Dice-XMBD: Deep Learning-Based Cell Segmentation for Imaging Mass Cytometry.
Ontology highlight
ABSTRACT:
SUBMITTER: Xiao X
PROVIDER: S-EPMC8480472 | biostudies-literature |
REPOSITORIES: biostudies-literature
Similar Datasets
Project description:BackgroundVisualizing and quantifying cellular heterogeneity is of central importance to study tissue complexity, development, and physiology and has a vital role in understanding pathologies. Mass spectrometry-based methods including imaging mass cytometry (IMC) have in recent years emerged as powerful approaches for assessing cellular heterogeneity in tissues. IMC is an innovative multiplex imaging method that combines imaging using up to 40 metal conjugated antibodies and provides distributions of protein markers in tissues with a resolution of 1 μm2 area. However, resolving the output signals of individual cells within the tissue sample, i.e., single cell segmentation, remains challenging. To address this problem, we developed MATISSE (iMaging mAss cyTometry mIcroscopy Single cell SegmEntation), a method that combines high-resolution fluorescence microscopy with the multiplex capability of IMC into a single workflow to achieve improved segmentation over the current state-of-the-art.ResultsMATISSE results in improved quality and quantity of segmented cells when compared to IMC-only segmentation in sections of heterogeneous tissues. Additionally, MATISSE enables more complete and accurate identification of epithelial cells, fibroblasts, and infiltrating immune cells in densely packed cellular areas in tissue sections. MATISSE has been designed based on commonly used open-access tools and regular fluorescence microscopy, allowing easy implementation by labs using multiplex IMC into their analysis methods.ConclusionMATISSE allows segmentation of densely packed cellular areas and provides a qualitative and quantitative improvement when compared to IMC-based segmentation. We expect that implementing MATISSE into tissue section analysis pipelines will yield improved cell segmentation and enable more accurate analysis of the tissue microenvironment in epithelial tissue pathologies, such as autoimmunity and cancer.
| S-EPMC8114487 | biostudies-literature
Project description:Motivation:Mass cytometry or CyTOF is an emerging technology for high-dimensional multiparameter single cell analysis that overcomes many limitations of fluorescence-based flow cytometry. New methods for analyzing CyTOF data attempt to improve automation, scalability, performance and interpretation of data generated in large studies. Assigning individual cells into discrete groups of cell types (gating) involves time-consuming sequential manual steps, untenable for larger studies. Results:We introduce DeepCyTOF, a standardization approach for gating, based on deep learning techniques. DeepCyTOF requires labeled cells from only a single sample. It is based on domain adaptation principles and is a generalization of previous work that allows us to calibrate between a target distribution and a source distribution in an unsupervised manner. We show that DeepCyTOF is highly concordant (98%) with cell classification obtained by individual manual gating of each sample when applied to a collection of 16 biological replicates of primary immune blood cells, even when measured across several instruments. Further, DeepCyTOF achieves very high accuracy on the semi-automated gating challenge of the FlowCAP-I competition as well as two CyTOF datasets generated from primary immune blood cells: (i) 14 subjects with a history of infection with West Nile virus (WNV), (ii) 34 healthy subjects of different ages. We conclude that deep learning in general, and DeepCyTOF specifically, offers a powerful computational approach for semi-automated gating of CyTOF and flow cytometry data. Availability and implementation:Our codes and data are publicly available at https://github.com/KlugerLab/deepcytof.git. Contact:yuval.kluger@yale.edu. Supplementary information:Supplementary data are available at Bioinformatics online.
| S-EPMC5860171 | biostudies-literature
Project description:Imaging mass cytometry (IMC) is a powerful technique capable of detecting over 30 markers on a single slide. It has been increasingly used for single-cell-based spatial phenotyping in a wide range of samples. However, it only acquires a rectangle field of view (FOV) with a relatively small size and low image resolution, which hinders downstream analysis. Here, we reported a highly practical dual-modality imaging method that combines high-resolution immunofluorescence (IF) and high-dimensional IMC on the same tissue slide. Our computational pipeline uses the whole-slide image (WSI) of IF as a spatial reference and integrates small-FOV IMC into a WSI of IMC. The high-resolution IF images enable accurate single-cell segmentation to extract robust high-dimensional IMC features for downstream analysis. We applied this method in esophageal adenocarcinoma of different stages, identified the single-cell pathology landscape via reconstruction of WSI IMC images, and demonstrated the advantage of the dual-modality imaging strategy.
