Adewole Adamson, MD, an assistant professor in the department of internal medicine at Austin, Texas-based Dell Medical School, and H. Gilbert Welch, MD, a senior investigator in the center for surgery and public health at Boston-based Brigham and Women’s Hospital, described this challenge in a Dec. 12 perspective article for The New England Journal of Medicine.
At the core of the issue is a lack of definition in the “gray zones” between tissue that is cancerous and not cancerous. Because pathologists still lack a “gold standard” to determine what constitutes a clinically meaningful lesion, the authors wrote, algorithms are trained only to recognize designated patterns in imaging data, and are therefore likely to classify as cancerous certain abnormalities that a human expert may conclude to be asymptomatic or completely benign.
“Machine learning cannot solve the gold-standard problem, but it could further expose it. Ultimately, what matters to patients and clinicians is whether the diagnosis of cancer has relevance to the length or quality of life,” the authors concluded. “We believe that the possibility of training machine-learning algorithms to recognize an intermediate category between ‘cancer’ and ‘not cancer’ should be given serious consideration before this technology is widely adopted. Highlighting the existence of gray areas could present an important opportunity for pathologists to discuss decisions about what constitutes cancer.”
More articles on AI:
Brigham and Women’s, Mass General test AI-driven patient safety platform
80% of healthcare professionals say AI has decreased burnout: 3 notes
Google is developing AI to do more with ‘much less data’
