The Adoption of Artificial Intelligence in Cancer Pathology and Imaging

Abstract

Artificial intelligence (AI) is revolutionizing how the world operates, in manufacturing, in transportation, in advertising, and in criminal justice, how we communicate, and how we practice medicine. AI had undergone cyclical interests and disinterests over the past few decades, but recent surge of interest in AI owing to the emerging practical deep learning technologies that have caught the healthcare industry’s attention. Obtaining key understanding of AI is essential in incorporating such disruptive technologies into healthcare. The medical field is diverse and complex leading to a myriad of barriers in optimal integration in the clinical workflow, with some specialties reaping better success incorporating AI into their practices than other specialties. Medical imaging, notably radiology, is one success story due to the digital nature of their workload, allowing simple conversion for AI to complement their practices and to automate mundane tasks. On the other hand, pathology has limited success in combining pathological practice and AI due to a lack of supporting digital infrastructure and a complex imaging workflow. Other complications include the fear of the ambiguity, legal representation, responsibility and culpability, and significant capital expenditures necessary to keep the industry up-to-date, hindering the ability of AI to transform the healthcare landscape. Should these conundrums be addressed, AI could prove to be an incredibility powerful tool for improving the practice and productivity of oncologists and imaging specialists.