| Ph-myeloproliferative neoplasm（MPN）is a group of hematopoietic stem cell clonal diseases。It usually causes abnormal increase of cell number of one or multiple lines.They mainly include Essential thrombocytosis（ET）,Polycythemia Vera（PV）and Primary Thrombocytosis（PMF）.At present,MPN is mainly diagnosed based on clinical manifestations,hematological changes,histopathology and typical gene mutations.Recently,many literatures have discussed that the occurrence and development of MPN is closely related to inflammation,and some scholars even believe that the damage of inflammation on hematopoietic stem cells in MPN is earlier than the occurrence of gene mutation.Dragana et al.[1]emphasized the importance of neutrophils in the occurrence and development of MPNS.In addition,the present study suggests the presence of three heterogeneous neutrophils in tumor patients.With the rapid development of artificial intelligence in recent years,this cutting-edge technology has been gradually applied to various aspects of the medical field.Deep learning in artificial intelligence has unparalleled information mining ability.It can automatically capture potential information without relying on human selection features.It plays an important role in image classification,object detection and semantic segmentation.More and more studies have also confirmed that the accuracy and sensitivity of deep learning algorithm in distinguishing different cell morphology in blood diseases can reach more than 90%,reaching or even surpassing the expert level.In this study,a new method was proposed to establish a deep learning model based on neutrophil morphology to identify myeloproliferative neoplasm and determine whether patients are MPN and MPN subtypes.The Pico Det deep learning object detection method selected by us achieves good classification prediction effect.Therefore,deep learning can realize auxiliary diagnosis and classification of MPNS. |
PDF Full Text Request