Integrating deep learning with microfluidics for biophysical classification of sickle red blood cells adhered to laminin.

Sickle cell disease, a genetic disorder affecting a sizeable global demographic, manifests in sickle red blood cells (sRBCs) with altered shape and biomechanics. sRBCs show heightened adhesive interactions with inflamed endothelium, triggering painful vascular occlusion events. Numerous studies employ microfluidic-assay-based monitoring tools to quantify characteristics of adhered sRBCs from high resolution channel images. The current image analysis workflow relies on detailed morphological characterization and cell counting by a specially trained worker. This is time and labor intensive, and prone to user bias artifacts. Here we establish a morphology based classification scheme to identify two naturally arising sRBC subpopulations-deformable and non-deformable sRBCs-utilizing novel visual markers that link to underlying cell biomechanical properties and hold promise for clinically relevant insights. We then set up a standardized, reproducible, and fully automated image analysis workflow designed to carry out this classification. This relies on a two part deep neural network architecture that works in tandem for segmentation of channel images and classification of adhered cells into subtypes. Network training utilized an extensive data set of images generated by the SCD BioChip, a microfluidic assay which injects clinical whole blood samples into protein-functionalized microchannels, mimicking physiological conditions in the microvasculature. Here we carried out the assay with the sub-endothelial protein laminin. The machine learning approach segmented the resulting channel images with 99.1±0.3% mean IoU on the validation set across 5 k-folds, classified detected sRBCs with 96.0±0.3% mean accuracy on the validation set across 5 k-folds, and matched trained personnel in overall characterization of whole channel images with R2 = 0.992, 0.987 and 0.834 for total, deformable and non-deformable sRBC counts respectively. Average analysis time per channel image was also improved by two orders of magnitude (∼ 2 minutes vs ∼ 2-3 hours) over manual characterization. Finally, the network results show an order of magnitude less variance in counts on repeat trials than humans. This kind of standardization is a prerequisite for the viability of any diagnostic technology, making our system suitable for affordable and high throughput disease monitoring.

Evaluation of the CellaVision Advanced RBC Application for Detecting Red Blood Cell Morphological Abnormalities.

BACKGROUND: The Advanced RBC Application of the CellaVision DM9600 system (CellaVision AB, Lund, Sweden) automatically characterizes and classifies red blood cells (RBCs) into 21 morphological categories based on their size, color, shape, and inclusions. We evaluated the diagnostic performance of the CellaVision Advanced RBC Application with respect to the classification and grading of RBC morphological abnormalities in accordance with the 2015 International Council for Standardization in Haematology (ICSH) guidelines. METHODS: A total of 223 samples, including 123 with RBC morphological abnormalities and 100 from healthy controls, were included. Seven RBC morphological abnormalities and their grading obtained with CellaVision DM9600 pre- and post-classification were compared with the results obtained using manual microscopic examination. The grading cut-off percentages were determined in accordance with the 2015 ICSH guidelines. The sensitivity and specificity of the CellaVision DM9600 system were evaluated using the manual microscopic examination results as a true positive. RESULTS: In pre-classification, >90% sensitivity was observed for target cells, tear drop cells, and schistocytes, while >90% specificity was observed for acanthocytes, spherocytes, target cells, and tear drop cells. In post-classification, the detection sensitivity and specificity of most RBC morphological abnormalities increased, except for schistocytes (sensitivity) and acanthocytes (specificity). The grade agreement rates ranged from 35.9% (echinocytes) to 89.7% (spherocytes) in pre-classification and from 46.2% (echinocytes) to 90.1% (spherocytes) in post-classification. The agreement rate of samples with within-one grade difference exceeded 90% in most categories, except for schistocytes and echinocytes. CONCLUSIONS: The Advanced RBC Application of CellaVision DM9600 is a valuable screening tool for detecting RBC morphological abnormalities.

The frequency of occurrence of fish-shaped red blood cells in different haematologic disorders.

