Domain-Based Analysis of Colon Polyp in CT Colonography Using Image-Processing Techniques

Area of Science:

• Medical Imaging
• Gastroenterology
• Computer-Aided Diagnosis

Background:

• CT Colonography (CTC) is a key tool for colorectal polyp detection.
• Improving polyp detection accuracy in CTC remains a significant clinical challenge.
• Existing CTC solutions can be enhanced through advanced image processing techniques.

Purpose of the Study:

• To enhance polyp detection accuracy in CT Colonography (CTC).
• To develop effective colon segmentation and virtual cleansing methods.
• To accurately measure smaller polyps using automated techniques.

Main Methods:

• Implemented boundary-based semi-automatic colon segmentation considering colon distension.
• Utilized an adaptive multistep method for virtual cleansing based on Hounsfield Units.
• Developed an automated polyp measurement method using skeletonization techniques.

Main Results:

• Achieved 95.826±0.6854% accuracy in colon segmentation with preserved structures.
• Demonstrated high performance in virtual cleansing, correcting pseudo-enhancements and removing air-contrast layers.
• Attained 90% True Positive Rate (TPR), 82.3% True Negative Rate (TNR), and 88.31% overall accuracy in polyp detection and measurement.

Conclusions:

• Developed an automated system for polyp measurement from colon segmentation to improve CTC.
• The domain-based approach to polyp analysis yielded significant improvements.
• Created a prototype software for low-cost polyp diagnosis.