skip to main content

Title: A Deep Learning Framework for Sickle Cell Disease Microfluidic Biomarker Assays
Introduction: Vaso-occlusive crises (VOCs) are a leading cause of morbidity and early mortality in individuals with sickle cell disease (SCD). These crises are triggered by sickle red blood cell (sRBC) aggregation in blood vessels and are influenced by factors such as enhanced sRBC and white blood cell (WBC) adhesion to inflamed endothelium. Advances in microfluidic biomarker assays (i.e., SCD Biochip systems) have led to clinical studies of blood cell adhesion onto endothelial proteins, including, fibronectin, laminin, P-selectin, ICAM-1, functionalized in microchannels. These microfluidic assays allow mimicking the physiological aspects of human microvasculature and help characterize biomechanical properties of adhered sRBCs under flow. However, analysis of the microfluidic biomarker assay data has so far relied on manual cell counting and exhaustive visual morphological characterization of cells by trained personnel. Integrating deep learning algorithms with microscopic imaging of adhesion protein functionalized microfluidic channels can accelerate and standardize accurate classification of blood cells in microfluidic biomarker assays. Here we present a deep learning approach into a general-purpose analytical tool covering a wide range of conditions: channels functionalized with different proteins (laminin or P-selectin), with varying degrees of adhesion by both sRBCs and WBCs, and in both normoxic and hypoxic environments. Methods: Our neural networks were trained on a repository of manually labeled SCD Biochip microfluidic biomarker assay whole channel images. Each channel contained adhered cells pertaining to clinical whole blood under constant shear stress of 0.1 Pa, mimicking physiological levels in post-capillary venules. The machine learning (ML) framework consists of two phases: Phase I segments pixels belonging to blood cells adhered to the microfluidic channel surface, while Phase II associates pixel clusters with specific cell types (sRBCs or WBCs). Phase I is implemented through an ensemble of seven generative fully convolutional neural networks, and Phase II is an ensemble of five neural networks based on a Resnet50 backbone. Each pixel cluster is given a probability of belonging to one of three classes: adhered sRBC, adhered WBC, or non-adhered / other. Results and Discussion: We applied our trained ML framework to 107 novel whole channel images not used during training and compared the results against counts from human experts. As seen in Fig. 1A, there was excellent agreement in counts across all protein and cell types investigated: sRBCs adhered to laminin, sRBCs adhered to P-selectin, and WBCs adhered to P-selectin. Not only was the approach able to handle surfaces functionalized with different proteins, but it also performed well for high cell density images (up to 5000 cells per image) in both normoxic and hypoxic conditions (Fig. 1B). The average uncertainty for the ML counts, obtained from accuracy metrics on the test dataset, was 3%. This uncertainty is a significant improvement on the 20% average uncertainty of the human counts, estimated from the variance in repeated manual analyses of the images. Moreover, manual classification of each image may take up to 2 hours, versus about 6 minutes per image for the ML analysis. Thus, ML provides greater consistency in the classification at a fraction of the processing time. To assess which features the network used to distinguish adhered cells, we generated class activation maps (Fig. 1C-E). These heat maps indicate the regions of focus for the algorithm in making each classification decision. Intriguingly, the highlighted features were similar to those used by human experts: the dimple in partially sickled RBCs, the sharp endpoints for highly sickled RBCs, and the uniform curvature of the WBCs. Overall the robust performance of the ML approach in our study sets the stage for generalizing it to other endothelial proteins and experimental conditions, a first step toward a universal microfluidic ML framework targeting blood disorders. Such a framework would not only be able to integrate advanced biophysical characterization into fast, point-of-care diagnostic devices, but also provide a standardized and reliable way of monitoring patients undergoing targeted therapies and curative interventions, including, stem cell and gene-based therapies for SCD. Disclosures Gurkan: Dx Now Inc.: Patents & Royalties; Xatek Inc.: Patents & Royalties; BioChip Labs: Patents & Royalties; Hemex Health, Inc.: Consultancy, Current Employment, Patents & Royalties, Research Funding.  more » « less
Award ID(s):
Author(s) / Creator(s):
; ; ; ; ; ; ; ;
Date Published:
Journal Name:
Supplement 1
Page Range / eLocation ID:
15 to 16
Medium: X
Sponsoring Org:
National Science Foundation
More Like this
  1. Fadlelmola, Faisal Mohamed (Ed.)
    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 R 2 = 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. 
    more » « less
  2. Abstract

