Images Of Red Blood Cells Research Articles

An analysis of decipherable red blood cell abnormality detection under federated environment leveraging XAI incorporated deep learning

In recent times, automated detection of diseases from pathological images leveraging Machine Learning (ML) models has become fairly common, where the ML models learn detecting the disease by identifying biomarkers from the images. However, such an approach requires the models to be trained on a vast amount of data, and healthcare organizations often tend to limit access due to privacy concerns. Consequently, collecting data for traditional centralized training becomes challenging. These privacy concerns can be handled by Federation Learning (FL), which builds an unbiased global model from local models trained with client data while maintaining the confidentiality of local data. Using FL, this study solves the problem of centralized data collection by detecting deformations in images of Red Blood Cells (RBC) in a decentralized way. To achieve this, RBC data is used to train multiple Deep Learning (DL) models, and among the various DL models, the most efficient one is considered to be used as the global model inside the FL framework. The FL framework works by copying the global model’s weight to the client’s local models and then training the local models in client-specific devices to average the weights of the local model back to the global model. In the averaging process, direct averaging is performed and alongside, weighted averaging is also done to weigh the individual local model’s contribution according to their performance, keeping the FL framework immune to the effects of bad clients and attacks. In the process, the data of the client remains confidential during training, while the global model learns necessary information. The results of the experiments indicate that there is no significant difference in the performance of the FL method and the best-performing DL model, as the best-performing DL model reaches an accuracy of 96% and the FL environment reaches 94%-95%. This study shows that the FL technique, in comparison to the classic DL methodology, can accomplish confidentiality secured RBC deformation classification from RBC images without substantially diminishing the accuracy of the categorization. Finally, the overall validity of the classification results has been verified by employing GradCam driven Explainable AI techniques.

Compressed sensing reconstruction for high-SNR, rapid dissolved 129Xe gas exchange MRI.

Three-dimensional hyperpolarized 129Xe gas exchange imaging suffers from low SNR and long breath-holds, which could be improved using compressed sensing (CS). The purpose of this work was to assess whether gas exchange ratio maps are quantitatively preserved in CS-accelerated dissolved-phase 129Xe imaging and to investigate the feasibility of CS-dissolved 129Xe imaging with reduced-cost natural abundance (NA) xenon. 129Xe gas exchange imaging was performed at 1.5 T with a multi-echo spectroscopic imaging sequence. A CS reconstruction with an acceleration factor of 2 was compared retrospectively with conventional gridding reconstruction in a cohort of 16 healthy volunteers, 5 chronic obstructive pulmonary disease patients, and 23 patients who were hospitalized following COVID-19 infection. Metrics of comparison included normalized mean absolute error, mean gas exchange ratio, and red blood cell (RBC) image SNR. Dissolved 129Xe CS imaging with NA xenon was assessed in 4 healthy volunteers. CS reconstruction enabled acquisition time to be halved, and it reduced background noise. Median RBC SNR increased from 6 (2-18) to 11 (2-100) with CS, and there was strong agreement between CS and gridding mean ratio map values (R2 = 0.99). Image fidelity was maintained for gridding RBC SNR > 5, but below this, normalized mean absolute error increased nonlinearly with decreasing SNR. CS increased the mean SNR of NA 129Xe images 3-fold. CS reconstruction of dissolved 129Xe imaging improved image quality with decreased scan time, while preserving key gas exchange metrics. This will benefit patients with breathlessness and/or low gas transfer and shows promise for NA-dissolved 129Xe imaging.

Evaluation of blood evacuation efficiency using 99m Tc-RBC imaging: a comparative study of exsanguination tourniquet rings and Esmarch bandages.

