Dynamics of shape recovery by stored red blood cells during washing at the single cell level

Deep ensemble learning enables highly accurate classification of stored red blood cell morphology

Changes in red blood cell (RBC) morphology distribution have emerged as a quantitative biomarker for the degradation of RBC functional properties during hypothermic storage. Previously published automated methods for classifying the morphology of stored RBCs often had insufficient accuracy and relied on proprietary code and datasets, making them difficult to use in many research and clinical applications. Here we describe the development and validation of a highly accurate open-source RBC morphology classification pipeline based on ensemble deep learning (DL). The DL-enabled pipeline utilized adaptive thresholding or semantic segmentation for RBC identification, a deep ensemble of four convolutional neural networks (CNNs) to classify RBC morphology, and Kalman filtering with Hungarian assignment for tracking changes in the morphology of individual RBCs over time. The ensembled CNNs were trained and evaluated on thousands of individual RBCs from two open-access datasets previously collected to quantify the morphological heterogeneity and washing-induced shape recovery of stored RBCs. Confusion matrices and reliability diagrams demonstrated under-confidence of the constituent models and an accuracy of about 98% for the deep ensemble. Such a high accuracy allowed the CNN ensemble to uncover new insights over our previously published studies. Re-analysis of the datasets yielded much more accurate distributions of the effective diameters of stored RBCs at each stage of morphological degradation (discocyte: 7.821 ± 0.429 µm, echinocyte 1: 7.800 ± 0.581 µm, echinocyte 2: 7.304 ± 0.567 µm, echinocyte 3: 6.433 ± 0.490 µm, sphero-echinocyte: 5.963 ± 0.348 µm, spherocyte: 5.904 ± 0.292 µm, stomatocyte: 7.080 ± 0.522 µm). The effective diameter distributions were significantly different across all morphologies, with considerable effect sizes for non-neighboring classes. A combination of morphology classification with cell tracking enabled the discovery of a relatively rare and previously overlooked shape recovery of some sphero-echinocytes to early-stage echinocytes after washing with 1% human serum albumin solution. Finally, the datasets and code have been made freely available online to enable replication, further improvement, and adaptation of our work for other applications.

Hypoxic storage of donor red cells preserves deformability after exposure to plasma from adults with sickle cell disease

BACKGROUND

Red cell (RBC) transfusions are beneficial for patients with sickle cell disease (SCD), but ex vivo studies suggest that inflamed plasma from patients with SCD during crises may damage these RBCs, diminishing their potential efficacy. The hypoxic storage of RBCs may improve transfusion efficacy by minimizing the storage lesion. We tested the hypotheses that (1) The donor RBCs exposed to the plasma of patients in crisis would have lower deformability and higher hemolysis than those exposed to non-crisis plasma, and (2) hypoxic storage, compared to standard storage, of donor RBCs could preserve deformability and reduce hemolysis.

STUDY DESIGN AND METHODS

18 SCD plasma samples from patients who had severe acute-phase symptoms (A-plasma; n = 9) or were at a steady-state (S = plasma; n = 9) were incubated with 16 RBC samples from eight units that were stored either under conventional(CRBC) or hypoxic(HRBC) conditions. Hemolysis and microcapillary deformability assays of these RBCs were analyzed using linear mixed-effect models after each sample was incubated in patient plasma overnight at 37°C RESULTS: Relative deformability was 0.036 higher (p < 0.0001) in HRBC pairs compared to CRBC pairs regardless of plasma type. Mean donor RBC hemolysis was 0.33% higher after incubation with A-plasma compared to S-plasma either with HRBC or CRBC (p = 0.04). HRBCs incubated with steady-state patient plasma demonstrated the highest deformability and lowest hemolysis.

CONCLUSION

Hypoxic storage significantly influenced RBC deformability. Patient condition significantly influenced post-incubation hemolysis. Together, HRBCs in steady-state plasma maximized donor red cell ex vivo function and survival.

Storage-Induced Micro-Erythrocytes Can Be Quantified and Sorted by Flow Cytometry

Refrigerated storage of red cell concentrates before transfusion is associated with progressive alterations of red blood cells (RBC). Small RBC (type III echinocytes, sphero-echinocytes, and spherocytes) defined as storage-induced micro-erythrocytes (SME) appear during pretransfusion storage. SME accumulate with variable intensity from donor to donor, are cleared rapidly after transfusion, and their proportion correlates with transfusion recovery. They can be rapidly and objectively quantified using imaging flow cytometry (IFC). Quantifying SME using flow cytometry would further facilitate a physiologically relevant quality control of red cell concentrates. RBC stored in blood bank conditions were stained with a carboxyfluorescein succinimidyl ester (CFSE) dye and incubated at 37°C. CFSE intensity was assessed by flow cytometry and RBC morphology evaluated by IFC. We observed the accumulation of a CFSE ^high RBC subpopulation by flow cytometry that accounted for 3.3 and 47.2% at day 3 and 42 of storage, respectively. IFC brightfield images showed that this CFSE ^high subpopulation mostly contains SME while the CFSE ^low subpopulation mostly contains type I and II echinocytes and discocytes. Similar numbers of SME were quantified by IFC (based on projected surface area) and by flow cytometry (based on CFSE intensity). IFC and scanning electron microscopy showed that ≥95% pure subpopulations of CFSE ^high and CFSE ^low RBC were obtained by flow cytometry-based sorting. SME can now be quantified using a common fluorescent dye and a standard flow cytometer. The staining protocol enables specific sorting of SME, a useful tool to further characterize this RBC subpopulation targeted for premature clearance after transfusion.

Microfluidic capillary networks are more sensitive than ektacytometry to the decline of red blood cell deformability induced by storage

Ektacytometry has been the primary method for evaluating deformability of red blood cells (RBCs) in both research and clinical settings. This study was designed to test the hypothesis that the flow of RBCs through a network of microfluidic capillaries could provide a more sensitive assessment of the progressive impairment of RBC deformability during hypothermic storage than ektacytometry. RBC units (n = 9) were split in half, with one half stored under standard (normoxic) conditions and the other half stored hypoxically, for up to 6 weeks. RBC deformability was measured weekly using two microfluidic devices, an artificial microvascular network (AMVN) and a multiplexed microcapillary network (MMCN), and two commercially available ektacytometers (RheoScan-D and LORRCA). By week 6, the elongation indexes measured with RheoScan-D and LORRCA decreased by 5.8–7.1% (5.4–6.9% for hypoxic storage). Over the same storage duration, the AMVN perfusion rate declined by 27.5% (24.5% for hypoxic) and the MMCN perfusion rate declined by 49.0% (42.4% for hypoxic). Unlike ektacytometry, both AMVN and MMCN measurements showed statistically significant differences between the two conditions after 1 week of storage. RBC morphology deteriorated continuously with the fraction of irreversibly-damaged (spherical) cells increasing significantly faster for normoxic than for hypoxic storage. Consequently, the number of MMCN capillary plugging events and the time MMCN capillaries spent plugged was consistently lower for hypoxic than for normoxic storage. These data suggest that capillary networks are significantly more sensitive to both the overall storage-induced decline of RBC deformability, and to the differences between the two storage conditions, than ektacytometry.