Deep Learning Techniques in Acute Lymphoblastic Leukemia

Abstract

Acute Lymphoblastic Leukemia (ALL) is a serious hematologic malignancy that
requires prompt and accurate diagnosis for effective treatment. Traditional diagnostic
approaches, such as microscopic inspection and flow cytometry, while reliable, are labor
intensive and time-consuming. Deep learning (DL), particularly Convolutional Neural
Networks (CNNs), has revolutionized ALL diagnosis by automatically extracting features
from medical images with an accuracy comparable to experienced hematologists. This
automation enhances diagnostic precision and significantly reduces diagnosis time,
ultimately improving clinical decision-making and patient outcomes. This study uses deep
learning through rigorous bibliometric analysis to examine research productivity indicators
related to the classification of Acute Lymphoblastic Leukemia. The dataset comprises 152
publications on ALL classifications using deep learning from 2009 to 2023. The analysis
is conducted using the R Bibliometrix library. These publications, authored by 654
researchers from 38 countries, are distributed across 99 sources, including journals and
books. Collectively, the corpus has received 1,770 citations, with an average of 11.8
citations per document. The most-cited paper, Leukemia Diagnosis in Blood Slides Using
Transfer Learning in CNNs and SVM for Classification by Luis H.S. Vogado et al., has
been cited 143 times. In terms of global scientific output, India leads with 26 publications
(17.3%), followed by China with 22 (14.7%) and the United States with 16 (10.7%).
Keywords: bibliometric analysis; R Bibliometrix; Acute Lymphoblastic Leukemia; Deep
Learning; Web of Science.