The objective of this study is to compare the effectiveness of the Esmarch bandage and exsanguination tourniquet rings (ETRs) in blood evacuation procedures using a controlled intra-subject design involving healthy volunteers. A total of 20 healthy adult volunteers (12 males, 8 females) were recruited from the community. Participants underwent blood evacuation procedures on both legs, using the Esmarch bandage on one leg and the ETR on the other. The order of the procedures was randomized. Blood evacuation time, overall blood evacuation rate, and calf blood evacuation rate were measured using 99m Tc-labeled red blood cell imaging. Paired t -tests were conducted to compare the effectiveness of the two methods. The ETRs demonstrated a significantly faster blood evacuation time compared to the Esmarch bandage (mean difference = -41.72 s, P < 0.0001). The overall blood evacuation rate was slightly higher for the ETRs (mean difference = 1.717%), though not statistically significant ( P = 0.3680). The calf blood evacuation rate was significantly higher for the ETRs (mean difference = 6.86%, P = 0.0225). No significant discomfort or adverse reactions were reported by any participants. ETRs are more efficient in terms of blood evacuation time and calf blood evacuation rate compared to the Esmarch bandage, without causing significant discomfort or adverse reactions. These findings suggest that ETRs could be a preferable option in clinical settings for blood evacuation procedures.

Eliminating the systematic error of spatial light interference microscopy by the wavefront comparison iterative algorithm

Spatial light interference microscopy (SLIM) is a unique quantitative phase imaging technology. Since the phase of direct light is simply assumed to be 0, it has a non-negligible systematic error. Based on the theory of partially coherent SLIM, here we propose a wavefront comparison iterative (WCI) algorithm to eliminate this error for accurate phase imaging. In this paper, we clarified the existence of this systematic error by phase imaging of oil-immersed polystyrene microspheres, and we demonstrated the feasibility of the WCI algorithm by phase imaging of red blood cells. Furthermore, through computer simulation, we have confirmed that for phase imaging of large adherent cells, if the SLIM system compresses the phase of direct light sufficiently small, the WCI algorithm will be feasible too.

Revolutionizing malaria diagnosis: deep learning-powered detection of parasite-infected red blood cells

Malaria is a significant global health issue, responsible for the highest rates of morbidity and mortality globally. This paper introduces a very effective and precise convolutional neural network (CNN) method that employs advanced deep learning techniques to automate the detection of malaria in images of red blood cells (RBC). Furthermore, we present an emerging and efficient deep learning method for differentiating between cells infected with malaria and those that are not infected. To thoroughly evaluate the efficiency of our approach, we do a meticulous assessment that involves comparing different deep learning models, such as ResNet-50, MobileNet-v2, and Inception-v3, within the domain of malaria detection. Additionally, we conduct a thorough comparison of our proposed approach with current automated methods for malaria identification. An examination of the most current techniques reveals differences in performance metrics, such as accuracy, specificity, sensitivity, and F1 score, for diagnosing malaria. Moreover, compared to existing models for malaria detection, our method is the most successful, achieving an accurate score of 1.00 in all statistical matrices, confirming its promise as a highly efficient tool for automating malaria detection.

Efficient deep learning-based approach for malaria detection using red blood cell smears

Malaria is an extremely malignant disease and is caused by the bites of infected female mosquitoes. This disease is not only infectious among humans, but among animals as well. Malaria causes mild symptoms like fever, headache, sweating and vomiting, and muscle discomfort; severe symptoms include coma, seizures, and kidney failure. The timely identification of malaria parasites is a challenging and chaotic endeavor for health staff. An expert technician examines the schematic blood smears of infected red blood cells through a microscope. The conventional methods for identifying malaria are not efficient. Machine learning approaches are effective for simple classification challenges but not for complex tasks. Furthermore, machine learning involves rigorous feature engineering to train the model and detect patterns in the features. On the other hand, deep learning works well with complex tasks and automatically extracts low and high-level features from the images to detect disease. In this paper, EfficientNet, a deep learning-based approach for detecting Malaria, is proposed that uses red blood cell images. Experiments are carried out and performance comparison is made with pre-trained deep learning models. In addition, k-fold cross-validation is also used to substantiate the results of the proposed approach. Experiments show that the proposed approach is 97.57% accurate in detecting Malaria from red blood cell images and can be beneficial practically for medical healthcare staff.

