The Use of Artificial Intelligence in Cancer Diagnosis.

Abstract

Artificial intelligence (AI) is increasingly influencing cancer diagnostics, with ongoing advancements into more refined and advanced applications. This review examines current applications of AI in cancer diagnosis. AI has demonstrated considerable success in screening and diagnosis of a range of cancer types, including breast, colorectal, and skin cancer, but challenges with specificity and false-positive rates remain. Within radiology and histopathology, AI shows promise in identifying cancers and augmenting clinician performance. Natural language processing and AI-driven clinical documentation tools can expedite diagnosis by reducing administrative burden and extracting clinically meaningful data from unstructured records. Additionally, AI models using unimodal and multimodal data can improve patient prognostication and risk stratification. Despite these advances, challenges relating to deployment, bias, interpretability, accountability and real-world efficacy persist. Prospective validation and careful implementation are required to ensure safe, equitable, and effective adoption of AI in cancer diagnostics.