| S-EPMC10626190 | biostudies-literature
Project description:Breast cancer analysis implies that radiologists inspect mammograms to detect suspicious breast lesions and identify mass tumors. Artificial intelligence techniques offer automatic systems for breast mass segmentation to assist radiologists in their diagnosis. With the rapid development of deep learning and its application to medical imaging challenges, UNet and its variations is one of the state-of-the-art models for medical image segmentation that showed promising performance on mammography. In this paper, we propose an architecture, called Connected-UNets, which connects two UNets using additional modified skip connections. We integrate Atrous Spatial Pyramid Pooling (ASPP) in the two standard UNets to emphasize the contextual information within the encoder-decoder network architecture. We also apply the proposed architecture on the Attention UNet (AUNet) and the Residual UNet (ResUNet). We evaluated the proposed architectures on two publically available datasets, the Curated Breast Imaging Subset of Digital Database for Screening Mammography (CBIS-DDSM) and INbreast, and additionally on a private dataset. Experiments were also conducted using additional synthetic data using the cycle-consistent Generative Adversarial Network (CycleGAN) model between two unpaired datasets to augment and enhance the images. Qualitative and quantitative results show that the proposed architecture can achieve better automatic mass segmentation with a high Dice score of 89.52%, 95.28%, and 95.88% and Intersection over Union (IoU) score of 80.02%, 91.03%, and 92.27%, respectively, on CBIS-DDSM, INbreast, and the private dataset.
| S-EPMC8640011 | biostudies-literature
Project description:Flow cytometry nowadays is among the main working instruments in modern biology paving the way for clinics to provide early, quick, and reliable diagnostics of many blood-related diseases. The major problem for clinical applications is the detection of rare pathogenic objects in patient blood. These objects can be circulating tumor cells, very rare during the early stages of cancer development, various microorganisms and parasites in the blood during acute blood infections. All of these rare diagnostic objects can be detected and identified very rapidly to save a patient's life. This review outlines the main techniques of visualization of rare objects in the blood flow, methods for extraction of such objects from the blood flow for further investigations and new approaches to identify the objects automatically with the modern deep learning methods.
| S-EPMC7177904 | biostudies-literature
Project description:Deep learning has achieved spectacular performance in image and speech recognition and synthesis. It outperforms other machine learning algorithms in problems where large amounts of data are available. In the area of measurement technology, instruments based on the photonic time stretch have established record real-time measurement throughput in spectroscopy, optical coherence tomography, and imaging flow cytometry. These extreme-throughput instruments generate approximately 1 Tbit/s of continuous measurement data and have led to the discovery of rare phenomena in nonlinear and complex systems as well as new types of biomedical instruments. Owing to the abundance of data they generate, time-stretch instruments are a natural fit to deep learning classification. Previously we had shown that high-throughput label-free cell classification with high accuracy can be achieved through a combination of time-stretch microscopy, image processing and feature extraction, followed by deep learning for finding cancer cells in the blood. Such a technology holds promise for early detection of primary cancer or metastasis. Here we describe a new deep learning pipeline, which entirely avoids the slow and computationally costly signal processing and feature extraction steps by a convolutional neural network that directly operates on the measured signals. The improvement in computational efficiency enables low-latency inference and makes this pipeline suitable for cell sorting via deep learning. Our neural network takes less than a few milliseconds to classify the cells, fast enough to provide a decision to a cell sorter for real-time separation of individual target cells. We demonstrate the applicability of our new method in the classification of OT-II white blood cells and SW-480 epithelial cancer cells with more than 95% accuracy in a label-free fashion.
| S-EPMC6668572 | biostudies-literature
Project description:Even though ghost imaging (GI), an unconventional imaging method, has received increased attention by researchers during the last decades, imaging speed is still not satisfactory. Once the data-acquisition method and the system parameters are determined, only the processing method has the potential to accelerate image-processing significantly. However, both the basic correlation method and the compressed sensing algorithm, which are often used for ghost imaging, have their own problems. To overcome these challenges, a novel deep learning ghost imaging method is proposed in this paper. We modified the convolutional neural network that is commonly used in deep learning to fit the characteristics of ghost imaging. This modified network can be referred to as ghost imaging convolutional neural network. Our simulations and experiments confirm that, using this new method, a target image can be obtained faster and more accurate at low sampling rate compared with conventional GI method.