White light diffraction phase microscopy for imaging of red blood cells for different storage times

In this study, the effects of different storage times on the surface morphology of red blood cells (RBCs) were investigated using white light diffraction phase microscopy (WDPM). Blood samples collected from 10 volunteer and stored for 56 days, were imaged on WDPM at every 7 days without any sample preparation. To obtain the phase profiles of RCBs, first the sample and then the reference interferograms were obtained from the experimental setup. Then, surface profiles were calculated from these interferogram images using Fourier transform (FT). With the experiment performed every 7 days, 10 RBC phase information were obtained from each sample and surface profiles were created. From these profiles, 7 parameters related with RBC morphology (average cell thickness—ACT; mean corpuscular volume—MCV; projected surface area—PSA; total surface area—SA; diameter—D; mean corpuscular haemoglobin—MCH; surface area to volume ratio - SAV) have been calculated. Therefore, changes in the morphology of RBCs during storage were evaluated quantitatively. Additionally, phase calibration target was used to confirm the accuracy of our experimental setup system. From the interferogram images, the depth of the phase target (GRP 9 and element 3) was obtained as 309 nm, in accordance with the produced depth. By this way, the reliability of the WDPM setup is demonstrated. This study suggests that the morphology of RBCs can be quantitatively obtained in a reliable manner at higher resolution with WDPM without sample preparation.

Study on Fine-Grained Visual Classification of Low-Resolution Urinary Erythrocyte.

The morphological analysis test item of urine red blood cells is referred to as "extracorporeal renal biopsy," which holds significant importance for medical department testing. However, the accuracy of existing urine red blood cell morphology analyzers is suboptimal, and they are not widely utilized in medical examinations. Challenges include low image spatial resolution, blurred distinguishing features between cells, difficulty in fine-grained feature extraction, and insufficient data volume. This article aims to improve the classification accuracy of low-resolution urine red blood cells. This paper proposes a super-resolution method based on category-aware loss and an RBC-MIX data enhancement approach. It optimizes the cross-entropy loss to maximize the classification boundary and improve intra-class tightness and inter-class difference, achieving fine-grained classification of low-resolution urine red blood cells. Experimental outcomes demonstrate that with this method, an accuracy rate of 97.8% can be achieved for low-resolution urine red blood cell images. This algorithm attains outstanding classification performance for low-resolution urine red blood cells with only category labels required. This method can serve as a practical reference for urine red blood cell morphology examination items.

A dataset of digital holograms of normal and thalassemic cells

Digital holographic microscopy (DHM) is an intriguing medical diagnostic tool due to its label-free and quantitative nature, providing high-contrast images of phase samples. By capturing both intensity and phase information, DHM enables the numerical reconstruction of quantitative phase images. However, the lateral resolution is limited by the diffraction limit, which prompted the recent suggestion of microsphere-assisted DHM to enhance the DHM resolution straightforwardly. The use of such a technique as a medical diagnostic tool requires testing and validation of the proposed assays to prove their feasibility and viability. This paper publishes 760 and 609 microsphere-assisted DHM images of normal and thalassemic red blood cells obtained from a normal and thalassemic male individual, respectively.

A Novel Deep Learning Approach to Malaria Disease Detection on Two Malaria Datasets

Malaria is a contagious febrile disease transmitted to humans by the bite of female mosquitoes. It is important to diagnose this disease in a short period of time. Finding the mathematically best numerical solution to a particular problem is the most important issue for most departments. In deep learning-based systems developed, the difference between the real data and the predicted result of the model is measured using loss functions. To minimize the error rate in the predictions during the training process of deep learning models, the weight values used in the model should be updated. This update process has a significant effect on the model prediction result. This article presents a new deep learning-based malaria detection method that will help diagnose malaria in a short time. A new 21-layer Convolutional Neural Network (CNN) model is designed and proposed to describe infected and uninfected thin red blood cell images. By using thin red blood cell sample images, 95% accuracy was achieved with Nadam and RMSprop optimization techniques. The results obtained show the efficiency of the proposed method according to each optimization algorithm.

