Automated segmentation of cells with IHC membrane staining

Area of Science:

• Biomedical Imaging
• Computational Pathology
• Digital Histopathology

Background:

• Accurate cell membrane segmentation is crucial for quantitative immunohistochemistry (IHC).
• Existing methods struggle with the discontinuous nature of membrane staining in IHC images.
• Cellular membrane visualization is limited to stained tracts, complicating analysis.

Purpose of the Study:

• To develop a fully automated technique for precise membrane segmentation in IHC images.
• To overcome challenges posed by unstained membrane regions in tissue samples.
• To enable accurate cell-by-cell morphological analysis and protein quantification.

Main Methods:

• An automated approach for segmenting stained membrane tracts in IHC images.
• Utilizing nuclear membranes as spatial references to reconstruct unstained membrane regions.
• Developing a cell-by-cell segmentation strategy for detailed analysis.

Main Results:

• The automated method achieves accurate segmentation of cellular membranes in stained tracts.
• Reconstruction of approximate unstained tract locations using nuclear membrane references.
• Demonstrated high accuracy and wide applicability across diverse IHC datasets.

Conclusions:

• The developed automated membrane segmentation technique enhances IHC analysis accuracy.
• Enables per-cell morphological analysis and quantification of membrane proteins.
• Supports critical medical applications including cancer characterization and personalized therapy.