BACKGROUND: Red blood cells (RBC) resembling the silhouette of a fish are rarely observed in peripheral blood (PB) smears. In this study, we determined the frequency of occurrence of fish-shaped RBC in different haematologic diseases. METHODS: We examined PB smears of patients with iron deficiency anaemia (IDA) (n=23), ß-thalassaemia minor (BTM) (n=30), sickle cell disease (SCD) (n=7), autoimmune haemolytic anaemia (AIHA) (n=13), microangiopathic haemolytic anaemia (MAHA) (n=11), hereditary sphaerocytosis (HS) (n=4), hereditary elliptocytosis (HE) (n=3), vitamin B12 and folate deficiency (n=15), anaemia in liver disease (LD) (n=17), myelodysplastic syndrome (MDS) (n=15), acute myeloid leukaemia (AML) (n=29), chronic myeloid leukaemia (CML) (n=18), primary myelofibrosis (PMF) (n=12), chronic myelo-monocytic leukaemia (CMML) (n=15) and 21 healthy controls by light microscopy for the occurrence of fish-shaped erythrocytes. The fish-shaped RBC were counted as cells per 20 high-power fields (HPF) at 1000-fold magnification, and slides containing ≥1 fish-shaped RBC/20 HPF were regarded as positive. RESULTS: Fish-shaped RBC were significantly found in HE, iron deficiency, vitamin B12/folate deficiency, LD and PMF. The highest numbers of fish-shaped RBC were seen in HE and vitamin B12/folate deficiency. In patients with BTM, MDS, AML and CMML, this RBC anomaly was only occasionally observed. Furthermore, a statistically significant negative correlation of haemoglobin with the occurrence of fish-shaped RBC was apparent (p<0.014). CONCLUSIONS: Our data show that the occurrence of fish-shaped RBC is suggestive of a pathologic condition, especially IDA, HE, vitamin B12 or folate deficiency, primary mylofibrosis or LD, and is significantly associated with severity of anaemia.

Quantitative microscopy and nanoscopy of sickle red blood cells performed by wide field digital interferometry.

We have applied wide-field digital interferometry (WFDI) to examine the morphology and dynamics of live red blood cells (RBCs) from individuals who suffer from sickle cell anemia (SCA), a genetic disorder that affects the structure and mechanical properties of RBCs. WFDI is a noncontact, label-free optical microscopy approach that can yield quantitative thickness profiles of RBCs and measurements of their membrane fluctuations at the nanometer scale reflecting their stiffness. We find that RBCs from individuals with SCA are significantly stiffer than those from a healthy control. Moreover, we show that the technique is sensitive enough to distinguish classes of RBCs in SCA, including sickle RBCs with apparently normal morphology, compared to the stiffer crescent-shaped sickle RBCs. We expect that this approach will be useful for diagnosis of SCA and for determining efficacy of therapeutic agents.

Metabolic depletion decreases the aggregability of erythrocytes.

We studied the influence of metabolic depletion on red blood cell (RBC) aggregability, which is a determinant of blood flow. Heparinized blood was stored at room temperature for 0, 24, and 48 h. RBCs were washed twice and resuspended in Tris-buffer containing 3% dextran 70 (hematocrit 30%). Suspension viscosities were measured at 37 °C and shear rates of 37.6 and 0.1 s(-1), RBC aggregability was analysed by the sedimentation rate, direct microscopic visualization and a Myrenne aggregometer. RBCs in autologous plasma showed an increasing echinocytic shape transformation, which was reversible in buffer. The viscosities of RBC suspensions in buffer remained unchanged at both low (0.1 s(-1)) and high shear rate (37.6 s(-1)), the latter result indicating an unchanged RBC deformability. RBC aggregability decreased: The RBC sedimentation rates were 40.7 ± 5.0, 29.3 ± 13.4, and 13.3 ± 11.2 mm/h (p < 0.001) at 0, 24, and 48 h, respectively, which correlated well with the visual aggregability index and the Myrenne aggregation parameters M and M1. We conclude that metabolic depletion for 48 h leads to RBC swelling and a reversible echinocytic shape transformation. These ATP-depleted, but normally shaped RBCs had a decreased aggregability. In contrast to all other methods used, low shear viscosity was inaccurate and should not be used to test RBC aggregability.

Novel hemoplasma species detected in free-ranging sika deer (Cervus nippon).

