Advancements and limitations of image-enhanced endoscopy in colorectal lesion diagnosis and treatment selection: A narrative review

Abstract

Colorectal cancer (CRC) is a leading cause of cancer-related mortality, highlighting the need for early detection and accurate lesion characterization. Traditional white-light imaging has limitations in detecting lesions, particularly those with flat morphology or minimal color contrast with the surrounding mucosa. It also struggles to distinguish neoplastic from non-neoplastic lesions. These limitations led to the development of image-enhanced endoscopy (IEE). Image-enhanced endoscopy modalities such as narrow-band imaging, blue laser imaging, linked color imaging, and texture and color enhancement imaging enhance mucosal surface and vascular pattern visualization, thereby improving lesion detection and characterization. In contrast, red dichromatic imaging is primarily designed to enhance the visibility of deep blood vessels, making it particularly useful during therapeutic endoscopies, such as identifying bleeding sources and monitoring post-treatment hemostasis. Although IEE enhances lesion detection and characterization, it remains limited in assessing submucosal invasion depth, which is a key factor in treatment decisions. Endoscopic submucosal dissection requires accurate prediction of invasion depth; however, IEE mainly reflects superficial features. Endoscopic ultrasound and artificial intelligence-assisted diagnostics have emerged as complementary techniques for improving depth assessment and lesion classification. Additionally, IEE plays a critical role in detecting ulcerative colitis-associated neoplasia (UCAN), which often presents with a flat morphology and indistinct borders. High-definition chromoendoscopy and IEE modalities enhance detection; however, inflammation-related changes limit diagnostic accuracy. Artificial intelligence and molecular biomarkers may improve UCAN diagnosis. This review examines the role of IEE in lesion detection and treatment selection, its limitations, and complementary techniques such as endoscopic ultrasound and artificial intelligence. We also explored pit pattern diagnosis using crystal violet staining and discussed emerging strategies to refine colorectal cancer screening and management.

Keywords: artificial intelligence; colorectal cancer; image‐enhanced endoscopy; invasion depth assessment; ulcerative colitis‐associated neoplasia.

FIGURE 1

Differences in lesion visibility among white‐light imaging, linked color imaging, and blue laser imaging. Linked color imaging enhances redness in tumor areas, improving lesion recognition and providing greater overall brightness, which increases the amount of information visible in the field of view. In contrast, blue laser imaging emphasizes vascular and surface patterns, making it particularly useful for identifying lesion characteristics.

FIGURE 2

Comparison of white‐light imaging, texture and color enhancement imaging, and narrow‐band imaging, demonstrating how each modality enhanced lesion visibility

FIGURE 3

Endoscopic and endoscopic ultrasound assessment of an elevated colorectal cancer lesion. Although surface evaluation alone was insufficient to confirm deep invasion due to the absence of exposed highly atypical areas, endoscopic ultrasound assessment confirmed muscularis propria invasion, leading to a decision for surgical treatment. Pathological examination confirmed a pT2 depth of invasion.