Computer analysis of upper gastrointestinal endoscope images

Area of Science:

• Gastroenterology
• Medical Imaging
• Computational Pathology

Background:

• Gastro-oesophageal reflux disease (GORD) affects up to 40% of Western adults.
• Current classification systems for oesophagitis, a key GORD determinant, are subjective.
• Reducing inter-observer variability in upper GI endoscopy is crucial for accurate diagnosis.

Purpose of the Study:

• To develop a computer image analysis system for objective interpretation of upper GI endoscopic examinations.
• To reduce inter-observer variability in the assessment of reflux oesophagitis.
• To explore the feasibility of quantifying endoscopic images for GORD diagnosis.

Main Methods:

• Digital video recordings were obtained from patients with clinical evidence of reflux oesophagitis.
• Cross-sectional profiles of hue and saturation data were analyzed from endoscopic images.
• Images from seven patients with Los Angeles (LA) grade B or C oesophagitis were studied.

Main Results:

• Significant changes in hue (p = 0.01) and saturation (p = 0.001) were observed in endoscopic images.
• The analysis demonstrated quantifiable differences in image data related to oesophagitis severity.
• These findings suggest objective assessment of upper GI endoscopic images is achievable.

Conclusions:

• Computer-based image analysis is a feasible method for the objective assessment of upper GI endoscopic findings.
• This approach may help standardize the interpretation of reflux oesophagitis, improving diagnostic accuracy.
• Quantification of endoscopic images offers a potential solution to the subjectivity in current GORD classification systems.