A deep learning method and device for bone marrow imaging cell detection

Abstract

Background: Morphological analysis of bone marrow cells is considered as the gold standard for the diagnosis of leukemia. However, due to the diverse morphology of bone marrow cells, extensive experience and patience are needed for morphological examination. automatic diagnosis system through the comprehensive application of image analysis and pattern recognition technology is urgently needed to reduce work intensity, error probability and improves work efficiency. Methods: In this article, we establish a new morphological diagnosis system for bone marrow cell detection based on the deep learning object detection framework. The model is based on the Faster Region-Convolutional Neural Network (R-CNN), a classical object detection model. The system automatically detects bone marrow cells and determines their types. As specimens have severe long-tail distribution, i.e., the frequency of different types of cells varies dramatically, we proposed a general score ranking loss to solve such a problem. The general score ranking loss considers the ranking relationship between positive and negative samples and optimizes the positive sample with a higher classification probability value.Results: We verified this system with 70 bone marrow specimens of leukemia patients, which proved that it can realize intelligent recognition with high efficiency. The software is finally integrated into the microscope system to build an augmented reality system.Conclusions: Clinical tests show that the response speed of the newly developed diagnostic system is faster than that of trained diagnostic experts.

A scoping review of deep learning in cancer nursing combined with augmented reality: The era of intelligent nursing is coming

“…J o u r n a l P r e -p r o o f ) and Italy (n = 1). Of 6 studies, four were diagnostic study [37][38][39][40] , one was report [41] and the study design of the rest one is unclear [42] . All of the included studies conducted the application development of DL+AR.…”

“…All of the included studies conducted the application development of DL+AR. Five studies [37][38][39][40][41] developed the microscopes based on the deep learning method and augmented reality module to improve the accuracy and efficiency of cancer diagnosis, and the rest one [42] established a two-steps automatic system to improve accuracy in locating the tumor…”

“…They can simplify workflow and facilitate clinical diagnosis. For example, Liu [40 ] established a new morphological diagnosis system of bone marrow cells based on the DL object detection framework and built an AR system, with clinical test showing that the newly developed diagnosis system can respond more rapidly compared with a well-trained expert for diagnosis. In addition, DL+AR is also applied well in surgeries to help accurately localize tumors.…”

“…In the stage of technology development in these studies, first of all, when the DL model was constructed, a certain number of samples were selected and the selected combined with AR. In the five studies that developed the AI microscope [37][38][39][40][41] , the results identified by the AI system were then projected into the original optical path by an AR technique. In one study that implemented a two-steps automatic system [42] ,…”

“…Tanzi [42] picked U-Net with MobileNet as the most performing ensemble for performance and speed and the frame rate of the real-time application was acceptable if they kept the number of iterations per second (it/s) greater than 10: with 11.04 it/s, on the hardware used for J o u r n a l P r e -p r o o f testing, it is able to reach a frame rate of approximately 8fps, which is a good value for a medical application. In addition to the characteristics of high-speed analysis, Liu [40] proposed a general score ranking loss, which would weaken the impact on the correctly ranked positive samples and provided extra optimization to make positive samples with larger prediction score, and improve the detection accuracy.…”