[Image processing of early gastric cancer cases]

Area of Science:

• Medical Imaging
• Oncology
• Computer Science

Background:

• Early detection of stomach cancer is crucial for patient survival.
• Traditional methods for detecting small gastric lesions can be challenging.
• Computer image processing offers potential for enhancing diagnostic accuracy.

Purpose of the Study:

• To evaluate the effectiveness of computer image processing in enhancing gastric lesions for improved stomach cancer detection.
• To assess specific image processing techniques for visualizing early gastric cancer.
• To determine the contribution of image enhancement to the detection of small gastric tumors.

Main Methods:

• Digitization of 25 surgically and pathologically confirmed early gastric cancer cases.
• Application of grayscale transformation, Sobel operator for edge enhancement, and unsharp masking for high-pass filtering.
• Comparative analysis of image processing techniques for lesion visualization.

Main Results:

• Grayscale transformation improved overall image quality for lesion detection.
• The Sobel operator effectively enhanced lesion margins.
• Unsharp masking superiorly visualized mucosal texture patterns.
• Eight out of ten small lesions (<2.0 cm) were successfully demonstrated.
• Tumor surface patterns and margins of larger lesions (>2.0 cm) were clearly visualized.

Conclusions:

• Computer image processing significantly contributes to the detection of small early gastric cancer lesions.
• Image enhancement techniques improve the visualization of pathological gastric lesions.
• Further refinement may be needed for optimal detection of lesions in specific anatomical locations like the antrum.