Whole-Section Tumor Micro-Architecture Analysis by a Two-Dimensional Phasor-Based Approach Applied to...

Area of Science:

• Biomedical Optics
• Histopathology
• Medical Image Analysis

Background:

• Second Harmonic Generation (SHG) microscopy offers label-free tissue imaging, revealing morphology and collagen structure crucial for disease detection.
• Increasing data volumes necessitate automated image analysis tools for reliable, user-independent histopathological diagnosis.
• Understanding collagen microarchitecture is key to identifying pathological changes in tissues.

Purpose of the Study:

• To develop and validate an automated image analysis method using phasor parameters for collagen microarchitecture characterization in histopathology.
• To assess the efficacy of the proposed method in discriminating between different collagen organizations and identifying tumor boundaries.
• To explore the potential of phasor-based SHG analysis as a supportive tool for standard histopathology.

Main Methods:

• Utilized a 2-dimensional phasor-based approach (μMAPPS) with polarization-dependent SHG signals.
• Employed a clustering algorithm to automatically recover collagen microarchitectures from phasor parameters.
• Analyzed mesoscopic collagen fibril orientation and anisotropy in mouse cancer xenograft sections.

Main Results:

• Successfully recovered distinct collagen microarchitectures within the extracellular matrix.
• Quantified local spatial heterogeneity of collagen fibrils to discriminate tissue organizations.
• "Fibril entropy" parameter effectively highlighted tumor edges by assessing tissue order.
• Demonstrated identification of tumor areas versus surrounding skin tissue based on collagen organization.

Conclusions:

• The μMAPPS method, leveraging phasor parameters, enables automated characterization of collagen microarchitecture in histopathology.
• The "fibril entropy" parameter shows significant potential for automatic segmentation of tumor boundaries.
• This approach, combined with morphological information, can serve as a valuable adjunct to conventional histopathology for disease diagnosis.