Clinical applications of hyperspectral imaging in gastroenterology and hepatology: A systematic review.

Abstract

Background: Hyperspectral imaging (HSI), a high-throughput, three-dimensional technology initially developed for remote sensing, has recently garnered increasing interest in applications in gastroenterology and hepatology. We focus on the aspects of HSI clinical research in gastroenterology and hepatology from a more clinically oriented perspective, discussing its prospective trajectories in this field and delving into the current challenges encountered with HSI.

Methods: A comprehensive systematic search was carried out in PubMed, Embase and Web of Science, encompassing studies published up to June 30, 2024. Following Preferred Reporting Items for Systematic reviews and Meta-Analyses (PRISMA) 2020 guidelines, a systematic review was then conducted to evaluate the identified studies. Two independent reviewers screened the retrieved articles based on pre-defined inclusion and exclusion criteria, followed by an independent data extraction process to ensure accuracy and reliability.

Results: Total 63 studies met the inclusion criteria. Among the 63 included studies, HSI demonstrated significant potential in cancer diagnosis (65.1%), intra-operative monitoring (33.3%) and H. pylori detection (1.6%), with most studies integrating artificial intelligence (AI) for enhanced data analysis. Most studies integrated HSI with AI, enabling efficient analysis and processing of complex hyperspectral data. Additionally, advancements in medical devices, including hyperspectral endoscopes and hyperspectral microscopes, have enhanced the practicality and clinical utility of HSI in gastroenterology and hepatology.

Discussion: HSI shows promising performance and significant potential in gastroenterology and hepatology by enhancing the detection of gastrointestinal cancers, improving intra-operative decision-making through real-time tissue characterization, and facilitating tissue assessment during surgery. HSI offers non-invasive, real-time tissue analysis, but its clinical adoption is hindered by high costs, technical complexity and lack of standardized protocols.