Dice-XMBD

Dice-XMBD performs deep learning-based single-cell segmentation of imaging mass cytometry (IMC) data to generate accurate single-cell masks across nuclear, membrane, and cytoplasmic markers for downstream analysis of tumor microenvironments.

Key Features:

• Deep learning-based segmentation: Uses a deep learning approach specifically tailored for cell segmentation in tissue multiplexed imaging data.
• Marker-agnostic processing: Processes IMC images with varying channel configurations without requiring marker-specific modifications.
• Multiplex IMC compatibility: Handles high-resolution IMC data with large numbers of channels, including datasets with up to 40 markers.
• Subcellular marker support: Produces segmentation compatible with nuclear, membrane, and cytoplasmic markers.
• Precise single-cell mask generation: Generates robust and accurate single-cell masks across different marker types for downstream analysis.

Scientific Applications:

• Tumor microenvironment analysis: Enables quantification of cell composition and spatial interactions within tumor microenvironments from IMC data.
• Single-cell phenotyping in multiplexed tissues: Facilitates extraction of marker signals at subcellular resolution for single-cell phenotype assignment.
• Basic and clinical research: Supports both basic research and clinical applications that rely on high-dimensional tissue imaging.

Methodology:

Applies a deep learning-based cell segmentation model to imaging mass cytometry (IMC) multiplexed images and outputs marker-agnostic single-cell masks.