Automated teniae coli detection and identification on computed tomographic colonography

Area of Science:

• Medical Imaging
• Gastroenterology
• Computer-Aided Diagnosis

Background:

• Computed tomographic colonography (CTC) is a minimally invasive screening tool for colonic polyps and cancer.
• Teniae coli, three longitudinal smooth muscle bands on the colon surface, are significant anatomical landmarks.
• Accurate identification of teniae coli can aid in colon surface navigation and analysis during CTC.

Purpose of the Study:

• To develop and evaluate an automated method for detecting teniae coli in computed tomographic colonography (CTC).
• To utilize teniae coli as anatomical landmarks for improved colon analysis in CTC.
• To assess the accuracy and potential clinical utility of the proposed automated detection system.

Main Methods:

• CTC slices were segmented and reconstructed into a 3D colon surface, then unfolded using a reversible projection technique.
• A 2D height map was generated, and teniae coli were detected using Gabor filter banks to analyze haustral fold features.
• Teniae coli were identified based on fold center detection, path creation, and spatial relationships with anatomical markers like the ileocecal valve.

Main Results:

• The automated method was tested on 47 CTC cases (37 patients).
• The system achieved an average normalized root mean square error (RMSE) of 4.87% (± 2.93%).
• Performance metrics indicate high accuracy in detecting and identifying teniae coli.

Conclusions:

• The proposed fully-automated teniae coli detection and identification method is accurate.
• This technique demonstrates significant promise for future clinical applications in colon cancer screening.
• Automated teniae coli detection can enhance the interpretation of CTC images.