Optofluidic-based cell multi-axis controllable rotation and 3D surface imaging

The controlled rotation of individual cells plays a crucial role in enabling three-dimensional multi-angle observation of cellular structure, function, and dynamic processes. Reported cell rotation techniques often struggle to strike a balance between high precision and simple control, and they exhibit limited control flexibility, typically achieving only uniaxial cell rotation. In this study, we propose a cell rotation technique in three dimensions based on optofluidics, which utilizes optical tweezers to immobilize the cell and exploits the asymmetry of the surrounding flow to drive cell rotation. By adjusting the focal position of the optical tweezers, cells can be positioned within various flow profiles, enabling control of the rotation speed, rotation direction, and rotation axis of cells. This approach simplifies the manipulation procedure, achieving desirable control precision and greater rotation flexibility. Using our approach, multi-angle surface imaging projections of target cells can be rapidly obtained, followed by capturing the cell contour data from the images. By combining the cell contour data with corresponding angular position information, we have reconstructed the 3D surface of the target cell. We have employed this technique in experiments for the analysis of red blood cell morphology. Based on the constructed 3D surface images of diverse-shaped red blood cells, we quantified structural parameters including cell surface area, volume, sphericity, and surface roughness, which demonstrates the potential application of this cell rotation technique for cellular morphological analysis.

A novel transfer learning-based model for diagnosing malaria from parasitized and uninfected red blood cell images

Malaria represents a potentially fatal communicable illness triggered by the Plasmodium parasite. This disease is transmitted to humans through the bites of Anopheles mosquitoes that carry the infection. This disease has significant and devastating consequences on the health systems of fragile countries, particularly in sub-Saharan Africa. Malaria affects red blood cells by invading and replicating within them, destroying them, and releasing toxic byproducts into the bloodstream. The parasite’s ability to stick and modify the surface of red blood cells can cause them to become sticky, obstructing blood flow in vital organs such as the brain and spleen. Therefore, efficient approaches for the early detection of malaria are critical to saving patients’ lives. The main aim of this study is to develop an efficient model for early malaria diagnosis. We used malaria images based on parasitized and uninfected red blood cells for the study experiments. We applied neural network-based approaches such as Neural Search Architecture Network (NASNet) and compared its performance with machine learning techniques. Moreover, we proposed a novel NNR (NASNet Random forest) method for feature engineering. The proposed NNR approach first extracts spatial features from input malaria images, then class prediction probability features are extracted from these spatial features. The feature set obtained from the data extraction trains machine learning models. Our comprehensive experiments show that the support vector machine outperformed state-of-the-art models, achieving a high-performance score of 99% and having an inference time near 0.025 s. We validated the performance using k-fold cross-validation and optimized the hyperparameters through tuning. Our proposed research has improved the early diagnosis of malaria and can assist medical specialists in reducing the mortality rate.

Correlation of Sysmex-XN9000 and CellaVision Advanced RBC software on anisocytosis

