Unsupported Browser
The American College of Surgeons website is not compatible with Internet Explorer 11, IE 11. For the best experience please update your browser.
Menu
Become a member and receive career-enhancing benefits

Our top priority is providing value to members. Your Member Services team is here to ensure you maximize your ACS member benefits, participate in College activities, and engage with your ACS colleagues. It's all here.

Become a Member
Become a member and receive career-enhancing benefits

Our top priority is providing value to members. Your Member Services team is here to ensure you maximize your ACS member benefits, participate in College activities, and engage with your ACS colleagues. It's all here.

Membership Benefits
ACS
From the Director's Desk

Artificial Intelligence

June 22, 2023

heidi-nelson.jpg

Heidi Nelson, MD, FACS
Medical Director, ACS Cancer Programs

No matter what news media you follow these days, it is hard to ignore the many conversations swirling around the future of artificial intelligence (AI). Is AI more likely to help us solve problems that exceed human cognitive capacity? Will AI take over tasks that humans now perform? While ACS Cancer Programs is not engaged in these public debates, we are engaged in the use of AI, so perhaps now is the time for us to join in the conversation.

Before proceeding, it may be helpful to clarify what we mean by AI and how it applies to the work of ACS Cancer Programs. In general terms, AI refers to the use of computer science to simulate human intelligence; one form of AI is referred to as “deep learning.” Over the course of the last year, the Statistical and Analytic team of the National Cancer Database (NCDB) has been using deep learning to evaluate and optimize the American Joint Committee on Cancer TNM staging for multiple disease types.

At first glance, this may not seem like a major advance since NCDB survival data have been employed in the past to substantiate staging standards. In reality, however, the new AI approach affords the opportunity for the machine to run all the data against the endpoint to inform us as to which tumor factors correlate most closely with survival. In other words, rather than testing multiple hypotheses, the machine can rank and prioritize tumor factors based on using all the available data upfront, rather than testing and validating individual observations. Standardizing deep learning methodologies has brought the added benefit of allowing us to compare results across disease types.

While applying AI to NCDB data has shown early wins and is an important step toward delivering a personalized approach to estimating survival, many of you may be wondering when AI might help on the other end, i.e., the data abstraction process. Stay tuned. This will be covered in a future column.