Project Code: PN-III-P1-1.1-TE-2019-1756
Contract Number: TE28/2020
Project Title: Integration of pixel-wise and whole image classification of second harmonic generation microscopy datasets for thyroid pathology
Duration: 15.09.2020 – 14.09.2022
Grant value: 431 900 Lei (~ 90 000 EURO)
Abstract: SHGThyPath proposes to bring the quantitative analysis of second harmonic generation (SHG) images beyond the current state-of-the-art by integrating pixel-level and whole-image level quantitative analysis approaches. More precisely, the project scope is to create a solution which would integrate different quantitative analysis methods for collagen fibers in order classify the experimental data obtained by using intensity-based SHG and polarization-resolved SHG (PSHG) microscopy techniques. We will consider currently available pixel-wise metrics like the second order susceptibility χ(2) tensor element ratios, SHG circular dichroism ratio, the anisotropy factor and whole image analysis parameters like first order statistics, second-order statistic (the gray-level co-occurrence matrix), fractal analysis or the Helmholtz analysis. Using this integrated approach one can obtain both ultrastructural information regarding the collagen fibers and large area collagen distribution information in tissue samples. To date an integrated solution has been lacking since the researchers involved in this field use either of the two approaches and even more, different parameters for each approach. The goal is to create a methodology which will be tested on mouse tail tendon samples and applied for the differentiation of benign and malignant tissue regions on different pathologies of the thyroid where the current gold standard in histology, the hematoxylin and eosin staining of thin tissue section has its limitations: differentiation of thyroid follicular adenoma/carcinoma and of noninvasive follicular thyroid neoplasm with papillary-like nuclear features (NIFTP) vs. encapsulated papillary carcinoma.
Project objective: The development of a methodology for quantitative analysis of images obtained on tissue sections by integrating pixel-level and whole-image level quantitative metrics.
Funding Agency: Executive Unit for Financing Higher Education, Research, Development and Innovation (UEFISCDI)