A Pulmonary Vascular Extraction Algorithm from Chest CT/CTA Images.

Area of Science:

• Medical Imaging
• Computer-Aided Diagnosis
• Pulmonary Medicine

Background:

• Accurate segmentation of pulmonary vessels in CT/CTA images is crucial for clinical diagnosis and treatment planning.
• Existing segmentation methods face challenges due to the complexity of CT/CTA image data, leading to limitations in accuracy and oversegmentation.
• The need for improved pulmonary vessel segmentation techniques is driven by their importance in assessing various cardiopulmonary conditions.

Purpose of the Study:

• To develop and evaluate a novel two-step method for enhanced pulmonary vessel segmentation in CT/CTA images.
• To improve the accuracy and reduce oversegmentation in pulmonary vessel segmentation by incorporating lung parenchyma extraction.
• To validate the proposed method's performance against conventional techniques using clinical CT/CTA data.

Main Methods:

• A two-step segmentation approach was implemented, beginning with lung parenchyma extraction using the Unet++ algorithm.
• Pulmonary vessels within the extracted lung parenchyma were subsequently segmented using the nnUnet algorithm.
• An "AND" operation combined the original image with lung parenchyma segmentation to isolate vessels, minimizing external tissue interference.

Main Results:

• The proposed method achieved high segmentation accuracy, with CT images reaching 85.1% and CTA images reaching 87.7%.
• The inclusion of lung parenchyma segmentation significantly reduced oversegmentation rates compared to methods without this step.
• Comparative analysis demonstrated superior accuracy of the proposed algorithm over conventional pulmonary vessel segmentation methods.

Conclusions:

• The developed two-step method effectively segments pulmonary vessels in CT/CTA images, offering improved accuracy and reduced oversegmentation.
• This technique enhances the reliability of pulmonary vessel segmentation, aiding physicians in better patient assessment and treatment decisions.
• The findings suggest that incorporating lung parenchyma segmentation is a valuable strategy for improving the performance of pulmonary vessel analysis in medical imaging.