ATMAD: robust image analysis for Automatic Tissue MicroArray De-arraying.

Area of Science:

• Histology and pathology
• Biomedical imaging
• Bioinformatics

Background:

• Tissue Microarray (TMA) technology enables multiplex analysis of numerous tissue samples under standardized conditions.
• Manufacturing process distortions can misalign samples from their design grid, leading to inaccurate assay results.
• A robust de-arraying method is crucial for accurate sample localization and matching in TMA analysis.

Purpose of the Study:

• To develop an Automatic, fast and robust TMA De-arraying (ATMAD) approach.
• To accurately localize and match tissue samples with their design grid positions despite manufacturing distortions.
• To provide quantitative information for assessing TMA manufacturing quality.

Main Methods:

• Utilizes wavelet transform and adaptive thresholding for initial tissue sample detection.
• Employs a parametric shape model for ellipse-shaped object segmentation and filtering.
• Applies thin-plate spline interpolation to estimate grid deformation for accurate sample matching.
• Incorporates a secondary segmentation step to recover initially rejected samples based on estimated deformation.

Main Results:

• The ATMAD approach successfully localizes and matches TMA samples with their design grid.
• The method is effective for images acquired with brightfield and fluorescence microscopes.
• Handles challenging imaging conditions including high dynamic range, poor signal-to-noise ratio, and complex backgrounds.
• Accurately corrects non-linear deformations in the TMA grid.

Conclusions:

• A novel de-arraying approach combining wavelet-based detection, active contour segmentation, and thin-plate spline interpolation has been developed.
• This method robustly handles various image complexities and TMA grid deformations.
• The deformation estimation provides valuable quantitative data for TMA manufacturing quality assessment.