Artificial‐Intelligence‐Enabled Reagent‐Free Imaging Hematology Analyzer. (2nd June 2021)

• Record Type: Journal Article
• Authors: Shu, Xin; Sansare, Sameera; Jin, Di; Zeng, Xiangxiang; Tong, Kai-Yu; Pandey, Rishikesh; Zhou, Renjie
• Is Part Of: Advanced intelligent systems. Volume 3:Number 8(2021)
• Page Start: n/a
• Page End: n/a
• Publication Date: 2021-06-02
• Subjects: cell classifications -- deep learning -- label-free imaging -- leukocyte classifications -- quantitative phase imaging; Artificial intelligence -- Periodicals; Robotics -- Periodicals; Control theory -- Periodicals; 006.3
• DOI: 10.1002/aisy.202000277
• Languages: English
• ISSNs: 2640-4567

Abstract

Abstract : Leukocyte differential test is a widely carried out clinical procedure for screening infectious diseases. Existing hematology analyzers require labor‐intensive work and a panel of expensive reagents. Herein, an artificial‐intelligence‐enabled reagent‐free imaging hematology analyzer (AIRFIHA) modality is reported that can accurately classify subpopulations of leukocytes with minimal sample preparation. AIRFIHA is realized through training a two‐step residual neural network using label‐free images of isolated leukocytes acquired from a custom‐built quantitative phase microscope. By leveraging the rich information contained in quantitative phase images, not only high accuracy is achieved in differentiating B and T lymphocytes, but also CD4 and CD8 T cells are classified, therefore outperforming the classification accuracy of most current hematology analyzers. The performance of AIRFIHA in a randomly selected test set is validated and is cross‐validated across all blood donors. Due to its easy operation, low cost, and accurate discerning capability of complex leukocyte subpopulations, AIRFIHA is clinically translatable and can also be deployed in resource‐limited settings, e.g., during pandemic situations for the rapid screening of infectious diseases. Abstract : A reagent‐free hematology analyzer is developed for classifying human leukocyte subpopulations. The system is realized through a cascaded residual neural network and a quantitative phase microscope. Human B and T cells are accurately classified without using chemical labels, outperforming bright‐field‐/dark‐field‐based imaging methods with machine learning. This low‐cost and ease‐of‐use method is suitable for deployment in resource‐limited settings.