21 May 2022

The method allows to conduct non-invasive and more accurate cancer diagnostics based on quantitative image analysis of tissue texture at a pixel level.

Researchers at Immanuel Kant Baltic Federal University(IKBFU) are developing a method to diagnose cancer without taking a biopsy. Researchers at Kant University are developing a "virtual biopsy" method of cancer diagnosis which will allow them to diagnose cancer in a non-invasive way, based on computer analysis of tissue characteristics at a pixel level, inaccessible to the human eye. This was reported on Wednesday by the IKBFU press service.  

The scientists have proposed an original technique of analyzing CT and MRI images using artificial neural networks and the anisotropy and radiomics indices of digital images to diagnose various cancers. 

"The "virtual biopsy" method allows to conduct non-invasive and more accurate cancer diagnostics without taking a biopsy, based on quantitative image analysis of tissue texture at a pixel level. Two new approaches are used: anisotropy - determination of changes in tissue properties depending on the direction of their measurement; radiomics - extraction of quantitative features from digital images using computer algorithms," Andrey Litvin, professor of the IKBFU Medical Institute and the scientific director of the project. 

Using such an analysis, a computer-based decision support system will be created for early non-invasive diagnosis of various tumor diseases. "The system will be a cloud-based service that can be used by all interested public and private medical centers and various medical research organizations. The system will also function as an installation program on a disk if it is impossible to connect to the cloud service via the Internet," explained project participant Dmitry Burkin, a graduate student at the Institute of Physics, Mathematics and Information Technology at IKBFU.

 The developers of the "virtual biopsy" method won a grant in the field of artificial intelligence in the Start-III-1 contest held by the Foundation for Assistance to Small Innovative Enterprises in Science and Technology. The size of the grant for the IKBFU scientists' development stood at 4 million rubles.

Researchers plan to complete the work to create the innovative system by the spring of 2023. The authors of the "virtual biopsy" method are Dmitry Burkin, a graduate student at the Institute of Physics, Mathematics and Information Technology; Andrey Litvin, MD, professor of the IKBFU Medical Institute and the scientific director of the project; Fyodor Paramzin, an oncologist, and Zakhar Ponimash, a software engineer.