A systematic review of algorithms for breast cancer diagnosis using thermography.

Abstract

Breast cancer tumorigenesis usually takes place in the cells of the mammary ducts and poses a significant health challenge globally, as it remains the most common cancer type. Early diagnosis and advancements in treatment have substantially enhanced survival rates for breast cancer. Currently, mammography, ultrasound, and magnetic resonance imaging are the primary techniques for early breast cancer detection. In contrast, thermography-based diagnostics, despite its decreased cost when compared to all other techniques, remains underused. This happens largely due to the challenges faced by current computational methods to correctly classify thermography-based images. This study conducts a systematic literature review to identify and understand which artificial intelligence algorithms are currently employed in diagnosing breast cancer using such thermographic images. Additionally, it seeks to map the main challenges stemming from this specific research area. To achieve this, a significant number of international studies were scrutinized, considering criteria such as: employed methodology, databases utilized, publication venues, impact factors, algorithms applied, overall contributions to the field, findings, and conclusions.