The deviations in blood cell concentration beyond reasonable ranges may indicate the presence of certain diseases within the body
Opto-Electronic Science, Volume. 2, Issue 12, 230018(2023)
Smart palm-size optofluidic hematology analyzer for automated imaging-based leukocyte concentration detection
A critical function of flow cytometry is to count the concentration of blood cells, which helps in the diagnosis of certain diseases. However, the bulky nature of commercial flow cytometers makes such tests only available in hospitals or laboratories, hindering the spread of point-of-care testing (POCT), especially in underdeveloped areas. Here, we propose a smart Palm-size Optofluidic Hematology Analyzer based on a miniature fluorescence microscope and a microfluidic platform to lighten the device to improve its portability. This gadget has a dimension of 35 × 30 × 80 mm and a mass of 39 g, less than 5% of the weight of commercially available flow cytometers. Additionally, automatic leukocyte concentration detection has been realized through the integration of image processing and leukocyte counting algorithms. We compared the leukocyte concentration measurement between our approach and a hemocytometer using the Passing-Bablok analysis and achieved a correlation coefficient of 0.979. Through Bland-Altman analysis, we obtained the relationship between their differences and mean measurement values and established 95% limits of agreement, ranging from ?0.93×103 to 0.94×103 cells/μL. We anticipate that this device can be used widely for monitoring and treating diseases such as HIV and tumors beyond hospitals.
Introduction
The deviations in blood cell concentration beyond reasonable ranges may indicate the presence of certain diseases within the body
The commonly used blood analyzers include imaging hematology analyzers and flow cytometers. Imaging hematology analyzers perform cell counting by analyzing cell images on glass slides and are easily integrable. For instance, the MSLBX01 white blood cell analyzer by Hangzhou Livie Technology Co., Ltd. has dimensions of only 138 mm × 140 mm × 116 mm and weighs merely 750 g. Furthermore, smartphones, with their excellent imaging capabilities, often incorporate meticulously designed optics, light sources, and mechanical components to achieve portable imaging hematology analyzers
On the other hand, the flow cytometers allow high-throughput, accurate, and rapid sorting and counting of blood cells in fluids
Microfluidic chips offer a retrofit solution for miniaturization and functional expansion of flow cytometry
Here, we have constructed a miniature fluorescence microscope and integrated it with a compact microfluidic platform to realize a Palm-size Optofluidic Hematology Analyzer, measuring 35 × 30 × 80 mm and weighing 39 g. To obtain the concentration of white blood cells in the sample, the miniaturized fluorescence microscope records the stained white blood cells pumped into the field of view per unit time and subsequently, a particle counting algorithm is employed to quantify the cell numbers. We used this Palm-size Optofluidic Hematology Analyzer to measure leukocyte concentration in blood samples and compared the results with the counting values from a benchtop hemocytometer. The reliability of our device has been demonstrated by Passing-Bablok regression analysis and Bland-Altman analysis. Taken together, we have achieved the tiniest optofluidic hematology analyzer to our best knowledge, and it overcomes the bulky limitations of traditional flow cytometers and the sample throughput constraints of imaging hematology analyzers.
Materials and methods
Principle of the Palm-size Optofluidic Hematology Analyzer
As shown in
Figure 1.
Pre-treatment of blood samples
Before performing the white blood cell concentration measurements using the device, pre-processing of the blood samples is required. To maintain the structural and physiological integrity of white blood cells, we took 100 μL of blood sample and placed it in a brown light-protected vial. First, 20 μL of acridine orange nucleic acid dye (AO Sigma) and 20 μL of red blood cell lysing solution (NP-40 Amresco) were added to the sample. Second, the mixture was diluted tenfold with 0.01 mol/L phosphate buffer solution and gently shaken. Finally, the sample was stained under ambient temperature and light-protected conditions for 15–20 minutes.
To avoid the cells clumping together and affecting identification, the sample was sufficient shaking and homogenization. The blood samples were then added to a syringe and driven by a high-precision pump into the channel of the microfluidic chip. The blood samples were diluted tenfold and flowed through the channel at a rate of 1 μL/min by adjusting the pump and recording the number of leucocytes passing through in 30 seconds. The flow rate and the number of leucocytes allow us to calculate the concentration of leucocytes in this sample, as shown in
Particle counting algorithm
The particle counting algorithm is based on Python and implemented through NumPy, SciPy, and OpenCV. The core of the particle counting algorithm is a centroid tracking method, where particles in adjacent frames are correlated by their minimum Euclidean distance from each other. The principle of centroid tracking is shown in
Figure 2.(
The particle counting algorithm is divided into image pre-processing and particle counting. The image pre-processing is shown in
The particle counting is depicted in
Statistical analyses
It is necessary to introduce another reliable method to test the same set of samples and perform consistency analysis to verify the validity and accuracy of the proposed method. Here, we utilized a commonly used benchtop hemocytometer for side-by-side comparisons of our device (more detailed experimental procedure in Supplementary Section 4). To assess the correlation and overall consistency between these two approaches, we applied the Passing-Bablok regression and the Bland-Altman analysis, respectively.
