Colorectal cancer poses a significant global health challenge, with over 100,000 individuals in the Netherlands alone undergoing colonoscopies annually to detect and prevent this disease. Traditionally, colonoscopies involve manual visual inspection by a physician to identify and remove polyps, which can develop into colorectal cancer. However, recent advancements in artificial intelligence (AI) have shown promising potential in enhancing the efficiency of this screening process. This article delves into the role of AI in colonoscopy and its potential implications for colorectal cancer prevention.Â
Medical researcher Michiel Maas underscores the limitations of conventional colonoscopies, where polyps may go undetected due to brief appearances in images or the physician’s divided attention. Here, AI emerges as a valuable asset, capable of analyzing colon camera photos swiftly and accurately. Maas emphasizes that while human attention may falter, AI remains vigilant, meticulously scanning every pixel of the image within milliseconds. This capability significantly improves the detection rate of polyps, thereby mitigating the risk of colorectal cancer in the future.Â
A groundbreaking study, spanning ten hospitals across various nations and involving nearly 1,000 patients eligible for colorectal cancer screening or with a history of polyps, showcased the effectiveness of AI-assisted colonoscopy. Patients were presented with the option of undergoing a live-monitored computer examination or a traditional colonoscopy. The AI system highlighted suspicious areas in purple, alerting physicians to regions warranting further examination.Â
Findings from the study revealed that AI-assisted colonoscopy detected and removed nearly forty percent more polyps compared to conventional methods. Importantly, this increased detection did not result in a higher rate of false positives or unnecessary tissue removal. Polyps identified and removed with AI assistance were equally likely to be hazardous as those identified during routine checks. Furthermore, despite the augmented detection, colonoscopies with digital support were completed expeditiously, underscoring the efficiency of AI integration.Â
The intelligent computer system utilized in the study was developed by Magentiq Eye LTD, leveraging a vast dataset of colonoscopy images and recordings from around the world to train the software in anomaly detection. While the system is commercially available, its widespread adoption is impeded by reimbursement challenges associated with medical examinations incorporating AI assistance.Â
Future research endeavors aim to explore additional functionalities of AI in colonoscopy, such as assessing the risk level of polyps based on their size and visual characteristics. Professor and gastroenterologist Peter Siersema underscores the potential to leave benign polyps in situ if their risk level can be accurately determined during examination, thereby reducing the necessity for further pathological assessment. Ultimately, such advancements hold promise in mitigating complications and curtailing healthcare expenditures linked with colorectal cancer screening and prevention.Â
The integration of AI into colonoscopy represents a significant stride in colorectal cancer screening, offering heightened detection rates and efficiency compared to conventional methods. The success of AI-assisted colonoscopy in pinpointing and extracting polyps underscores its potential to enhance patient outcomes and alleviate the global burden of colorectal cancer. Ongoing research and development in this domain are poised to yield further enhancements in screening precision and effectiveness, ultimately fostering improved outcomes for individuals at risk of colorectal cancer.Â
Journal Reference Â
Michiel H J Maas et al, A computer-aided polyp detection system in screening and surveillance colonoscopy: an international, multicentre, randomised, tandem trial, The Lancet Digital Health (2024). DOI: 10.1016/S2589-7500(23)00242-XÂ