ABSTRACTObjectiveThis study aims to evaluate the consistency of heterogeneity degree of erythrocyte volume parameters between the blood automated analyzer Sysmex-XN9000 and the advanced red blood cell software CellaVisionDI-60.Method500 blood samples of volunteers were analyzed by Sysmex-XN9000 and CellaVision-DI60. The sensitivity, specificity, positive predictive value, negative predictive value, false positive rate, and false negative rate were evaluated. The consistency of all parameters was tested.ResultTaking the standard RBC group as the control group, the RBC parameters of the macrocytic and the microcytic group were compared. There was a statistical difference between the groups. ROC curve analysis showed that the best cutoff value of microcytic and of macrocytic affecting MCV were 4.1% and 5.7%, respectively. The best cutoff value of anisocytosis was 15.0%. The correlation coefficient between anisocytosis and red blood cell distribution width (RDW-CV) was 0.756. The sensitivity, specificity, positive predictive value and coincidence rate of anisocytosis were high. The false negative rate was 10.0%, and the false positive rate was 7.4%.ConclusionAll parameters of the degree of heterogeneity have good accuracy and consistency in the two instruments. Anisocytosis is with higher coincidence rate and positive predictive value. MIC and MAC have a good prediction on the increase or decrease of MCV. Although advanced RBC software's false negative and false positive rates are high, the red blood cell image system is more intuitive and time-saving in observing cells. Consequently, CellaVision-DI60 is suggested to combine with XN-9000 for judging the anisocytosis in daily work comprehensively.

AI-based analysis of 3D position and orientation of red blood cells using a digital in-line holographic microscopy.

The morphological and mechanical characteristics of red blood cells (RBCs) largely vary depending on the occurrence of hematologic disorders. Variations in the rheological properties of RBCs affect the dynamic motions of RBCs, especially their rotational behavior. However, conventional techniques for measuring the orientation of biconcave-shaped RBCs still have some technical limitations, including complicated optical setups, complex post data processing, and low throughput. In this study, we propose a novel image-based technique for measuring 3D position and orientation of normal RBCs using digital in-line holographic microscopy (DIHM) and artificial intelligence (AI). Formaldehyde-fixed RBCs are immobilized in coagulated polydimethylsiloxane (PDMS). Holographic images of RBCs positioned at various out-of-plane angles are acquired by precisely manipulating the PDMS-trapped RBC sample attached to a 4-axis optical stage. With the aid of deep learning algorithms for data augmentation and regression analysis, the out-of-plane angle of RBCs is directly predicted from the captured holographic images. The 3D position and in-plane angle of RBCs are acquired by employing numerical reconstruction and ellipse detection methods. Combining these digital image processing techniques, the 3D positional and orientational information of each RBC recorded in a single holographic image is measured within 23.5 and 3.07s, respectively. The proposed AI-based DIHM technique that can extract the 3D position, orientation, and morphology of individual RBCs would be utilized to analyze the dynamic translational and rotational motions of abnormal RBCs with hematologic disorders in shear flows through further research.

RedTell: an AI tool for interpretable analysis of red blood cell morphology.

Introduction: Hematologists analyze microscopic images of red blood cells to study their morphology and functionality, detect disorders and search for drugs. However, accurate analysis of a large number of red blood cells needs automated computational approaches that rely on annotated datasets, expensive computational resources, and computer science expertise. We introduce RedTell, an AI tool for the interpretable analysis of red blood cell morphology comprising four single-cell modules: segmentation, feature extraction, assistance in data annotation, and classification. Methods: Cell segmentation is performed by a trained Mask R-CNN working robustly on a wide range of datasets requiring no or minimum fine-tuning. Over 130 features that are regularly used in research are extracted for every detected red blood cell. If required, users can train task-specific, highly accurate decision tree-based classifiers to categorize cells, requiring a minimal number of annotations and providing interpretable feature importance. Results: We demonstrate RedTell's applicability and power in three case studies. In the first case study we analyze the difference of the extracted features between the cells coming from patients suffering from different diseases, in the second study we use RedTell to analyze the control samples and use the extracted features to classify cells into echinocytes, discocytes and stomatocytes and finally in the last use case we distinguish sickle cells in sickle cell disease patients. Discussion: We believe that RedTell can accelerate and standardize red blood cell research and help gain new insights into mechanisms, diagnosis, and treatment of red blood cell associated disorders.