Compared to conventional linear fitting, the Passing-Bablok regression analysis is non-parametric, robust, and insensitive to errors and outliers and is suitable for evaluating the correlation of the two testing devices. Thus, the Passing-Bablok regression analysis
The a and b are the medians of the slope and intercept determined at any two points in the Cartesian coordinate system, respectively. If the 95% confidence interval of parameter a includes 1, and the 95% confidence interval of parameter b contains 0, there is no systematic or proportional difference between these two methods, and they can be considered interchangeable.
To evaluate the consistency of these two devices, and visually illustrate the differences between two measurement methods and their relationship with the mean measurement value, the Bland-Altman analysis
Results and discussion
Results
Leukocytes within the microfluidic channel captured by the miniature fluorescence microscope are shown in a distinct outline in
Figure 3.(
For side-by-side comparisons between the proposed device and a conventional benchtop hemocytometer, we took blood samples from 40 groups of patients for white blood cell counting. With 10 repetitive measurements of each sample, we verified the reliability of the device by Passing-Bablok regression analysis, as shown in
Figure 4.(
The standard values, average values, and errors of the white blood cell concentration for partial our data are listed in
|
Discussion
There are several factors that affect the accuracy of blood cell count. Firstly, whole blood left in vitro for too long can lead to apoptosis of some blood cells and reduce the staining efficiency. In addition, cells adhering together or attaching to the channel wall can cause a loss of count. Therefore, the key to further improving the accuracy of the Palm-size Optofluidic Hematology Analyzer is to ensure that the blood sample is fresh and active, to perform a rigorous and standardized staining process, and to agitate it adequately. Furthermore, regarding the mass of the gadget, it could be further reduced, for example, by changing the optical resin shelf to aluminum or plastic.
Increasing the numerical aperture of the objective of the miniature fluorescence microscope and incorporating deep learning allow the classification and count of specific species of leukocytes. Moreover, microfluidic chips, in combination with other types of miniaturized microscopic imaging systems, such as miniaturized confocal microscopy or miniaturized phase contrast microscopy, also have a wide range of applications and potential for development.
To drive the sample through the microfluidic chip at a constant flow rate, a peristaltic pump is also required during operation. The data processing is conducted on a laptop computer. The peristaltic pump is essentially a variable-speed motor, and the palm-sized versions are also commercially available. Furthermore, considering the widespread availability of laptop computers, we argue that the proposed device is portable enough and fully meets the needs of POCT.
Conclusions
We built a miniature fluorescence microscope and combined it with microfluidics to construct a Palm-size Optofluidic Flow Hematology Analyzer 35 × 30 × 80 mm with a mass of 39 g and validated the gadget by Passing-Bablok regression analysis and Bland-Altman analysis. This device calculates leukocyte concentration with an error of less than 10%, meeting the requirements of UK NEQAS and CLIA-88 for white blood cell count accuracy.
Further improvements in our device are expected. To avoid issues such as out-of-focus caused by oscillations during movement, we will upgrade the casing to integrate the miniaturized fluorescence microscope system more securely with the microfluidic chip. Additionally, to further optimize the overall size and weight of the system, we intend to reduce the dimensions of the microfluidic chips. Regarding the analysis approaches, we will develop particle identification and classification algorithms, and integrate them with deep learning techniques. This will enable real-time classification of different types of white blood cells, providing more comprehensive information for disease diagnosis.
This allows POCT of patients’ blood cells away from the hospital or laboratory environment and enhances medical diagnosis in remote or deprived areas. In addition, performing blood cell counts by astronauts in space environments holds significant importance in the fields of radiation biology and microgravity biology. However, in such resource-limited and energy-intensive environments, the use of conventional equipment requiring chemical fuels and occupying substantial space significantly increases costs. In this context, the development of the Palm-size Optofluidic Hematology Analyzer offers a potential solution by addressing the issues of volume and weight.
Acknowledgements
This work was supported by the National Natural Science Foundation of China (grant no. 62305083 to W. Z., grant no. T2222009 to H. L., grant no. 32227802 to H. L.), China Postdoctoral Science Foundation (grant no. 2023T160163 to W. Z. grant no. 2022M720971 to W. Z.), the Heilongjiang Provincial Postdoctoral Science Foundation (grant no. LBH-Z22027 to W. Z.), the National Key Research and Development Program of China (grant no. 2022YFC3400600 to H. L.), and the Natural Science Foundation of Heilongjiang Province (grant no. YQ2021F013 to H. L.).
The authors declare no competing financial interests.
Supplementary information for this paper is available at
[26] T Maleki, T Fricke, JT Quesenberry, PW Todd, JF Leary. Point-of-care, portable microfluidic blood analyzer system. Proc SPIE, 8251, 82510C(2012).
[32] V Kumar, AK Abbas, JC Aster.
[33] L Bilić-Zulle. Comparison of methods: passing and Bablok regression. Biochem Med, 21, 49-52(2011).
Get Citation
Copy Citation Text
Deer Su, Xiangyu Li, Weida Gao, Qiuhua Wei, Haoyu Li, Changliang Guo, Weisong Zhao. Smart palm-size optofluidic hematology analyzer for automated imaging-based leukocyte concentration detection[J]. Opto-Electronic Science, 2023, 2(12): 230018
Category: Research Articles
Received: Jul. 30, 2023
Accepted: Oct. 9, 2023
Published Online: Mar. 19, 2024
The Author Email: Guo Changliang (CLGuo), Zhao Weisong (WSZhao)