Artificial intelligence in cancer diagnosis: a systematic review

Keywords

artificial intelligence - cancer diagnosis - screening

Introduction: Cancer is one of the leading causes of global mortality, highlighting the need for early diagnosis to improve clinical outcomes. Artificial intelligence (AI) has emerged as a promising tool that offers precise interventions in medical practice. In recent years, AI has established itself as a significant innovation, such as in mammography and tomosynthesis devices, revolutionizing medical diagnosis through advances in detailed and predictive analyses.

Objective: Analyze AI interventions in cancer diagnosis and contribute to early screenings.

Method: Articles from the PubMed database were reviewed, the inclusion criteria being: topics related to the use of AI in cancer diagnosis from 2019 to 2023, in English and Portuguese. The search terms used were “artificial intelligence”, “cancer”, and “diagnosis”.

Results: Among the 33 articles analyzed, 21 showed that AI is essential in the early diagnosis of cancer, due to its precision and lower failure rates, as well as its ability to direct and reduce treatment costs. AI handles large amounts of data, aiding in the prediction of potential protein interactions and gene expression. This tool facilitates the discovery of new drugs at lower costs and in less time. For example, the smartphone app based on the Skinvision algorithm enables patients to receive a quicker diagnosis, as this technology evaluates skin lesions in 30 seconds and detects 95% of skin cancer at an early stage - although medical intervention is still necessary afterward. Nine articles indicated that AI is promising, but lacks concrete evidence compared to specialized doctors. Finally, three articles did not conclude that AI is superior for cancer diagnosis.

Conclusion: The use of AI in oncology has proven promising for diagnosing and monitoring cellular changes, given its ease, precision and safety in healthcare. This has already been demonstrated in robotic and nanorobotic surgeries, drugs administration and the identification of abnormalities in radiological images, showing a high sensitivity compared to conventional technologies. Therefore, although AI does not replace human intelligence or interaction, it is a valuable ally in medical practice, working across various areas of oncology such as screening, early diagnosis and treatment.

Corresponding author: Larissa Maria Moraes Rodrigues De Souza (e-mail: [e-mail]).