Multi-scale evaluation of structural changes in murine peritonitis models
ORAL
Abstract
Multi-scale imaging compares data from different scales of size to obtain insights into relationships between macroscopic tissue changes and underlying tissue architecture. Here, we present a multi-scale analysis of induced murine peritonitis models using optical hyperspectral imaging (HSI).
We studied the changes in the organization of cell nuclei in histology using HSI, which enabled the decoupling of individual hematoxylin and eosin stains. Hematoxylin maps were evaluated for structural changes in the Fourier space. Two metrics, namely the total image energy (TIE) and energy of high spatial frequency components (HSFE), exhibited high statistical significance in discriminating between healthy and diseased subjects (p=0.02) with an almost twofold increase in the values. The TIE and HSFE metrics were compared to light scattering macroscopic metrics, achieving a moderate correlation of 0.55 and 0.51 for TIE and HSFE, respectively. Furthermore, TIE and HSFE correlated strongly (linear correlation greater than -0.7) with metrics of cellular architecture in Fourier space.
The results present a successful application of multi-scale analysis that will pave the way in the future for tailored longitudinal monitoring of pathologies.
We studied the changes in the organization of cell nuclei in histology using HSI, which enabled the decoupling of individual hematoxylin and eosin stains. Hematoxylin maps were evaluated for structural changes in the Fourier space. Two metrics, namely the total image energy (TIE) and energy of high spatial frequency components (HSFE), exhibited high statistical significance in discriminating between healthy and diseased subjects (p=0.02) with an almost twofold increase in the values. The TIE and HSFE metrics were compared to light scattering macroscopic metrics, achieving a moderate correlation of 0.55 and 0.51 for TIE and HSFE, respectively. Furthermore, TIE and HSFE correlated strongly (linear correlation greater than -0.7) with metrics of cellular architecture in Fourier space.
The results present a successful application of multi-scale analysis that will pave the way in the future for tailored longitudinal monitoring of pathologies.
*Funding: ARRS P1-0389 and Z1-4384
–
Publication: We are planning a paper that will include these results, among others.
Presenters
-
Jost Stergar
- J. Stefan Institute
- Jožef Stefan Institute