Cancer Genetics and Epigenetics 2024, Vol.12, No.4, 210-222 http://medscipublisher.com/index.php/cge 219 diagnostic accuracy, improving detection rates, and aiding in the prognostication and treatment of colorectal cancer (CRC). AI-based systems have been particularly effective in increasing adenoma detection rates (ADRs) and polyp detection rates (PDRs) during colonoscopy, as evidenced by meta-analysis which showed that AI-assisted colonoscopy significantly outperformed traditional methods in detecting small non-advanced adenomas and polyps. Additionally, AI has proven to be highly accurate in histology prediction and detection of colorectal polyps, with a pooled sensitivity of 92.3% and specificity of 89.8%. In the realm of pathology, AI has made strides in CRC image analysis, particularly in gland segmentation, tumor classification, and tumor microenvironment characterization. However, the clinical application of these techniques is still limited by the scale and quality of available datasets. AI has also shown potential in reducing the miss rate of colorectal neoplasia, thereby enhancing CRC prevention. Furthermore, AI has been applied in the surgical treatment of CRC, aiding in various intraoperative steps such as phase and action recognition, excision plane navigation, and real-time circulation analysis. Despite these advancements, the use of AI in CRC surgery is still in its nascent stages6. The implications of these findings for colon cancer prediction are profound. The use of AI in multi-modal data fusion can lead to earlier and more accurate detection of colorectal cancer, which is crucial for improving patient outcomes. Enhanced detection rates of adenomas and polyps during colonoscopy can lead to more effective screening programs and potentially reduce the incidence of CRC. AI's ability to accurately predict histology and detect colorectal polyps can streamline the diagnostic process, allowing for quicker and more precise treatment decisions. This is particularly important for the management of diminutive polyps, which are often challenging to diagnose accurately. In pathology, AI's capabilities in image analysis can provide more detailed and accurate assessments of tumor characteristics, aiding in the development of personalized treatment plans. However, the need for larger and higher-quality datasets remains a significant barrier to the widespread clinical application of these techniques. The reduction in miss rates of colorectal neoplasia through AI-assisted colonoscopy highlights the potential for AI to mitigate perceptual errors and improve the overall quality of colonoscopy procedures. This can lead to more comprehensive screening and surveillance programs, ultimately enhancing CRC prevention efforts. The integration of AI in multi-modal data fusion for colon cancer prediction represents a significant advancement in the field of oncology. The promising results from various studies underscore the potential of AI to revolutionize the diagnosis, treatment, and management of colorectal cancer. However, several challenges remain. The variability in AI models and study designs, as well as the need for larger and more diverse datasets, are critical issues that need to be addressed. Future research should focus on standardizing AI methodologies and developing robust, high-quality datasets to validate AI models in clinical practice. Moreover, the ethical considerations and regulatory frameworks surrounding the use of AI in healthcare must be carefully examined to ensure patient safety and data privacy. As AI technologies continue to evolve, it is essential to foster collaboration between researchers, clinicians, and policymakers to harness the full potential of AI in improving colon cancer prediction and treatment. While AI has shown remarkable potential in enhancing the prediction and management of colorectal cancer, ongoing research and development are necessary to overcome existing limitations and fully integrate AI into routine clinical practice. The future of AI in colon cancer prediction is promising, with the potential to significantly improve patient outcomes and transform the landscape of cancer care. Acknowledgments The authors thank two anonymous peer reviewers for their suggestions on the first draft of this study. Funding This work was funded by National Natural Science Foundation of China [U20A20376].
RkJQdWJsaXNoZXIy MjQ4ODYzNQ==