Hemoplasma infections in wild ungulates have not been reported yet in Japan. We examined presence of hemoplasmas in blood samples collected from 147 sika deer (Cervus nippon) in the Iwate prefecture by real-time PCR, and found 13 (9%) were positive. Almost entire region of the 16S rRNA gene of the representative strains from positive samples was amplified by conventional PCR. The nucleotide sequences of the 16S rRNA gene were further determined and compared with those of other hemoplasmas. Our examinations 1st revealed the presence of 2 distinct hemoplasma species in sika deer, which are previously not described. One of them was closely related to M. ovis by the 16S rRNA sequence analysis, but was found distinct by comparison of the RNase P RNA gene sequences. Pathogenicity of these two hemoplasma species in sika deer is currently unknown.

Discriminant value of % microcytic/% hypochromic ratio in the differential diagnosis of microcytic anemia.

BACKGROUND: % Microcytic and % hypochromic red cells show an inverse trend in beta thalassemia trait and in iron deficiency anemia (IDA). The aim of this study was to assess the discriminant value of % microcytic/% hypochromic ratio, as can be obtained from the Advia 2120 (Siemens) analyzer, in the differential diagnosis of microcytic anemia, compared to the conventional England and Fraser, Green and King, Mentzer, Ricerca, Shine and Lal and Srivastava indices. METHODS: Receiver operating characteristic (ROC) curves, sensitivity, specificity and Youden index of indices were calculated for a set of 148 IDA patients and 170 beta thalassemia carriers. RESULTS: The Green and King index showed the best area under the curve (AUC, 0.962), with a Youden index of 80.9%; the % microcytic/% hypochromic ratio ranked second (AUC 0.943), with a Youden index of 76.3% and a sensitivity of 99.2%. CONCLUSIONS: This index, with high sensitivity for beta thalassemia screening, can be a useful tool in the differential diagnosis of microcytic anemia, so samples can be chosen for HbA(2) analysis, to confirm the presumptive diagnosis of the disease. As specificity is 77.1%, we assume it will be 22.9% for false positives, but these samples will have HbA(2) values within the reference range, for correct diagnosis.

Rate of deoxygenation modulates rheologic behavior of sickle red blood cells at a given mean corpuscular hemoglobin concentration.

Although the mean corpuscular hemoglobin concentration (MCHC) plays a dominant role in the rheologic behavior of deoxygenated density-defined sickle red blood cells (SS RBCs), previous studies have not explored the relationship between the rate of deoxygenation and the bulk viscosity of SS RBCs at a given MCHC. In the present study, we have subjected density-defined SS classes (i.e., medium-density SS4 and dense SS5 discocytes) to varying deoxygenation rates. This approach has allowed us to minimize the effects of SS RBC heterogeneity and investigate the effect of deoxygenation rates at a given MCHC. The results show that the percentages of granular cells, classic sickle cells and holly leaf forms in deoxygenated samples are significantly influenced by the rate of deoxygenation and the MCHC of a given discocyte subpopulation. Increasing the deoxygenation rate using high K+ medium (pH 6.8), results in a greater percentage of granular cells in SS4 suspensions, accompanied by a pronounced increase in the bulk viscosity of these cells compared with gradually deoxygenated samples (mainly classic sickle cells and holly leaf forms). The effect of MCHC becomes apparent when SS5 dense cells are subjected to varying deoxygenation rates. At a given deoxygenation rate, SS5 dense discocytes show a greater increase in the percentage of granular cells than that observed for SS4 RBCs. Also, at a given deoxygenation rate, SS5 suspensions exhibit a higher viscosity than SS4 suspensions with fast deoxygenation resulting in maximal increase in viscosity. Although MCHC is the main determinant of SS RBC rheologic behavior, these studies demonstrate for the first time that at a given MCHC, the rate of deoxygenation (hence HbS polymerization rates) further modulates the rheologic behavior of SS RBCs. Thus, both MCHC and the deoxygenation rate may contribute to microcirculatory flow behavior of SS RBCs.

Classification of red blood cells as normal, sickle, or other abnormal, using a single image analysis feature.

