On March 7, 2022, Requirements of pathological datasets for assisted diagnosis of chronic myelogenous leukemia (CML) (standard number: T/CSBME 051—2022) was officially published on National Platform for Group Standards Information. This group standard, led by Union Hospital of Huazhong University of Science and Technology and initiated by Chinese Society of Biomedical Engineering, was jointly drafted and verified by 47 clinicians, pathologists and artificial intelligence experts from 32 hospitals, 2 research institutes and 7 enterprises in China. Professor He Pengcheng from Department of Hematology of the First Affiliated Hospitalof Xi’an Jiaotong University (XJTU) participated in the formulation of this group standard.

Requirements of pathological datasets for assisted diagnosis of chronic myelogenous leukemia (CML)is the first group standard of artificial intelligence-assisted diagnosis datasets of hematological pathology.In this group standard, the requirements of dataset design, sample collection and processing, data collection, data preprocessing and dataset management and maintenance of assisted diagnosis of CML pathology are specified, which are applicable to the algorithm model training, model verification, performance testing, clinical evaluation, product quality control and other steps in the research and development of products related to assisted diagnosis of CML pathology.

The advancement of artificial intelligence represents deepening application of big data. Data is not only the basis of the application of artificial intelligence, but also the resource of artificial intelligence. At present, lack of high-quality data has become one of the bottlenecks restricting the development of artificial intelligence. The formulation of this group standard provides a direction for screening and constructing effective standardized datasets, brings novel breakthroughs for the research and development and application of artificial intelligence in the field of hematological diseases. It represents deep integration of artificial intelligence counting and medical and health fields, which is expected to further accelerate rapid development of dataset industry standards for hematological diseases, and promote the research and clinical translation of intelligent diagnostic products of hematological diseases.