Automatic Cell Detection in Bright-Field Microscope Images Using SIFT, Random Forests, and Hierarchical Clustering

Deep Learning-Based Microscopic Cell Detection Using Inverse Distance Transform and Auxiliary Counting

Microscopic cell detection is a challenging task due to significant inter-cell occlusions in dense clusters and diverse cell morphologies. This paper introduces a novel framework designed to enhance automated cell detection. The proposed approach integrates a deep learning model that produces an inverse distance transform-based detection map from the given image, accompanied by a secondary network designed to regress a cell density map from the same input. The inverse distance transform-based map effectively highlights each cell instance in the densely populated areas, while the density map accurately estimates the total cell count in the image. Then, a custom counting-aided cell center extraction strategy leverages the cell count obtained by integrating over the density map to refine the detection process, significantly reducing false responses and thereby boosting overall accuracy. The proposed framework demonstrated superior performance with F-scores of 96.93%, 91.21%, and 92.00% on the VGG, MBM, and ADI datasets, respectively, surpassing existing state-of-the-art methods. It also achieved the lowest distance error, further validating the effectiveness of the proposed approach. These results demonstrate significant potential for automated cell analysis in biomedical applications.

A machine learning-based prediction model pre-operatively for functional recovery after 1-year of hip fracture surgery in older people

Background

Machine learning (ML) has been widely utilized for constructing high-performance prediction models. This study aimed to develop a preoperative machine learning-based prediction model to identify functional recovery one year after hip fracture surgery.

Methods

We collected data from 176 elderly hip fracture patients admitted to the Department of Orthopaedics and Oncology at Shenzhen Second People's Hospital between May 2019 and December 2019, who met the inclusion criteria. Patient's functional recovery was monitored for one year after surgery. We selected 26 factors, comprising 12 preoperative indicators, 8 surgical indicators, and 6 postoperative indicators. Eventually, 77 patients were included based on the exclusion criteria. Random allocation divided them into the training set (70%) and test set (30%) for internal validation. The Lasso method was employed to screen prognostic variables. We conducted comparisons among various common machine learning classifiers to determine the best prediction model. Prediction performance was evaluated using the area under the receiver operating characteristic curve (ROC), calibration curve, and decision curve analysis. To identify the importance of the predictor variables, we performed the recursive feature elimination (RFE) algorithm based on Shapley Additive Explanations (SHAP) values.

Results

The AUCs for the testing dataset were as follows: logistic regression (Logit) model = 0.934, k-nearest neighbors (KNN) model = 0.930, support vector machine (SVM) model = 0.910, Gaussian naive Bayes (GNB) model = 0.926, decision tree (DT) model = 0.730, random forest (RF) model = 0.957, and Extreme Gradient Boosting (XGB) model = 0.902. Among the seven ML-based models tested, the RF model demonstrated the best prediction performance, incorporating four features: postoperative rehabilitation compliance, marital status, age-adjusted Charlson comorbidity score (aCCI), and clinical frailty scale (CFS).

Conclusion

We developed a prediction model for the functional recovery following hip fracture surgery in elderly patients after one year, based on the Random Forest (RF) algorithm. This model exhibited superior prediction performance (ROC) compared to other models. The software application is available for use. External validation in a larger patient cohort or diverse hospital settings is necessary to assess the clinical utility of this tool.

Cell segmentation from telecentric bright-field transmitted light microscopy images using a Residual Attention U-Net: A case study on HeLa line

Living cell segmentation from bright-field light microscopy images is challenging due to the image complexity and temporal changes in the living cells. Recently developed deep learning (DL)-based methods became popular in medical and microscopy image segmentation tasks due to their success and promising outcomes. The main objective of this paper is to develop a deep learning, U-Net-based method to segment the living cells of the HeLa line in bright-field transmitted light microscopy. To find the most suitable architecture for our datasets, a residual attention U-Net was proposed and compared with an attention and a simple U-Net architecture. The attention mechanism highlights the remarkable features and suppresses activations in the irrelevant image regions. The residual mechanism overcomes with vanishing gradient problem. The Mean-IoU score for our datasets reaches 0.9505, 0.9524, and 0.9530 for the simple, attention, and residual attention U-Net, respectively. The most accurate semantic segmentation results was achieved in the Mean-IoU and Dice metrics by applying the residual and attention mechanisms together. The watershed method applied to this best - Residual Attention - semantic segmentation result gave the segmentation with the specific information for each cell.

