Scientists from the Perm Tech have created a project for a device for automatical histological analysis. The development will make it possible to identify cellular pathologies, including oncopathology, in humans and animals promptly. In the future, the device will be able to produce the actions faster and cheaper than it is performed now by several technicians. Researchers conducted a series of experiments that showed an accuracy of disease recognition of up to 92%.
Together with colleagues from Perm State Aviation Technical University and the Perm Institute of the Federal Penitentiary Service of Russia , the project is being implemented. The scientific group published the results of the work in the IOP Conference Series: Earth and Environmental Science and Journal of Physics: Conference Series.
–Structural changes in cell tissues detected by histological analysis give an accurate picture of a person's disease. Now this process takes longer time, and several specialists are involved in the work, who carry out research manually. The goal of our development is to make the procedure of histological analysis to accelerate the recognition of diseases hands off. The decision support module will be able to eliminate the human factor as much as possible and increase the accuracy of the analyzes. In addition, our device will help to reduce the operating stress on medical personnel, - says Sergei Kostarev, professor of the of Information Technologies and Automated Systems Department of Perm Tech, Doctor of Technical Sciences, Associate Professor
According to scientists, morphostructural changes in cells have a tree-like structure, which makes it possible to determine indicators of pathologies automatically. The research process begins with loading material samples into the apparatus, preparing a histological specimen and scanns them. Then specially developed computer algorithms process the resulting image. The device uses neural networks to "decode" it, finding indicators of tissue destruction that characterize various pathological processes.
To develop an automated system, scientists suggested using the theory of finite automata, in particular, Mealy machine. They presented calculations for the pathology of structural changes in tissues and the coding of the necessary indicators. The researchers developed the method of automated diagnosis based on the analysis of pathology indicators. Scientists also conducted simulation modeling to determine these indicators. They programmed the device under design using the CX-One for the Omron PLC.
–We carried out a series of experiments on the recognition of cellular pathologies, the results of which were evaluated by a laboratory assistant-histologist. In the process, we used a system of histological guiding - a Leica histological processor, a rotary microtome ThermoScientific НМ 325 and other equipment. The results showed that the device will be able to achieve fairly high accuracy rates - from 82% to 92%. In the future, our device will make it possible to identify various types of pathologies: from dystrophy to oncological diseases, ”explains Sergey Kostarev.
The goal of scientists is to create a "robot histologist" that will perform the function of loading samples of material before issuing a result. Researchers are currently developing a prototype device and are looking for funding for the project.