Objective assessment of stored blood quality by deep learning

Abstract

Stored red blood cells (RBCs) are needed for life-saving blood transfusions, but they undergo continuous degradation. RBC storage lesions are often assessed by microscopic examination or biochemical and biophysical assays, which are complex, time-consuming, and destructive to fragile cells. Here we demonstrate the use of label-free imaging flow cytometry and deep learning to characterize RBC lesions. Using brightfield images, a trained neural network achieved 76.7% agreement with experts in classifying seven clinically relevant RBC morphologies associated with storage lesions, comparable to 82.5% agreement between different experts. Given that human observation and classification may not optimally discern RBC quality, we went further and eliminated subjective human annotation in the training step by training a weakly supervised neural network using only storage duration times. The feature space extracted by this network revealed a chronological progression of morphological changes that better predicted blood quality, as measured by physiological hemolytic assay readouts, than the conventional expert-assessed morphology classification system. With further training and clinical testing across multiple sites, protocols, and instruments, deep learning and label-free imaging flow cytometry might be used to routinely and objectively assess RBC storage lesions. This would automate a complex protocol, minimize laboratory sample handling and preparation, and reduce the impact of procedural errors and discrepancies between facilities and blood donors. The chronology-based machine-learning approach may also improve upon humans’ assessment of morphological changes in other biomedically important progressions, such as differentiation and metastasis.

Deep Learning Image Classification of Red Blood Cell Deformability

“…Deep learning has been used previously to characterize other cellular properties of RBCs. For example, Doan et al 38 trained a deep learning model to classify unlabeled images of stored RBCs into seven morpho-types with 76.7% accuracy, which was comparable to 82.5% agreement in manual classification by experts. Other studies trained deep learning models to identify RBCs from patients with malaria, [39][40][41][42][43][44] sickle cell disease, [45][46][47][48][49][50] and thalassemia, [51][52][53] based on visually identifiable changes in RBC morphology.…”

“…RBCs typically exhibit a highly deformable biconcave discoid morphology, and deviation from this morphology may correspond with changes in cell deformability. 36,37 In fact, deep learning methods have been developed to assess changes in RBC morphology during cold storage, 38 malaria, [39][40][41][42][43][44] sickle cell disease, [45][46][47][48][49][50] and thalassemia. [51][52][53] However, RBC morphology varies over the life cycle of the cell and this variability may obscure efforts to infer deformability from cell morphology.…”

An image-based flow cytometric approach to the assessment of the nucleus-to-cytoplasm ratio

“…Here, no gating strategies are implemented to limit the loss of IFC images. Our group combined computer vision and machine learning strategies in the context of red blood cell storage lesions [52,53]. We look to further incorporate techniques used by Doan et al [54], Blasi et al [55], and Hennig et al [56] that use an unsupervised or weakly supervised, deep-learning method [57] for the assessment of the N:C ratio in both non-malignant and malignant cell lines.…”

Image- and Fluorescence-Based Test Shows Oxidant-Dependent Damages in Red Blood Cells and Enables Screening of Potential Protective Molecules

“…New approaches like microfluidic devices could be used to evaluate the impact of aging or chemical treatments on RBC deformability [ 8 , 9 ]. Finally, cell morphology using imaging flow cytometry or digital holographic microscopy (DHM) are useful tools to provide information of RBC phenotype [ 10 , 11 , 12 , 13 ].…”