| S-EPMC5915577 | biostudies-other
Project description:When analyzing data from in situ RNA detection technologies, cell segmentation is an essential step in identifying cell boundaries, assigning RNA reads to cells, and studying the gene expression and morphological features of cells. We developed a deep-learning-based method, GeneSegNet, that integrates both gene expression and imaging information to perform cell segmentation. GeneSegNet also employs a recursive training strategy to deal with noisy training labels. We show that GeneSegNet significantly improves cell segmentation performances over existing methods that either ignore gene expression information or underutilize imaging information.
| S-EPMC10585768 | biostudies-literature
Project description:Single-cell mass cytometry, also known as cytometry by time of flight (CyTOF) is a powerful high-throughput technology that allows analysis of up to 50 protein markers per cell for the quantification and classification of single cells. Traditional manual gating utilized to identify new cell populations has been inadequate, inefficient, unreliable, and difficult to use, and no algorithms to identify both calibration and new cell populations has been well established. A deep learning with graphic cluster (DGCyTOF) visualization is developed as a new integrated embedding visualization approach in identifying canonical and new cell types. The DGCyTOF combines deep-learning classification and hierarchical stable-clustering methods to sequentially build a tri-layer construct for known cell types and the identification of new cell types. First, deep classification learning is constructed to distinguish calibration cell populations from all cells by softmax classification assignment under a probability threshold, and graph embedding clustering is then used to identify new cell populations sequentially. In the middle of two-layer, cell labels are automatically adjusted between new and unknown cell populations via a feedback loop using an iteration calibration system to reduce the rate of error in the identification of cell types, and a 3-dimensional (3D) visualization platform is finally developed to display the cell clusters with all cell-population types annotated. Utilizing two benchmark CyTOF databases comprising up to 43 million cells, we compared accuracy and speed in the identification of cell types among DGCyTOF, DeepCyTOF, and other technologies including dimension reduction with clustering, including Principal Component Analysis (PCA), Factor Analysis (FA), Independent Component Analysis (ICA), Isometric Feature Mapping (Isomap), t-distributed Stochastic Neighbor Embedding (t-SNE), and Uniform Manifold Approximation and Projection (UMAP) with k-means clustering and Gaussian mixture clustering. We observed the DGCyTOF represents a robust complete learning system with high accuracy, speed and visualization by eight measurement criteria. The DGCyTOF displayed F-scores of 0.9921 for CyTOF1 and 0.9992 for CyTOF2 datasets, whereas those scores were only 0.507 and 0.529 for the t-SNE+k-means; 0.565 and 0.59, for UMAP+ k-means. Comparison of DGCyTOF with t-SNE and UMAP visualization in accuracy demonstrated its approximately 35% superiority in predicting cell types. In addition, observation of cell-population distribution was more intuitive in the 3D visualization in DGCyTOF than t-SNE and UMAP visualization. The DGCyTOF model can automatically assign known labels to single cells with high accuracy using deep-learning classification assembling with traditional graph-clustering and dimension-reduction strategies. Guided by a calibration system, the model seeks optimal accuracy balance among calibration cell populations and unknown cell types, yielding a complete and robust learning system that is highly accurate in the identification of cell populations compared to results using other methods in the analysis of single-cell CyTOF data. Application of the DGCyTOF method to identify cell populations could be extended to the analysis of single-cell RNASeq data and other omics data.
| S-EPMC9060369 | biostudies-literature
Project description:Recent advancements in deep learning have revolutionized the way microscopy images of cells are processed. Deep learning network architectures have a large number of parameters, thus, in order to reach high accuracy, they require a massive amount of annotated data. A common way of improving accuracy builds on the artificial increase of the training set by using different augmentation techniques. A less common way relies on test-time augmentation (TTA) which yields transformed versions of the image for prediction and the results are merged. In this paper we describe how we have incorporated the test-time argumentation prediction method into two major segmentation approaches utilized in the single-cell analysis of microscopy images. These approaches are semantic segmentation based on the U-Net, and instance segmentation based on the Mask R-CNN models. Our findings show that even if only simple test-time augmentations (such as rotation or flipping and proper merging methods) are applied, TTA can significantly improve prediction accuracy. We have utilized images of tissue and cell cultures from the Data Science Bowl (DSB) 2018 nuclei segmentation competition and other sources. Additionally, boosting the highest-scoring method of the DSB with TTA, we could further improve prediction accuracy, and our method has reached an ever-best score at the DSB.
| S-EPMC7081314 | biostudies-literature