Multiplane Image Restoration Using Multivariate Curve Resolution: An Alternative Approach to Deconvolution in Conventional Brightfield Microscopy

Three-dimensional reconstruction in brightfield microscopy is challenging since a 2D image includes from in-focus and out-of-focus light which removes the details of the specimen’s structures. To overcome this problem, many techniques exist, but these generally require an appropriate model of Point Spread Function (PSF). Here, we propose a new images restoration method based on the application of Multivariate Curve Resolution (MCR) algorithms to a stack of brightfield microscopy images to achieve 3D reconstruction without the need for PSF. The method is based on a statistical reconstruction approach using a self-modelling mixture analysis. The MCR-ALS (ALS for Alternating Least Square) algorithm under non-negativity constraints, Wiener, Richardson–Lucy, and blind deconvolution algorithms were applied to silica microbeads and red blood cells images. The MCR analysis produces restored images that show informative structures which are not noticeable in the initial images, and this demonstrates its capability for the multiplane reconstruction of the amplitude of 3D objects. In comparison with 3D deconvolution methods based on a set of No Reference Images Quality Metrics (NR-IQMs) that are Standard Deviation, ENTROPY Average Gradient, and Auto Correlation, our method presents better values of these metrics, showing that it can be used as an alternative to 3D deconvolution methods.

Label-free imaging of red blood cells and oxygenation with color third-order sum-frequency generation microscopy

Mapping red blood cells (RBCs) flow and oxygenation is of key importance for analyzing brain and tissue physiology. Current microscopy methods are limited either in sensitivity or in spatio-temporal resolution. In this work, we introduce a novel approach based on label-free third-order sum-frequency generation (TSFG) and third-harmonic generation (THG) contrasts. First, we propose a novel experimental scheme for color TSFG microscopy, which provides simultaneous measurements at several wavelengths encompassing the Soret absorption band of hemoglobin. We show that there is a strong three-photon (3P) resonance related to the Soret band of hemoglobin in THG and TSFG signals from zebrafish and human RBCs, and that this resonance is sensitive to RBC oxygenation state. We demonstrate that our color TSFG implementation enables specific detection of flowing RBCs in zebrafish embryos and is sensitive to RBC oxygenation dynamics with single-cell resolution and microsecond pixel times. Moreover, it can be implemented on a 3P microscope and provides label-free RBC-specific contrast at depths exceeding 600 µm in live adult zebrafish brain. Our results establish a new multiphoton contrast extending the palette of deep-tissue microscopy.

Diffraction-limited hyperspectral mid-infrared single-pixel microscopy

In this contribution, we demonstrate a wide-field hyperspectral mid-infrared (MIR) microscope based on multidimensional single-pixel imaging (SPI). The microscope employs a high brightness MIR supercontinuum source for broadband (1.55 upmu hbox {m}–4.5 upmu hbox {m}) sample illumination. Hyperspectral imaging capability is achieved by a single micro-opto-electro-mechanical digital micromirror device (DMD), which provides both spatial and spectral differentiation. For that purpose the operational spectral bandwidth of the DMD was significantly extended into the MIR spectral region. In the presented design, the DMD fulfills two essential tasks. On the one hand, as standard for the SPI approach, the DMD sequentially masks captured scenes enabling diffraction-limited imaging in the tens of millisecond time-regime. On the other hand, the diffraction at the micromirrors leads to dispersion of the projected field and thus allows for wavelength selection without the application of additional dispersive optical elements, such as gratings or prisms. In the experimental part, first of all, the imaging and spectral capabilities of the hyperspectral microscope are characterized. The spatial and spectral resolution is assessed by means of test targets and linear variable filters, respectively. At a wavelength of 4.15 upmu hbox {m} a spatial resolution of 4.92 upmu hbox {m} is achieved with a native spectral resolution better than 118.1 nm. Further, a post-processing method for drastic enhancement of the spectral resolution is proposed and discussed. The performance of the MIR hyperspectral microsopce is demonstrated for label-free chemical imaging and examination of polymer compounds and red blood cells. The acquisition and reconstruction of Hadamard sampled 64 times 64 images is achieved in 450 ms and 162 ms, respectively. Thus, combined with an unprecedented intrinsic flexibiliy gained by a tunable field of view and adjustable spatial resolution, the demonstrated design drastically improves the sample throughput in MIR chemical and biomedical imaging.

