Retinal Vascular Features and Risk of Colorectal Cancer Incidence and Mortality: A Population-Based Cohort Study

Area of Science:

• Ophthalmology
• Oncology
• Medical Imaging
• Artificial Intelligence

Background:

• Current colorectal cancer (CRC) screening methods face limitations due to invasiveness and suboptimal sensitivity.
• The retinal microvasculature, offering a noninvasive view of systemic vascular health, may indicate CRC-related angiogenesis and inflammation.

Purpose of the Study:

• To investigate the association between artificial intelligence (AI)-derived retinal vascular features (RVFs) and colorectal cancer (CRC) risk and prognosis.
• To explore the potential of retinal imaging as a noninvasive tool for CRC risk stratification.

Main Methods:

• Utilized UK Biobank participants with baseline fundus imaging.
• Quantified multidimensional RVFs (tortuosity, branching angle, caliber, density, complexity) using the Retina-based Microvascular Health Assessment System (RMHAS).
• Employed univariable and multivariable Cox regression models to assess associations of RVFs with incident CRC and post-CRC mortality.

Main Results:

• Over 11.32 years of follow-up in 55,798 participants, 652 incident CRC cases and 171 post-CRC deaths were recorded.
• Four RVFs were significantly associated with CRC risk, including increased arterial tortuosity and wider venular minimum branching angle.
• Specific RVF patterns differed between colon and rectal cancers, and certain RVFs correlated with CRC mortality risk.

Conclusions:

• AI-derived RVFs are independently linked to CRC risk, its subtypes, and mortality.
• Retinal microvascular changes may reflect systemic processes involved in carcinogenesis.
• Retinal imaging presents a promising noninvasive approach for assessing CRC risk.