Sickle cell anemia is a disease for which there is currently no effective treatment. One method of evaluating clinical status is the counting of cell types based on morphology. There is a need for a rapid, reproducible method, superior to human inspection, for classification of these cells. Quantitative digital-image analysis is being applied to this need. Blood from 24 patients with sickle cell anemia (SS) and SC disease and ten hematologically normal volunteers (AA) was stressed by bubbling with nitrogen. One hundred fifty cells were analyzed from each sickle specimen, and 100 were analyzed from each nonsickle specimen. Expert observers classified each cell as normal (N), sickle (S), or other abnormal (A). Cells were analyzed with a custom, high-resolution image-analysis instrument. A total of 42 features including metric, optical density-derived, and textural features were extracted. The metric feature Form Factor (4 pi Area/Perimeter2) was selected by recursive partitioning analysis as the sole feature needed for segregating cells into the classes of N, A, and S. The agreement of automated classification (using cutpoints determined by recursive partitioning analysis) with a human expert for specimens from individuals with sickle cell anemia was 89% for N-, 73% for A-, and 92% for S-classified cells. For specimens from AA individuals, the agreement was 92% for N and 76% for A. For specimens from individuals with sickle cell anemia, rates of agreement between two human experts were compared and found to be 86% for N, 84% for A, and 80% for S. For specimens from AA individuals, the agreement was 90% for N and 87% for A.

Differential susceptibility of type III erythrocytes of paroxysmal nocturnal hemoglobinuria to lysis mediated by complement and perforin.

Previous reports have suggested that a 65 kDa membrane protein, termed homologous restriction factor (HRF), in addition to protecting erythrocytes (E) against lysis by homologous complement (C), may also be involved in protecting cytolytic lymphocytes against lysis mediated by a pore-forming protein (PFP/perforin), one of their own lytic mediators. Here, we used HRF-deficient type III E of patients with paroxysmal nocturnal hemoglobinuria (PNH) to study their susceptibility to lysis mediated by homologous C and perforin, and compared it with lysis of HRF-bearing control or PNH type I E. We show that type III E of PNH patients are indeed more susceptible to lysis mediated by homologous C than control or type I E, but they are as susceptible to perforin-mediated lysis as type I E. In addition, all human E (type I or III) tested here are equally susceptible to lysis mediated by either human (homologous) or murine (heterologous) perforin. By immunoblot analysis, we confirm that type III E, in contrast to type I E, were deficient in the 65 kDa HRF. These results support the notion that homologous species restriction is seen in the C- but not in the lymphocyte perforin-system and argue against an active participation of HRF in protecting cells from perforin-mediated lysis.

Nondiscocytic erythrocytes in myalgic encephalomyelitis.

Blood samples from 102 volunteers who believed they suffered from myalgic encephalomyelitis were photographed in a scanning electron microscope at 500x. All identifiable cells were counted and classified on the basis of their shape. The frequency of each cell shape was expressed as a percentage of the total number of cells counted in the sample. The resulting data were compared with that from 52 healthy controls and 99 cases of multiple sclerosis which had been selected randomly by a computer from a panel of 229 cases in a concurrent study. Samples from subjects with myalgic encephalomyelitis had the lowest percentages of normal red cells and the highest incidence of cup forms. The results provide evidence that myalgic encephalomyelitis has an organic cause. Quantitative analysis of red cell shape may assist in the diagnosis of myalgic encephalomyelitis.

Membrane studies on rod-shaped red cells in hereditary elliptocytosis: least haemolysis and normal sodium influx with decreased membrane lipids.

The clinical features of hereditary elliptocytosis were studied in 25 cases and compared with 20 normal individuals. Based on morphological features, these patients were classified into two groups: those with a rod-shaped type of elliptocytosis (nine cases), and those with a non-rod type (16 cases). Most of the cases with overt haemolysis were detected among cases of the non-rod type. Overt haemolysis typically tended to be accompanied by stomatocytic changes, which appear to be superimposed on elliptic transformation of the red cells. Sodium influx increased in eight of nine HE patients with overt haemolysis, but in none of the HE with rod-shaped red cells. Significant quantitative and qualitative differences in red cell membrane lipids were observed between the two types of HE. No spectrin abnormalities in domain composition, dimer-dimer association, and thermal stability were observed in 11 of the HE patients studied, including four cases with overt haemolysis.

