A Review of Artificial Intelligence Applications in Hematology Management: Current Practices and Future Prospects

Abstract

Background Machine learning (ML) and deep learning (DL) methods have recently garnered a great deal of attention in the field of cancer research by making a noticeable contribution to the growth of predictive medicine and modern oncological practices. Considerable focus has been particularly directed toward hematologic malignancies because of the complexity in detecting early symptoms. Many patients with blood cancer do not get properly diagnosed until their cancer has reached an advanced stage with limited treatment prospects. Hence, the state-of-the-art revolves around the latest artificial intelligence (AI) applications in hematology management. Objective This comprehensive review provides an in-depth analysis of the current AI practices in the field of hematology. Our objective is to explore the ML and DL applications in blood cancer research, with a special focus on the type of hematologic malignancies and the patient’s cancer stage to determine future research directions in blood cancer. Methods We searched a set of recognized databases (Scopus, Springer, and Web of Science) using a selected number of keywords. We included studies written in English and published between 2015 and 2021. For each study, we identified the ML and DL techniques used and highlighted the performance of each model. Results Using the aforementioned inclusion criteria, the search resulted in 567 papers, of which 144 were selected for review. Conclusions The current literature suggests that the application of AI in the field of hematology has generated impressive results in the screening, diagnosis, and treatment stages. Nevertheless, optimizing the patient’s pathway to treatment requires a prior prediction of the malignancy based on the patient’s symptoms or blood records, which is an area that has still not been properly investigated.

Manually-established abnormal karyotype dataset based on normal chromosomes effectively train artificial intelligence model for better cytogenetic abnormalities prediction

“…The application of AI has been well-known to contribute to hematology management 5 , ranging from prediction, screening, diagnosis and treatment. For instance, machine learning models are generated to stratify disease subclassi cation in acute myeloid leukemia (AML) 6 , predict minimal residual disease (MRD) prognostication in multiple myeloma 7 and recognize the risk level of plasma cell myeloma (PCM) 8 , making AI model a signi cant step towards automation of hematological analysis with high accuracy.…”

“…For instance, machine learning models are generated to stratify disease subclassi cation in acute myeloid leukemia (AML) 6 , predict minimal residual disease (MRD) prognostication in multiple myeloma 7 and recognize the risk level of plasma cell myeloma (PCM) 8 , making AI model a signi cant step towards automation of hematological analysis with high accuracy. Indeed, the manual diagnosis of a blood cancer is costly, time-consuming and even with high rate of misdiagnosis 5,9 , highlighting the potential use of AI models to solve these unbearable problems.…”

“…Indeed, consensus regarding karyotype complexity suggesting malignant prognosis has even been proposed, emphasizing its signi cance 13 . However, similar to the challenges of AI implementation in other medicine-related applications 5 , the lack of representative abnormal karyotype images of hematological malignancies hinder the data analysis and further accurate diagnosis in clinic. Even there are databases existing, they fail to provide a comprehensive dataset covering almost all the abnormal karyogram images 14 .…”

Machine Learning for Diagnosis and Screening of Chronic Lymphocytic Leukemia Using Routine Complete Blood Count (CBC) Results

“…Furthermore, many AI methods reported so far in the context of CLL mainly investigate disease prognosis after CLL diagnosis [5]. Hence, reviews point out that the applicability of AI in the diagnosis of CLL is the least explored areas in hematology management, wherein further research is essential [6]. To bridge this gap, we demonstrate in our paper the potential of a ML algorithm to leverage the power of simple blood count test outcome in classifying patients with chronic lymphocytic leukemia (CLL).…”

An Artificial Intelligence-Based Diagnostic System for Acute Lymphoblastic Leukemia Detection

“…Moreover, Israet et al [5] employed Convolutional Neural Networks (CNNs) for effective blood image classification and leukocyte segmentation, which Roy et al [6] also achieved by alternatively developing a system based on DeepLabv3 and ResNet-50 to extract deep feature maps from the rest of the blood images. While the power of automation and Artificial Intelligence (AI) was greatly exploited through different applications of Machine Learning (ML) and Deep Learning (DL) on blood and bone marrow images for abnormality identification, microscopic examination was still considered challenging and computationally expensive [7]. In majority of cases, the input blood cell raw images require intensive enhancement and preprocessing to achieve higher quality images for future classifications.…”