DeepSeed Local Graph Matching for Densely Packed Cells Tracking

The tracking of densely packed plant cells across microscopy image sequences is very challenging, because their appearance change greatly over time. A local graph matching algorithm was proposed to track such cells by exploiting the tight spatial topology of neighboring cells, and then an iterative searching strategy was used to grow the correspondence from a seed cell pair. Thus, the performance of the existing tracking approach heavily relies on the robustness of finding seed cell pair. However, the existing local graph matching algorithm cannot guarantee the correctness of the seed cell pair, especially in unregistered image sequences or image sequences with large time intervals. In this paper, we propose a DeepSeed local graph matching model to find seed cell pair robustly, by combining local graph matching and CNN-based similarity learning, which uses cells' spatial-temporal contextual information and cell pairs' similarity information. The CNN-based similarity learning is designed to learn cells' deep feature and measure cell pairs' similarity. Compared with the existing plant cell matching methods, the experimental results show that the DeepSeed local graph matching method can track most cells in unregistered image sequences. Moreover, the DeepSeed tracking algorithm can accurately track cells across image sequences with large time intervals.

High Throughput and Highly Controllable Methods for In Vitro Intracellular Delivery

In vitro and ex vivo intracellular delivery methods hold the key for releasing the full potential of tissue engineering, drug development, and many other applications. In recent years, there has been significant progress in the design and implementation of intracellular delivery systems capable of delivery at the same scale as viral transfection and bulk electroporation but offering fewer adverse outcomes. This review strives to examine a variety of methods for in vitro and ex vivo intracellular delivery such as flow-through microfluidics, engineered substrates, and automated probe-based systems from the perspective of throughput and control. Special attention is paid to a particularly promising method of electroporation using micro/nanochannel based porous substrates, which expose small patches of cell membrane to permeabilizing electric field. Porous substrate electroporation parameters discussed include system design, cells and cargos used, transfection efficiency and cell viability, and the electric field and its effects on molecular transport. The review concludes with discussion of potential new innovations which can arise from specific aspects of porous substrate-based electroporation platforms and high throughput, high control methods in general.

Optimized High-Contrast Brightfield Microscopy Application for Noninvasive Proliferation Assays of Human Cell Cultures

High-contrast brightfield (HCBF) microscopy has emerged as a strong tool for noninvasive counting of cells in culture. HCBF imaging delivers precise cell growth data and is completely label free rendering it an attractive alternative to common cell counting procedures that often adversely affect cell growth. With computational image analysis, HCBF achieves efficient high-throughput automated workflows, extremely relevant for drug and chemical screens in pharmaceutical, toxicological, and biomedical research. We demonstrate the applicability of HCBF microscopy to count three common cell types (HEK293, Huh7, and primary human dermal fibroblasts) with diverse morphology challenging the method. The three cell types required different analysis settings, and we identified two parameters of the computational image analysis, which after cell-specific optimization significantly improved the cell counting accuracy, namely the lower size limit and the intensity threshold. Three-dimensional (3D) imaging approaches, which have obtained great attention in recent years, were an interesting prospect to combine with HCBF microscopy. We optimized the analysis of two 3D outputs but found 3D HCBF imaging to be inferior to the optimized single-layer HCBF imaging for cell counting. HCBF cell counts were highly linearly correlated with (R² > 0.99) and highly similar (<15% difference) to cell counts obtained through Hoechst staining, over a broad range of densities allowing at least this level of accuracy for two to three cell generations in Huh7 cells and fibroblasts. Counts of HEK293 cells correlated somewhat less. In conclusion, the HCBF cell counting method is excellently suited for cell proliferation assays and cytotoxicity assays.

Automated Neuron Detection in High-Content Fluorescence Microscopy Images Using Machine Learning

The study of neuronal morphology in relation to function, and the development of effective medicines to positively impact this relationship in patients suffering from neurodegenerative diseases, increasingly involves image-based high-content screening and analysis. The first critical step toward fully automated high-content image analyses in such studies is to detect all neuronal cells and distinguish them from possible non-neuronal cells or artifacts in the images. Here we investigate the performance of well-established machine learning techniques for this purpose. These include support vector machines, random forests, k-nearest neighbors, and generalized linear model classifiers, operating on an extensive set of image features extracted using the compound hierarchy of algorithms representing morphology, and the scale-invariant feature transform. We present experiments on a dataset of rat hippocampal neurons from our own studies to find the most suitable classifier(s) and subset(s) of features in the common practical setting where there is very limited annotated data for training. The results indicate that a random forests classifier using the right feature subset ranks best for the considered task, although its performance is not statistically significantly better than some support vector machine based classification models.

Label-free classification of cells based on supervised machine learning of subcellular structures

It is demonstrated that cells can be classified by pattern recognition of the subcellular structure of non-stained live cells, and the pattern recognition was performed by machine learning. Human white blood cells and five types of cancer cell lines were imaged by quantitative phase microscopy, which provides morphological information without staining quantitatively in terms of optical thickness of cells. Subcellular features were then extracted from the obtained images as training data sets for the machine learning. The built classifier successfully classified WBCs from cell lines (area under ROC curve = 0.996). This label-free, non-cytotoxic cell classification based on the subcellular structure of QPM images has the potential to serve as an automated diagnosis of single cells.

