Measurement of smaller colon polyp in CT colonography images using morphological image processing

Area of Science:

• Medical Imaging
• Image Processing
• Computational Anatomy

Background:

• Accurate measurement of colon polyps in computed tomography colonography (CTC) is crucial for diagnosis and treatment.
• Automated measurement of smaller polyps (<5 mm) presents a significant challenge in CTC analysis.

Purpose of the Study:

• To enhance the sensitivity and specificity of smaller polyp measurement in CTC.
• To develop and validate an automated image processing technique for accurate polyp size and shape determination.

Main Methods:

• A domain knowledge-based approach was implemented, combining hybrid colon segmentation and morphological image processing.
• A decision-making system utilizing a priori knowledge was employed for delineating smaller polyps.
• The method was applied to 45 CTC datasets and validated using 2D Multiplanar Reconstruction (MPR) and 3D views.

Main Results:

• The developed method accurately measured smaller polyps, including those less than 5 mm, in addition to the 6-9 mm range.
• Quantitative and qualitative validation confirmed the accuracy of the measurements against ground truth.
• The processing time for a dataset of 500 CTC images was approximately [Formula: see text] minutes on a high-performance computer with parallel processing.

Conclusions:

• The domain-based approach incorporating morphological image processing yields accurate measurements of smaller colon polyps.
• Accurate polyp measurement facilitates improved clinical decision-making in CTC.
• The method demonstrates acceptable qualitative and quantitative results compared to ground truth.