PubTrend: General Overview of Artificial Intelligence for Colorectal cancer diagnosis from 2010-2022

TL;DR

This paper reviews AI applications in colorectal cancer diagnosis from 2010 to 2022, highlighting key trends, methodologies, and limitations through analysis of highly cited literature and automated tools.

Contribution

It provides a comprehensive overview of AI advancements in CRC diagnosis, utilizing PubTrends to identify influential papers, keywords, and methodological trends over a 12-year period.

Findings

Most cited papers focus on image analysis and polyps detection.

AI enhances diagnostic accuracy but requires active medical team involvement.

Limitations of colonoscopy for therapeutic use are highlighted.

Abstract

Colorectal cancer (CRC) is among the most prevalent cancers in the world. Due to numerous scholarly papers and broad enquiries about specific use cases for artificial intelligence (AI) in colorectal cancer, researchers find it challenging to explore relevant papers on the current knowledge, comprehensive knowledge, and past methodologies in the literature review. This review extracts recent AI technology advances for diagnosing colorectal cancer from January 2010 to March 2022. PubTrends was used to identify and automate the intellectual structure and comparable papers on the use of AI in colorectal cancer diagnosis using the most cited papers, keywords, and similar papers. Papers with quantitative results were represented with a tabular summary, and other paper contributions were in a sentence summary. Twenty-four (24) out of the forty-nine (49) top-cited papers were quantitative results, with one (1) outlier about lung cancer comprehensive screening. The most frequently used words were: "polyps," "detected", "image," and "colonoscopy." In addition, 83 per cent of the terms frequently used shortly before 2022 were image, polyps, detected, colonoscopy, and learning. In addition, 16 per cent are preparation, variant, classification, sample, and surgery. The review showcases 49 of the 50 most cited papers, their notable contributions, objectives, specific AI methods, results, conclusions, and further recommendations. These papers highlight the limitations of colonoscopy for therapeutic use. The review concluded that despite the enormous benefits of using artificial intelligence, from improving diagnosis, the medical AI programmer still needs to be actively involved in the diagnosis team for effective results in CRC diagnosis.