3D Histopathology-a Lung Tissue Segmentation Workflow for Microfocus X-ray-Computed Tomography Scans

Area of Science:

• Pulmonary Medicine
• Medical Imaging
• Computational Biology

Background:

• Current lung histopathology relies on 2D tissue sections, limiting structural analysis.
• Microfocus X-ray-computed tomography (micro-CT) offers high-resolution 3D soft tissue imaging (2-10 μm) compatible with diagnostic workflows.
• Manual segmentation of 3D datasets for airway and blood vessel networks is time-consuming and challenging.

Purpose of the Study:

• To develop a semi-automatic workflow for segmenting 3D micro-CT lung images.
• To create a user-friendly ImageJ plugin (LungJ) integrating key segmentation steps.
• To enable more comprehensive 3D quantitative analysis of lung structures.

Main Methods:

• Utilized and extended open-source ImageJ software for image segmentation.
• Developed a modular workflow allowing for independent optimization of algorithms.
• Integrated workflow steps into a new ImageJ plugin named LungJ.
• Applied the workflow to segment tubular networks in human lung samples.

Main Results:

• Demonstrated an improved, modular workflow for semi-automatic 3D lung image segmentation.
• Achieved faster segmentation through incremental and independent algorithm optimization.
• Successfully represented 3D tubular networks of airways and blood vessels in human lung samples.
• Validated the potential for novel quantitative measurements from 3D tissue analysis.

Conclusions:

• The LungJ plugin and associated workflow facilitate efficient and accurate segmentation of 3D lung micro-CT data.
• This approach enhances the understanding of lung tissue structure, function, and interrelationships.
• 3D analysis offers a more complete picture of heterogeneous lung samples and enables new quantitative insights beyond 2D extrapolations.