ARTIFICIAL INTELLIGENCE IN ONCOLOGY: CURRENT CAPABILITIES,FUTURE OPPORTUNITIES AND ETHICAL CONSIDERATIONS

Abstract

Artificial intelligence in oncology is no longer a hypothetical concept, and its U.S. Food and Drug Administration (FDA) – approved use is expanding in several clinical scenarios, most notably in cancer diagnostics and computer vision applications. AI applications are in various stages of development across the continuum of oncology and in multidisciplinary practice, with some algorithms and advanced clinical decision support systems (CDSS) demonstrating capabilities equivalent to or even surpassing expert interventions. There are unique ethical and legal considerations associated with AI models, which limit their widespread application and reproducibility, including inherent bias when trained on datasets that disproportionately exclude underrepresented populations. Barriers to the broad adoption of AI involve ideological and organizational challenges, as well as a limited number of prospective validation studies; however, healthcare modernization is gradually reducing these obstacles and supporting its responsible use. The future of precision oncology, where “living” databases of multimodal data types are recursively used to improve clinical models, may lead to unprecedented outcomes for patients.