An automated neural network-based stage-specific malaria detection software using dimension reduction: The malaria microscopy classifier

Due to climate change and the COVID-19 pandemic, the number of malaria cases and deaths, caused by the Plasmodium genus, of which P. falciparum is the most common and lethal to humans, increased between 2019 and 2020. Reversing this trend and eliminating malaria worldwide requires improvements in malaria diagnosis, in which artificial intelligence (AI) has recently been demonstrated to have a great potential. One of the main reasons for the use of neural networks (NNs) is the time saving through automatising the process and the elimination of human error. When classifying with two-dimensional images of red blood cells (RBCs), the number of parameters fitted by the NN for the classification of RBCs is extremely high, which strongly influences the performance of the network, especially for training sets of moderate size. The complicated handling of malaria culturing and sample preparation does not only limit the efficiency of NNs due to small training sets, but also because of the uneven distribution of red blood cell (RBC) categories. To boost the performance of microscopy techniques in malaria diagnosis, our approach aims at resolving these drawbacks by reducing the dimension of the input data and by data augmentation, respectively. We assess the performance of our approach on images recorded by light (LM), atomic force (AFM), and fluorescence microscopy (FM). Our tool, the Malaria Stage Classifier, provides a fast, high-accuracy recognition by (1) identifying individual RBCs in multi-cell microscopy images, (2) extracting characteristic one-dimensional cross-sections from individual RBC images. These cross-sections are selected by a simple algorithm to contain key information about the status of the RBCs and are used to (3) classify the malaria blood stages. We demonstrate that our method is applicable to images recorded by various microscopy techniques and available as a software package.•Identifying individual RBCs in multi-cell microscopy images.•Extracting characteristic one-dimensional cross-sections from individual RBC images. These cross-sections are selected by a simple algorithm to contain key information about the status of the RBCs and are used to.•Classify the malaria blood stages. We demonstrate that our method is applicable to images recorded by various microscopy techniques and available as a software package.

The Value of Dual-Energy Computed Tomography Angiography-Derived Parameters in the Evaluation of Clot Composition

We aimed to assess the value of dual-energy computed tomography angiography (DE-CTA) derived parameters as a quantitative biomarker of thrombus composition in acute ischemic stroke (AIS). AIS patients who underwent DE-CTA before thrombectomy between August 2016 and September 2022 were included in this study. We assessed the relative proportionof red blood cells (RBCs) and the fibrin/platelet ratio (F/P) of the retrieved clots and categorized the clots as RBC-dominant (RBCs > F/P) or F/P-dominant (F/P > RBCs). The thrombus based parameters were measured on polyenergetic images (PEI), virtual monoenergetic (VM), virtual non-contrast (VNC), iodine concentration (IC), and effective atomic number (Zeff) images respectively, and the slope of the spectral Hounsfield unit curve (λHU) was calculated. These parameters were compared in the DE-CTA images of RBC- and F/P-dominant thrombi. The diagnostic performance of the parameters was analyzed using the ROC curve. Correlations between thrombus composition and DE-CTA-derived parameters were assessed. The retrieved clots in 54 of 88 patients (61.36%) were RBC-dominant. The RBC-dominant thrombi showed significantly higher VNC values and lower IC, λHU, and Zeff values than the F/P-dominant thrombi (p < 0.05). The CT density measured on IC images showed the largest AUC value (AUC, 0.94; sensitivity, 77.78%; specificity, 100.00%). The Spearman rank-order correlation coefficient values showed that CT density measured on IC images of the thrombus showed the strongest association with the proportion of RBCs (r=-0.64, p < 0.001) and F/P (r=0.65, p < 0.001). DE-CTA-derived parameters, especially the CT density measured on IC images, could be associated with thrombus composition and allow for personalized thrombectomy strategies.