ARTIFICIAL INTELLIGENCE TECHNIQUES FOR EARLY CANCER DIAGNOSIS: A LITERATURE REVIEW(Atena Editora)

Abstract

Objective: To evaluate how Artificial Intelligence (AI) techniques can be applied to facilitate the early diagnosis of cancer. Methods: Narrative bibliographic review using the PubMed database, applying the search strategy: (artificial intelligence) AND (diagnosis) AND (cancer). 4,119 articles were found and after applying the inclusion and exclusion criteria, only 17 articles were selected to compose the study. Discussion: The role of Artificial Intelligence (AI) in the early diagnosis of cancer is highlighted, with an emphasis on Machine Learning (ML) and Deep Learning (DL) techniques. ML uses algorithms to analyze patterns in large data sets, improving cancer diagnosis and treatment. DL, a subset of ML, uses multi-layer neural networks to interpret complex data, such as medical images, improving accuracy in identifying neoplasms. Although promising, these technologies face challenges such as the applicability of AI processes and the interpretation of genomic data, aiming to advance precision oncology and improve clinical practice. Such procedures have also been criticized for not explicitly highlighting how the model analyzes the data and makes decisions based on certain inputs, raising ethical questions about their applicability. Final considerations: Therefore, Artificial Intelligence and Deep Learning (DL) and Machine Learning (ML) emerge as crucial tools in the diagnosis and early treatment of cancer, improving patient survival and the efficiency of radiotherapy. Although promising, these technologies face accessibility challenges and the need for greater understanding in their application in oncology.