Comprehensive Review of the Current State-of-the-Art in AI-Driven Colorectal Cancer Screening

Abstract

The explosive growth of artificial intelligence (AI) marks a pivotal shift in colorectal cancer (CRC) screening, offering an avenue to bolster early detection and heighten the quality of patient care. This thorough review charts the progression of AI and its branches—machine learning (ML) and deep learning (DL) with an emphasis on their application within colonoscopy to aid in recognizing and diagnosing CRC. Presently, AI application in colonoscopy can be divided into two classes as follows: computer-aided detection (CADe) and computer-aided diagnosis (CADx). In randomized controlled trials, CADe has shown favorable outcomes, significantly increasing adenoma detection rate (ADR) and adenoma per colonoscopy (APC), while decreasing adenoma miss rate (AMR). For instance, a meta-analysis by Hassan et al. involving 4,354 participants demonstrated that CADe notably improved ADR (36.6% vs. 25.2%; relative risk [RR]: 1.44, 95% confidence interval [CI]: 1.27–1.62, p < 0.01) and APC (0.58 vs. 0.36; RR: 1.70, 95% CI: 1.53–1.89, p < 0.01). Conversely, Wei et al.’s review of real-world studies with 11,660 patients showed a statistically significant but clinically minimal improvement in ADR with CADe (36.3% vs. 35.8%; RR: 1.13, 95% CI: 1.01–1.28, p = 0.04) and no notable differences in APC highlighting the need for further research to understand the factors affecting these mixed results. AI in colonoscopy is not limited to spotting polyps; it can also aid in estimating polyp size, evaluating the quality of bowel preparation, and appraising CRC risk, streamlining the entire screening process. Nevertheless, the adoption of AI in CRC screening encounters several obstacles. The presence of false positives, concerns over data privacy, inherent biases in training data, and susceptibility to cyber threats are matters that warrant vigilant consideration. Furthermore, there are ethical and regulatory dilemmas regarding AI’s role in health care, issues of transparency, accountability for diagnostic errors, and the potential for reducing physicians’ diagnostic expertise that need to be resolved. Economic barriers, such as the lack of defined reimbursement methods for AI applications in endoscopy, stand as challenges too. Looking forward, the advancement of AI in CRC screening requires deep-seated collaboration among various fields, enhancements in medical training programs, and initiatives to ensure fair access to AI, averting health care disparities.