Celsus neural network developers for the analysis of medical images have completed the first version of the system which is used for the analysis of histological studies. Now the service's artificial intelligence will be able to examine breast tissue samples and identify the presence of tumor cells.
Histological examination is essential for cancer desease diagnosing as it helps to identify the type of neoplasm and determine its characteristics. In order to perform this analysis, the pathologist should examine a tissue fragment taken from a patient. Then the sample is scanned, digitized, and displayed on the screen. At this point, artificial intelligence can improve the process. Celsus independently locates the areas with pathological cells and separate the affected areas from the healthy ones. The service programmers also plan to train the system to mark suspicious areas.
The research work is being carried out in collaboration with Nizhny Novgorod Regional Clinical Oncology Center. Qualified doctors assisted us in preparing and marking the data. To train the system, we used databases with verified cases of cancer.
Artificial intelligence is helpful in conducting histological analysis, which will not only reduce the risks of incorrect and inaccurate diagnosis, but also the burden on pathomorphologists, as well as boost the research process. Currently, histological examination is a difficult study which is not available in every region of the country. Tissue samples are often sent to the laboratories of bigger cities which not only requires more time but also places an additional burden on the specialists.