This project has been shortlisted for the DPH 2025 Innovation Prize – Best Data Driven Innovation
Team: Martin Chi-Sang Wong, Jason Junjie Huang, Claire Zhong, Thomas Lam, Louis Lau and Philip Chiu (Chinese University of Hong Kong)
Outline: Colorectal cancer (CRC) constitutes a substantial proportion of disease burden globally, and its incidence continues to out-rank other cancers in developed countries. The overall prevalence rates of adenoma (23.9%; 95% CI, 22.2%-25.8%), advanced adenoma (4.6%; 95% CI, 3.8%-5.5%), and CRC (0.4%, 95% CI, 0.3%-0.5%) remain high worldwide. Over the past decade, there has been an increasing trend of CRC incidence in a large number of countries, which necessitates more intensive preventive initiatives. A recent study revealed that CRC screening by colonoscopy could reduce CRC-related incidence and mortality by approximately 18% and 11%, respectively. However, conventional colonoscopy bears a major limitation where up to 27% of adenomatous polyps could be missed due to cognitive and technical capabilities. Our research team introduced AI-assisted colonoscopy for CRC detection in early 2021, when the technology was new. After analysing a large number of endoscopic images, the AI system can evaluate new endoscopic images instantly during colonoscopy to alert doctors to identify adenomas and tumours in real time accurately via digital technology.
Therefore, our research team conducted a study from 2021 to 2022, recruiting 22 junior endoscopists-in-training with <3 years of experience who had performed less than 500 endoscopies to study their performance using the AI-assisted endoscopic system. They performed colonoscopies on 766 patients, among whom 386 were assigned AI-assisted colonoscopy and the rest received conventional colonoscopy without AI. The findings showed that junior endoscopists-in-training achieved about a 40% increase in adenoma detection rate relatively, with the use of AI. The performance was even more impressive among those endoscopists at beginner level, with adenoma detection rate of small adenomas increasing from 25% (traditional colonoscopy) to more than 40% (AI-assisted colonoscopy), implying its substantial success in identifying and characterising premalignant or malignant colorectal lesions.
In addition, we have performed a cost-effectiveness analysis on the adoption of AI-assisted colonoscopy as compared to other CRC screening modalities. A hypothetical population of 100,000 average-risk individuals aged 50 years with no past history or symptoms of CRC was set up. The data were entered into a decision analysis algorithm based on a Markov model. The entire population received four distinct screening strategies. These screening participants were followed up until the age of 75 years. When compared with [Faecal Immunochemical Tests (FIT) + colonoscopy], use of [FIT + AI colonoscopy] could lead to significantly smaller total loss of cancer-related life-years (5,355 vs. 5,327); higher number and proportion of CRC cases prevented (120 vs. 132, 3.7% vs. 4.1%), more life-years saved (280 vs. 308), and lower total cost per life-year saved (US$944,008 vs. US$854,367). [FIT + AI colonoscopy] had the lowest ICER [US$122,539] and dominated across all other strategies (-US36,462 vs. FIT + colonoscopy). When colonoscopy is adopted as a primary screening test, AI colonoscopy dominated conventional colonoscopy (ICER -39,040). These findings showed that AI-assisted colonoscopy can assist identification of early colorectal cancer and save lives.
The findings have been published in prestigious international journals including the Journal of Medical Internet Research (JMIR), Nature Communications, Gastroenterology, and Clinical Gastroenterology and Hepatology.
