Mr. Viktoras Mazeika Profile

Viktoras Mazeika

at Vilnius Univ

SPIE Involvement:

Author

Publications (3)

Proceedings Article | 30 May 2022 Presentation

Investigation of melanoma tissue with nonlinear multimodal polarimetric microscopy using texture analysis and machine learning

Martynas Riauka, Viktoras Mazeika, Mykolas Maciulis, Lukas Kontenis, Edvardas Zurauskas, Mehdi Alizadeh, Kamdin Mirsanaye, Virginijus Barzda

Proceedings Volume PC12144, PC121440S (2022) https://doi.org/10.1117/12.2621489

KEYWORDS: Polarimetry, Tissues, Melanoma, Collagen, Polarization, Microscopy, Second-harmonic generation, Machine learning, Statistical analysis, Microscopes

Read Abstract +

Development and metastasis of cancer are known to change the structure of extracellular matrix (ECM), which affect the tumor's further growth and spread. A substantial part of ECM is comprised of collagen, which is a noncentrosymmetric structure. As a result, it generates second harmonic signals, dependent on the polarization of incoming light. This property of collagen led to the applications of polarization-resolved second-harmonic generation (P-SHG) microscopy in investigating collagen ultrastructure changes in different cancers. In this work, multiphoton absorption fluorescence (MPF), third-harmonic generation (THG) and polarimetric second-harmonic generation (P-SHG) measurements were performed on various types and staging of human melanoma histological sections. Reduced polarimetry techniques, employing linear and circular polarization states, were used to obtain polarimetric SHG parameters of collagen in both normal and cancerous tissues. These parameters provide important information about the structural properties of collagen. The parameter distributions were analyzed using a grey-level co-occurrence matrix (GLCM), which allows to obtain statistical parameters, such as correlation, contrast, entropy, angular second moment and inverse difference moment. Statistical tests were performed on polarimetric and texture analysis data in order to determine whether parameter distribution differences in normal and cancerous tissues are statistically significant. Furthermore, a machine learning classifier algorithm was trained to distinguish normal tissues from cancerous using aforementioned polarimetric and texture parameters as predictors. Firstly, separate training and testing datasets were formed from each sample and classification was carried out for each of them individually and afterwards, a common training dataset was used for all samples. The results suggest that normal and cancerous skin tissues can be distinguished from each other with the help of multimodal nonlinear polarimetric microscopy. Also, depending on the type and stage of melanoma, the differences in some polarimetric and texture parameters are more pronounced, suggesting its possible application in melanoma diagnostics and differentiation.

Proceedings Article | 7 December 2021 Paper

Nonlinear multicontrast microscopy with texture analysis for colon cancer diagnostics

Viktoras Mažeika, Mykolas Mačiulis, Martynas Riauka, Lukas Kontenis, Edvardas Žurauskas, Virginijus Barzda

Proceedings Volume 11919, 1191913 (2021) https://doi.org/10.1117/12.2614800

KEYWORDS: Second-harmonic generation, Microscopy, Statistical analysis, Analytical research, Tissues, Polarimetry, Diagnostics, Colorectal cancer, Colon, Collagen

Read Abstract +

Proceedings Article | 3 December 2021 Paper

Characterization of lentiginous melanoma tissue with multimodal polarimetric harmonic generation microscopy

Martynas Riauka, Viktoras Mazeika, Mykolas Maciulis, Edvardas Zurauskas, Lukas Kontenis, Virginijus Barzda

Proceedings Volume 11921, 1192104 (2021) https://doi.org/10.1117/12.2614313

KEYWORDS: Tissues, Polarimetry, Collagen, Second-harmonic generation, Polarization, Microscopy, Melanoma, Harmonic generation, Polysomnography, Cancer

Read Abstract +

View contact details

UPDATE YOUR PROFILE

Is this your profile? Update it now.

Sign into your SPIE.org account

Don’t have a profile and want one?

Create an account on SPIE.org

Keywords/Phrases

Search In:

Publication Years