscoring its potential as a highly reliable diagnostic tool[3]. By distinguishing different ALL stages with precision, the model offers a significant advantage in facilitating timely intervention, improving prognosis, and reducing mortality rates. Implementing this model in real-world clinical settings could revolutionize the diagnostic landscape for ALL, enabling healthcare practitioners to achieve early detection and optimize remedy strategies. This research contributes to the growing field of AI-led healthcare solutions, offering a transformative approach to combating hematologic malignancies.