Segmentation of juxtapleural pulmonary nodules using a robust surface estimate

Area of Science:

• Medical Imaging
• Radiology
• Computer-Aided Diagnosis

Background:

• Accurate segmentation of solid pulmonary nodules, especially those juxtapleural, is challenging in computed tomography (CT) scans.
• Existing methods like plane-fitting algorithms have limitations in segmenting nodules attached to the chest wall.

Purpose of the Study:

• To develop and evaluate a novel algorithm for segmenting solid pulmonary nodules directly attached to the chest wall.
• To improve the accuracy and success rate of juxtapleural pulmonary nodule segmentation compared to previous methods.

Main Methods:

• Developed an algorithm that estimates the pleural surface to segment nodules from the chest wall.
• Employed a robust approach to identify pleural surface points not overlapping with the nodule.
• Estimated a 3D surface from identified points and evaluated performance on 150 solid juxtapleural nodules.

Main Results:

• The novel algorithm achieved a 98.0% success rate in segmenting juxtapleural pulmonary nodules.
• This represents a significant improvement over the 81.3% success rate of a previously published plane-fitting algorithm.
• Segmentation performance was rated acceptable (scores 3-4) for the majority of cases.

Conclusions:

• The developed algorithm offers a substantial advancement in the automated segmentation of juxtapleural pulmonary nodules.
• This improved segmentation capability is crucial for developing more robust automated nodule measurement methods.
• The algorithm shows high efficacy and potential for clinical application in lung nodule analysis.