Pseudosickling of hemoglobin Setif.

Hemoglobin Setif produces pseudosickling of red cells in vitro; the nature of the process and the conditions that "trigger" it are unknown. Studies of red cells, hemolysates, purified hemoglobin solutions, and artificial mixtures of Hb A and Setif suggest that pseudosickling is produced by intracellular crystallization of insoluble hemoglobin. Increased tonicity of the suspending medium accentuates the process, probably by causing a rise in intracellular hemoglobin concentration. If precipitates from A/Setif mixtures are analyzed, they always contain Hb A, suggesting an unusual mechanism for the process. Despite the fact that osmolality in the renal medulla is similar to that which produces pseudosickling in vitro, carriers do not have renal dysfunction of the type found in patients with sickle cell disease.

The influence of serum from patients infected with Mycoplasma pneumoniae on the osmotic property of normal human erythrocytes.

Venous blood sera from children infected with Mycoplasma pneumoniae (MP) responded with a high titer to MP antibodies, affected healthy normal erythrocytes in vitro, and gave rise to changes in erythrocyte osmotic fragility. When serum was inactivated at 56 degrees C for 30 min or preincubated with anti-human C1 esterase inhibitor, the changes in the osmotic properties were suppressed at the lower level. The total sialic acid content and the intracellular ATP concentration of the treated erythrocytes were analyzed.

Alterations in red blood cell morphology during a 500 metre dive.

Following compression to 500 m in a simulated chamber dive, the blood samples of the six divers were all found to contain several types of non-discoid erythrocytes. Compression to this depth induced a pressure stress and sensitisation in a proportion of each divers' erythrocyte population. Long in vitro decompression procedures further stressed these red cells and resulted in additional morphological changes. The formation of stomatocytes was increased by an acidic-buffered fixative, conversely, an alkaline medium caused echinocytosis. Cell counts of each morphological cell type showed that as echinocyte stage III & IV numbers were reduced a simultaneous decrease in mean haemoglobin concentration occurred. Decompressions of blood samples for routine haematology should be at a rate of 3 m/min so as to be completed within four hours from venesection. Hyperbaric exposure time explicitly influence these red cell anomalies and development of a subclinical anaemia.

Abnormality of phospholipid transverse diffusion in sickle erythrocytes.

We have used spin-labeled analogues of phosphatidylcholine, phosphatidylethanolamine, and phosphatidylserine to compare the transverse diffusion rates of lipids in normal and sickle erythrocytes. The beta-chain of the spin-labeled lipids was a short chain (five carbons) providing the spin labels with a relative water solubility, and hence permitting their rapid incorporation into cell membranes. The orientation of the labeled lipids in the membranes was assayed by selective chemical reduction of the nitroxide labels embedded in the outer leaflet. We have found that all three spin-labeled phospholipids are initially incorporated in the outer leaflet. Upon incubation at 4 degrees C the aminophospholipids, not the phosphatidylcholine, diffuse toward the inner leaflet within 3 h. The transverse diffusion rate of aminophospholipids is reduced by 41% (phosphatidylserine) and 14% (phosphatidylethanolamine) in homozygote sickle cells (SS) when compared with normal cells (AA) or heterozygote cells (AS or SC). At equilibrium the asymmetric distribution of spin-labeled phospholipids resulting from this selective diffusion is also reduced in SS cells when compared with AA, SC, or AS cells. This reduced asymmetry was not found in a reticulocyte-rich blood sample (hemoglobin A), indicating that the age of the cell cannot be responsible for this phenomenon. Moreover, because at low temperatures the sickling process does not occur, the observed perturbations in phospholipid organization reflect preexisting membrane abnormalities in sickle cells. Ghosts loaded with ATP give the same results. Varying the concentration of intracellular calcium had no effect on lipid diffusion, except at very high free calcium concentrations (3 microM) when diffusion was practically abolished. We suggest that membrane protein alterations may be part of the explanation of the observed abnormalities.