Separation of left and right lungs using 3-dimensional information of sequential computed tomography images and a...

Area of Science:

• Medical Imaging
• Radiology
• Computer-Aided Diagnosis

Background:

• Accurate separation of left and right lungs is crucial for analyzing thoracic computed tomography (CT) images.
• Complex anatomical variations and severe connections can challenge automated lung separation techniques.

Purpose of the Study:

• To develop and validate a novel computerized scheme for precise and reliable separation of left and right lungs in CT examinations.
• To improve the robustness and efficiency of lung separation algorithms.

Main Methods:

• A new method utilizing sequential CT data and a guided dynamic programming algorithm was developed.
• The algorithm featured adaptively selected start/end points to manage severe and multiple lung connections.
• Testing was performed on an independent dataset of 4034 CT images.

Main Results:

• The scheme successfully identified and separated all 827 connections across the tested CT images.
• Multiple connections were accurately separated irrespective of their location.
• The guided dynamic programming approach reduced computation time to approximately 4.6% of traditional methods.
• The separation boundary integrity was maintained, preventing permeation into normal lung tissue.

Conclusions:

• The proposed method demonstrates robust and accurate disconnection of all lung connections.
• The guided dynamic programming algorithm effectively eliminates redundant processing, enhancing computational efficiency.