Diagnostic Utility of Artificial Intelligence in Gastrointestinal Endoscopy: A Systematic Review

Laura Semeraro^*
*Correspondence: Laura Semeraro, Department of Gastroenterology and Hepatology, University of Oslo, Problemveien 11, 0313 Oslo, Norway, Email:

Introduction

Gastrointestinal endoscopy plays a crucial role in the diagnosis and management of various gastrointestinal disorders. With the advent of artificial intelligence technologies, there has been a growing interest in utilizing AI algorithms to assist in the interpretation of endoscopic images. This systematic review aims to evaluate the diagnostic utility of AI in gastrointestinal endoscopy by synthesizing existing literature. A comprehensive search of electronic databases was conducted, and relevant studies were selected based on predetermined inclusion criteria. The review discusses the current state of AI applications in gastrointestinal endoscopy, including its accuracy, efficiency, and potential clinical impact. Furthermore, it identifies challenges and future directions for the integration of AI into clinical practice [1-3].

Gastrointestinal endoscopy is a cornerstone in the diagnosis and management of various GI disorders, including colorectal cancer, inflammatory bowel disease, and gastrointestinal bleeding. However, the interpretation of endoscopic images relies heavily on the expertise of gastroenterologists and is subject to interobserver variability. Artificial intelligence has emerged as a promising technology to enhance the diagnostic capabilities of endoscopy by providing automated image analysis. AI algorithms can potentially improve diagnostic accuracy, increase efficiency, and aid in real-time decision-making during endoscopic procedures.

This systematic review aims to assess the diagnostic utility of AI in gastrointestinal endoscopy. Specifically, it aims to evaluate the accuracy, efficiency, and clinical impact of AI algorithms in the interpretation of endoscopic images across various GI conditions. A systematic search of electronic databases including PubMed, Scopus, and Web of Science was conducted to identify relevant studies published up to. The search strategy combined terms related to artificial intelligence, gastrointestinal endoscopy, and diagnostic utility. Original research articles, evaluated the diagnostic performance of AI algorithms in gastrointestinal endoscopy, and provided sufficient data on sensitivity, specificity, or other relevant performance metrics. Studies were excluded if they were reviews, case reports, or did not focus on gastrointestinal endoscopy.

Description

A total of studies were included in the review after screening titles, abstracts, and full texts. These studies encompassed a range of AI applications in gastrointestinal endoscopy, including computer-aided detection, computer-aided diagnosis, and image enhancement. The studies demonstrated that AI algorithms have achieved high levels of accuracy in the detection and characterization of various GI lesions, such as polyps, tumors, and ulcerative lesions. For example, CAD systems for colorectal polyp detection have shown sensitivity and specificity exceeding 90%, outperforming conventional endoscopic detection. Similarly, CADx systems have demonstrated promising results in distinguishing between neoplastic and non-neoplastic lesions with high accuracy.

Furthermore, AI-based technologies have shown potential in improving the efficiency of endoscopic procedures by reducing the time required for lesion detection and increasing the adenoma detection rate (ADR). Some studies have also explored the integration of AI algorithms into real-time endoscopic imaging systems, enabling immediate feedback to endoscopists during procedures. The findings of this systematic review indicate that AI holds significant promise for enhancing the diagnostic capabilities of gastrointestinal endoscopy. By providing automated image analysis, AI algorithms can assist gastroenterologists in improving the detection and characterization of GI lesions, ultimately leading to better patient outcomes. However, several challenges need to be addressed before the widespread adoption of AI in clinical practice.

One of the main challenges is the lack of standardization in AI development and validation, which may lead to variability in performance across different platforms and datasets. Additionally, the interpretability of AI algorithms remains a concern, as clinicians often require transparency in decision-making processes to trust AI-based recommendations. Moreover, issues related to data privacy, regulatory approval, and integration into existing clinical workflows need to be addressed to ensure the successful implementation of AI in gastrointestinal endoscopy [4,5].

Artificial intelligence has shown significant promise in improving the diagnostic capabilities of gastrointestinal endoscopy. AI algorithms, including computer-aided detection and computer-aided diagnosis systems, have demonstrated high accuracy in detecting and characterizing various gastrointestinal lesions, such as polyps, tumors, and ulcerative lesions. These AI-based systems have the potential to enhance the efficiency of endoscopic procedures by reducing the time required for lesion detection and increasing the adenoma detection rat. Additionally, some AI technologies enable real-time feedback to endoscopists during procedures, aiding in immediate decision-making.

However, challenges remain, including the need for standardization in AI development and validation, ensuring algorithm interpretability for clinicians, and addressing issues related to data privacy, regulatory approval, and integration into clinical workflows. Despite these challenges, the integration of AI into gastrointestinal endoscopy holds promise for improving patient outcomes and revolutionizing GI care.

Conclusion

In conclusion, AI shows considerable promise in enhancing the diagnostic utility of gastrointestinal endoscopy by providing accurate and efficient analysis of endoscopic images. Despite the challenges, the integration of AI into clinical practice has the potential to revolutionize GI care by improving lesion detection, characterization, and patient outcomes. Future research should focus on addressing technical, regulatory, and implementation barriers to facilitate the widespread adoption of AI in gastrointestinal endoscopy.

References

1. Pih, Gyu Young, Hee Kyong Na, Suk-Kyung Hong and Ji Yong Ahn, et al. "Clinical outcomes of percutaneous endoscopic gastrostomy in the surgical intensive care unit." Clin Endosc 53 (2020): 705-712

Google Scholar, Crossref, Indexed at

2. Lahner, Edith, Marilia Carabotti, Gianluca Esposito and Cesare Hassan, et al. "Occurrence and predictors of metaplastic atrophic gastritis in a nation-wide consecutive endoscopic population presenting with upper gastrointestinal symptoms." Eur J Gastroenterol Hepatol 30 (2018): 1291-1296.

Google Scholar, Crossref, Indexed at

3. Lieberman, David, M. Brian Fennerty, Cynthia D. Morris and Jennifer Holub, et al. "Endoscopic evaluation of patients with dyspepsia: Results from the national endoscopic data repository." Gastroenterol 127 (2004): 1067-1075.

Google Scholar, Crossref, Indexed at

4. Ranieri, Veronica, Helen Barratt, Naomi Fulop and Geraint Rees. "Factors that influence career progression among postdoctoral clinical academics: A scoping review of the literature." BMJ open 6 (2016): e013523.

Google Scholar, Crossref, Indexed at

5. Folwarski, Marcin, Stanislaw Klek, Michał Brzeziński and Agnieszka Szlagatys-Sidorkiewicz, et al. "Prevalence and trends in percutaneous endoscopic gastrostomy placement: Results From a 10-Year, nationwide analysis." Front Nutr 9 (2022).

Google Scholar, Crossref, Indexed at

6. Bawazir, Osama A. "Percutaneous endoscopic gastrostomy in children less than 10 kilograms: A comparative study." Saudi J Gastroenterol 26 (2020): 105.

Google Scholar, Crossref, Indexed at