Morphometric data set characteristics and structure substantiation with prostate gland magnetic resonance imaging results

The purpose is to substantiate the characteristics, develop a data set for testing and monitoring the quality of artificial intelligence (AI) services for prostate morphometry on magnetic resonance imaging (MRI) results. We carried out a search of scientific literature in PubMed and RINC databases with the search depth, mostly not exceeding 10 years. In the structure of malignant neoplasm morbidity among the male population, tumors steadily rank first in Russia and second in the global perspective. In Russia in 2013, the corresponding morbidity was 34.62 cases per 100 thousand people, and in 2023, it increased to 50.33 cases. There is a steady annual increase in the absolute number of first-time diagnoses of prostate cancer. There was a certain decline in this indicator during the COVID-19 pandemic, but now the growth has restarted. This type of pathology ranks third (9.0%) in the structure of mortality from malignant neoplasms among the male population. MRI has a special role in screening and diagnosis of prostate diseases. The varying levels of doctors’ competence and the labor-intensive processing, interpretation, and measurements often limit the implementation of MRI, because of the time needed to describe the research results. Computer vision technologies can serve as one of the potential ways of solving this problem. Both the lack of data itself and defects in data markup limit the introduction of AI technologies in practical healthcare.