Artificial Intelligence in Early Cancer Detection ARTIFICIAL INTELLIGENCE The ability of a computer to perform tasks commonly associated with human intelligence, such as how to act, reason, and learn. MACHINE LEARNING A type of AI that is programmed to learn over time from the data provided to make predictions or decisions; the more comprehensive and inclusive data provided to an ML model, the better it will perform. DEEP LEARNING A type of ML that learns from huge amounts of data using complex algorithms, called artificial neural networks, that are modeled after how the human brain processes information. BENEFITS POTENTIAL DRAWBACKS Potential benefits of AI-based approaches in early detection are speed and accuracy to detect existing cancers or to rule out that cancer is present. This may allow for better surveillance and intervention if/when needed. Two of the most common AIenhanced approaches for cancer early detection and diagnosis include: • Detecting and classifying cancerous tumors using various scans from radiological or pathological imaging. • Combining conventional blood tests as well as liquid biopsies with AI-powered analyses for cancer diagnosis. • The lack of large, well-annotated cancer datasets that are representative of the diversity of the population as well as of the distinct cancer burdens of various population groups is a significant barrier for the use of AI in cancer research and patient care. • Because the use of AI in medicine is a newer technological advance, there may be concerns among patients about the use of their clinical information, including images and tissues. • The use of AI in early detection and diagnosis of cancer is still in its infancy; inequitable use of AI may exacerbate disparities. All stakeholders must work together to ensure that the evidence is developed to show the benefit beyond detection alone, potential biases are minimized, and, when proven, there is equitable access of this technological advance. SELECTED EXAMPLES OF AI-BASED DEVICES AND SOFTWARE IN CANCER DETECTION ProstatID Approved by FDA in July 2022, is an AI-based software that uses AI to measure the volume of the prostate gland from images of the prostate and detect suspicious cancerous lesions in scans obtained using traditional magnetic resonance imaging. SKOUT Approved by FDA in September 2022, is a medical device that uses advanced computer vision technology designed to recognize polyps and abnormal tissue and provide real-time feedback to gastroenterologists during colonoscopy (313). Developed from (314). AI ML DL JULY 2022 SEPTEMBER 2022 SIDEBAR 24 AACR Cancer Progress Report 2023 Screening for Early Detection 66
RkJQdWJsaXNoZXIy NTkzMzk=