    Sickle cell disease (SCD) is a recessive genetic blood disorder exhibiting abnormal blood rheology. Polymerization of sickle hemoglobin, due to a point mutation in theβ‐globingene of hemoglobin, results in aberrantly adhesive and stiff red blood cells (RBCs). Hemolysis, abnormal RBC adhesion, and abnormal blood rheology together impair endothelial health in people with SCD, which leads to cumulative systemic complications. Here, we describe a microfluidic assay combined with a micro particle image velocimetry technique for the integrated in vitro assessment of whole blood viscosity (WBV) and RBC adhesion. We examined WBV and RBC adhesion to laminin (LN) in microscale flow in whole blood samples from 53 individuals with no hemoglobinopathies (HbAA, N = 10), hemoglobin SC disease (HbSC, N = 14), or homozygous SCD (HbSS, N = 29) with mean WBV of 4.50 cP, 4.08 cP, and 3.73 cP, respectively. We found that WBV correlated with RBC count and hematocrit in subjects with HbSC or HbSS. There was a significant inverse association between WBV and RBC adhesion under both normoxic and physiologically hypoxic (SpO2of 83%) tests, in which lower WBV associated with higher RBC adhesion to LN in subjects with HbSS. Low WBV has been found by others to associate with endothelial activation. Altered WBV and abnormal RBC adhesion may synergistically contribute to the endothelial damage and cumulative pathophysiology of SCD. These findings suggest that WBV and RBC adhesion may serve as clinically relevant biomarkers and endpoints in assessing emerging targeted and curative therapies in SCD.

    more » « less
  3. Abstract

    Sickle cell disease (SCD) is the most prevalent inherited blood disorder in the world. But the clinical manifestations of the disease are highly variable. In particular, it is currently difficult to predict the adverse outcomes within patients with SCD, such as, vasculopathy, thrombosis, and stroke. Therefore, for most effective and timely interventions, a predictive analytic strategy is desirable. In this study, we evaluate the endothelial and prothrombotic characteristics of blood outgrowth endothelial cells (BOECs) generated from blood samples of SCD patients with known differences in clinical severity of the disease. We present a method to evaluate patient‐specific vaso‐occlusive risk by combining novel RNA‐seq and organ‐on‐chip approaches. Through differential gene expression (DGE) and pathway analysis we find that BOECs from SCD patients exhibit an activated state through cell adhesion molecule (CAM) and cytokine signaling pathways among many others. In agreement with clinical symptoms of patients, DGE analyses reveal that patient with severe SCD had a greater extent of endothelial activation compared to patient with milder symptoms. This difference is confirmed by performing qRT‐PCR of endothelial adhesion markers like E‐selectin, P‐selectin, tissue factor, and Von Willebrand factor. Finally, the differential regulation of the proinflammatory phenotype is confirmed through platelet adhesion readouts in our BOEC vessel‐chip. Taken together, we hypothesize that these easily blood‐derived endothelial cells evaluated through RNA‐seq and organ‐on‐chips may serve as a biotechnique to predict vaso‐occlusive episodes in SCD patients and will ultimately allow better therapeutic interventions.

    more » « less
  4. Abstract

    White blood cell (WBC) differential counting is an established clinical routine to assess patient immune system status. Fluorescent markers and a flow cytometer are required for the current state‐of‐the‐art method for determining WBC differential counts. However, this process requires several sample preparation steps and may adversely disturb the cells. We present a novel label‐free approach using an imaging flow cytometer and machine learning algorithms, where live, unstained WBCs were classified. It achieved an average F1‐score of 97% and two subtypes of WBCs, B and T lymphocytes, were distinguished from each other with an average F1‐score of 78%, a task previously considered impossible for unlabeled samples. We provide an open‐source workflow to carry out the procedure. We validated the WBC analysis with unstained samples from 85 donors. The presented method enables robust and highly accurate identification of WBCs, minimizing the disturbance to the cells and leaving marker channels free to answer other biological questions. It also opens the door to employing machine learning for liquid biopsy, here, using the rich information in cell morphology for a wide range of diagnostics of primary blood. © 2019 The Authors.Cytometry Part Apublished by Wiley Periodicals, Inc. on behalf of International Society for Advancement of Cytometry.

    more » « less
  5. Summary

    Individuals with sickle cell disease (SCD) have persistently elevated thrombin generation that results in a state of systemic hypercoagulability. Antithrombin‐III (ATIII), an endogenous serine protease inhibitor, inhibits several enzymes in the coagulation cascade, including thrombin. Here, we utilize a biomimetic microfluidic device to model the morphology and adhesive properties of endothelial cells (ECs) activated by thrombin and examine the efficacy of ATIII in mitigating the adhesion of SCD patient‐derived red blood cells (RBCs) and EC retraction. Microfluidic devices were fabricated, seeded with ECs, and incubated under physiological shear stress. Cells were then activated with thrombin with or without an ATIII pretreatment. Blood samples from subjects with normal haemoglobin (HbAA) and subjects with homozygous SCD (HbSS) were used to examine RBC adhesion to ECs. Endothelial cell surface adhesion molecule expression and confluency in response to thrombin and ATIII treatments were also evaluated. We found that ATIII pretreatment of ECs reduced HbSS RBC adhesion to thrombin‐activated endothelium. Furthermore, ATIII mitigated cellular contraction and reduced surface expression of von Willebrand factor and vascular cell adhesion molecule‐1 (VCAM‐1) mediated by thrombin. Our findings suggest that, by attenuating thrombin‐mediated EC damage and RBC adhesion to endothelium, ATIII may alleviate the thromboinflammatory manifestations of SCD.

    more » « less