Using machine learning for the diagnosis of non-Hodgkin's lymphoma: Systematic review

Introduction: Non-Hodgkin's lymphoma (NHL) is a type of hematological malignancy. It affects the lymphaticsystem. It is a difficult task that requires expertise, considerable experience and meticulous morphologicalexamination to identify and subtype NHL .Accurately identifying the type of lymphoma is crucial for early treatmentand a positive prognosis. With the advancement of technology and the increasing growth of artificial intelligence, theuse of this technology in the diagnosis of non-Hodgkin's cancer has been considered. In this systematic review article,various studies conducted in the field of using artificial intelligence to diagnose non-Hodgkin's cancer wereexamined.Method: This systematic review was conducted in ۲۰۲۳. The search for relevant studies included electronic databasessuch as Web of Science, Cochrane, Scopus and PubMed using the keywords "machine learning" [Mesh], "deeplearning" [Mesh] and "non-Hodgkin's lymphoma" [Mesh]. The inclusion criteria were limited to articles with fulltext available from ۲۰۱۵ to ۲۰۲۳ and articles not meeting the research topic were excluded. Ultimately, ۲۷ articlesrelated to the topic were included in the study using entry and exit criteria (following the PRISMA checklist). Thestudies were reviewed based on the inclusion criteria (the Englishness, the availability, and the related of the studies)and those studies whose full text was not available and were not related to the topic were excluded from the review.And finally, to avoid biasing the final studies by the tools of CASP were evaluated.Results: It can be concluded that there have been significant advances in the application of artificial intelligence tothe diagnosis of non-Hodgkin's lymphoma, based on an analysis of the reviewed literature high levels of accuracyand precision have been demonstrated using techniques such as artificial neural networks, genetic algorithms andexpert systems to determine the presence of NHL. Furthermore, using medical images and clinical data to inputartificial intelligence has led to improved diagnosis of NHLConclusion: Artificial intelligence-aided sequencing holds potential to enhance diagnosis accuracy and efficiency,detect novel biomarkers, and forecast treatment outcomes. Nonetheless, restricted databases, lacking standardizationand validation, systematic errors, and partiality pose obstacles to AI replacing manual diagnosis in hematology. Theuse of AI to process large amounts of patient data and personal information raises privacy issues in clinical AIsystems, requiring the creation of regulations to evaluate AI systems and ethical concerns. However, with continuedresearch and development, AI has the potential to revolutionize the field of hematology and improve patient outcomes