A rapid white blood cell classification system based on multimode imaging technology

Meng Lv,Xi Zhao,Feng Chen,Ming Yu,Chao Li,Jinggong Sun

DOI: https://doi.org/10.1002/jbio.202000197

IF: 3.3899

2020-09-02

Journal of Biophotonics

Abstract:<p>In order to simplify the complexity of white blood cell classification in existing POCT testing equipment, a white blood cell classification detection system based on microfluidic and multimode imaging was constructed. Microfluidic chip was used in the system. A multimodal optical imaging system based on the characteristics of blood samples was designed to obtain eigenvalue extraction of cells. Afterwards, a BP neural network model was constructed to realize automatic classification of white blood cells. Finally, 80 human blood samples were classified and detected by this system and compared with the results of Sysmex XE‐5000. The consistency correlation coefficients of white blood cells, lymphocytes, monocytes, neutrophils and eosinophils are 1.038, 0.907, 0.549, 0.922, and 1.028, respectively and the CV values of the four types of white blood cells in the stability test were all below 10%. In this study, a white blood cell classification and detection system with small size, simple operation, fast single‐sample detection, high accuracy, and no maintenance is required. It will provide a solid technical support for the further development of POCT blood cell analysis equipment.</p><p>This article is protected by copyright. All rights reserved.</p>

optics,biochemical research methods,biophysics

What problem does this paper attempt to address?

### Problems Addressed by the Paper

The paper aims to simplify the complexity of leukocyte classification in existing point-of-care testing (POCT) devices and proposes a leukocyte classification system based on microfluidic chips and multimodal imaging technology.

Specifically, the paper addresses the following issues:
1. **Simplifying the leukocyte classification process**: Traditional leukocyte classification methods require professionals to perform microscopic observations. Although existing automated blood cell analyzers have high accuracy and large sample volumes, they face challenges such as large equipment size, poor portability, complex operation, and high maintenance costs in complex environments like battlefield rescue and earthquake rescue. Therefore, the paper proposes a new system to simplify this process.
2. **Improving portability and ease of operation**: By using microfluidic chips and multimodal imaging technology, the new system features small size, simple operation, fast single-sample detection speed, high accuracy, and no maintenance requirements.
3. **Enhancing detection accuracy**: The paper constructs a BP neural network model to achieve automatic leukocyte classification. Experimental results show that the system performs well in classifying lymphocytes, neutrophils, and eosinophils, but the classification of monocytes is less effective and requires further optimization.
4. **Validating system performance**: Through comparative experiments with the Sysmex XE-5000 device, the accuracy and stability of the system were verified. Experimental results indicate that, except for monocytes, the classification accuracy of the other three types of leukocytes exceeds 90%, and the coefficient of variation (CV value) in repeatability tests is below 10%.

In summary, the paper proposes a new leukocyte classification system aimed at improving portability, ease of operation, and detection accuracy to meet the demands of complex environments such as battlefield rescue and earthquake rescue.