National Taiwan University, Graduate Institute of Biomedical Electronics and Bioinformatics, Taipei, Taiwan.
Feng Chia University, College of Information and Electrical Engineering, Taichung, Taiwan.
J Biomed Opt. 2020 Nov;25(11). doi: 10.1117/1.JBO.25.11.116502.
Label-free quantitative phase imaging is a promising technique for the automatic detection of abnormal red blood cells (RBCs) in real time. Although deep-learning techniques can accurately detect abnormal RBCs from quantitative phase images efficiently, their applications in diagnostic testing are limited by the lack of transparency. More interpretable results such as morphological and biochemical characteristics of individual RBCs are highly desirable.
An end-to-end deep-learning model was developed to efficiently discriminate thalassemic RBCs (tRBCs) from healthy RBCs (hRBCs) in quantitative phase images and segment RBCs for single-cell characterization.
Two-dimensional quantitative phase images of hRBCs and tRBCs were acquired using digital holographic microscopy. A mask region-based convolutional neural network (Mask R-CNN) model was trained to discriminate tRBCs and segment individual RBCs. Characterization of tRBCs was achieved utilizing SHapley Additive exPlanation analysis and canonical correlation analysis on automatically segmented RBC phase images.
The implemented model achieved 97.8% accuracy in detecting tRBCs. Phase-shift statistics showed the highest influence on the correct classification of tRBCs. Associations between the phase-shift features and three-dimensional morphological features were revealed.
The implemented Mask R-CNN model accurately identified tRBCs and segmented RBCs to provide single-RBC characterization, which has the potential to aid clinical decision-making.
无标记定量相位成像是一种很有前途的技术，可实时自动检测异常的红细胞（RBC）。虽然深度学习技术可以高效地从定量相位图像中准确检测异常 RBC，但它们在诊断测试中的应用受到缺乏透明度的限制。更具可解释性的结果，如单个 RBC 的形态和生化特征，是非常需要的。
开发了一种端到端的深度学习模型，以有效地从定量相位图像中区分地中海贫血 RBC（tRBC）和健康 RBC（hRBC），并对 RBC 进行分割以进行单细胞特征描述。
使用数字全息显微镜采集 hRBC 和 tRBC 的二维定量相位图像。基于掩模区域的卷积神经网络（Mask R-CNN）模型被训练来区分 tRBC 并分割单个 RBC。利用 SHapley Additive exPlanation 分析和典型相关分析对自动分割的 RBC 相位图像进行 tRBC 特征描述。
所实现的模型在检测 tRBC 方面达到了 97.8%的准确率。相移统计显示对 tRBC 的正确分类有最高的影响。揭示了相移特征与三维形态特征之间的关联。
所实现的 Mask R-CNN 模型准确地识别了 tRBC 并对 RBC 进行了分割，以提供单个 RBC 的特征描述，这有可能辅助临床决策。
