Tissue Cytometry With Machine Learning in Kidney: From Small Specimens to Big Data.

Area of Science:

• Nephrology
• Computational Biology
• Biomedical Imaging

Background:

• Understanding kidney disease pathogenesis requires detailed cellular and molecular analysis.
• Classifying cells in situ, including injury-induced subtypes, is crucial for identifying therapeutic targets.
• High-resolution 3D imaging preserves tissue architecture, providing spatial context for cellular analysis.

Purpose of the Study:

• To review advancements in analyzing 3D kidney tissue imaging.
• To highlight the synergy of machine learning and cytometry for analyzing high-resolution imaging data.
• To demonstrate the utility of imaging-based cell classification for kidney disease research.

Main Methods:

• Utilizing Volumetric Tissue Exploration and Analysis (VTEA) cytometry for interactive analysis and classification of cells in 3D image volumes.
• Employing deep learning for imaging-based classification of cells in intact tissue using 3D nuclear staining (DAPI).
• Leveraging high-resolution 3D fluorescence imaging to capture cellular and spatial information.

Main Results:

• Demonstrated semiautomated classification of labeled cells in 3D image volumes using VTEA cytometry.
• Established and validated an imaging-based deep learning classification method for cells in intact kidney tissue.
• Generated large-scale data from small tissue specimens, enabling cell classification within a spatial context.

Conclusions:

• Combining machine learning with cytometry on 3D imaging data provides a powerful analytical output for kidney tissue.
• This approach facilitates cell classification in a spatial context, yielding novel insights into kidney disease pathology.
• Advances in imaging and computational methods are transforming the study of kidney disease pathogenesis and therapeutic target identification.