Cell Counting and Segmentation of Immunohistochemical Images in the Spinal Cord: Comparing Deep Learning and Traditional Approaches

Estimation of cell nuclei in images stained for the c-fos protein using immunohistochemistry (IHC) is infeasible in large image sets. Use of multiple human raters to increase throughput often creates variance in the data analysis. Machine learning techniques for biomedical image analysis have been explored for cell-counting in pathology, but their performance on IHC staining, especially to label activated cells in the spinal cord is unknown. In this study, we evaluate different machine learning techniques to segment and count spinal cord neurons that have been active during stepping. We present a qualitative as well as quantitative comparison of algorithmic performance versus two human raters. Quantitative ratings are presented with cell-count statistics and Dice (DSI) scores. We also show the degree of variability between multiple human raters' segmentations and observe that there is a higher degree of variability in segmentations produced by classic machine learning techniques (SVM and Random forest) as compared to the newer deep learning techniques. The work presented here, represents the first steps towards addressing the analysis time bottleneck of large image data sets generated by c-fos IHC staining techniques, a task that would be impossible to do manually.

Biomedical image classification based on a cascade of an SVM with a reject option and subspace analysis

Automated biomedical image classification could confront the challenges of high level noise, image blur, illumination variation and complicated geometric correspondence among various categorical biomedical patterns in practice. To handle these challenges, we propose a cascade method consisting of two stages for biomedical image classification. At stage 1, we propose a confidence score based classification rule with a reject option for a preliminary decision using the support vector machine (SVM). The testing images going through stage 1 are separated into two groups based on their confidence scores. Those testing images with sufficiently high confidence scores are classified at stage 1 while the others with low confidence scores are rejected and fed to stage 2. At stage 2, the rejected images from stage 1 are first processed by a subspace analysis technique called eigenfeature regularization and extraction (ERE), and then classified by another SVM trained in the transformed subspace learned by ERE. At both stages, images are represented based on two types of local features, i.e., SIFT and SURF, respectively. They are encoded using various bag-of-words (BoW) models to handle biomedical patterns with and without geometric correspondence, respectively. Extensive experiments are implemented to evaluate the proposed method on three benchmark real-world biomedical image datasets. The proposed method significantly outperforms several competing state-of-the-art methods in terms of classification accuracy.

Measuring multiple parameters of CD8+ tumor-infiltrating lymphocytes in human cancers by image analysis

BACKGROUND

Immuno-oncology and cancer immunotherapies are areas of intense research. The numbers and locations of CD8+ tumor-infiltrating lymphocytes (TILs) are important measures of the immune response to cancer with prognostic, pharmacodynamic, and predictive potential. We describe the development, validation, and application of advanced image analysis methods to characterize multiple immunohistochemistry-derived CD8 parameters in clinical and nonclinical tumor tissues.

METHODS

Commercial resection tumors from nine cancer types, and paired screening/on-drug biopsies of non-small-cell lung carcinoma (NSCLC) patients enrolled in a phase 1/2 clinical trial investigating the PD-L1 antibody therapy durvalumab (NCT01693562), were immunostained for CD8. Additional NCT01693562 samples were immunostained with a CD8/PD-L1 dual immunohistochemistry assay. Whole-slide scanning was performed, tumor regions were annotated by a pathologist, and images were analyzed with customized algorithms using Definiens Developer XD software. Validation of image analysis data used cell-by-cell comparison to pathologist scoring across a range of CD8+ TIL densities of all nine cancers, relying primarily on 95% confidence in having at least moderate agreement regarding Lin concordance correlation coefficient (CCC = 0.88-0.99, CCC_lower = 0.65-0.96).

RESULTS

We found substantial variability in CD8+ TILs between individual patients and across the nine types of human cancer. Diffuse large B-cell lymphoma had several-fold more CD8+ TILs than some other cancers. TIL densities were significantly higher in the invasive margin versus tumor center for carcinomas of head and neck, kidney and pancreas, and NSCLC; the reverse was true only for prostate cancer. In paired patient biopsies, there were significantly increased CD8+ TILs 6 weeks after onset of durvalumab therapy (mean of 365 cells/mm2 over baseline; P = 0.009), consistent with immune activation. Image analysis accurately enumerated CD8+ TILs in PD-L1+ regions of lung tumors using the dual assay and also measured elongate CD8+ lymphocytes which constituted a fraction of overall TILs.

CONCLUSIONS

Validated image analysis accurately enumerates CD8+ TILs, permitting comparisons of CD8 parameters among tumor regions, individual patients, and cancer types. It also enables the more complex digital solutions needed to better understand cancer immunity, like analysis of multiplex immunohistochemistry and spatial evaluation of the various components comprising the tumor microenvironment.

TRIAL REGISTRATION

ClinicalTrials.gov identifier: NCT01693562 . Study code: CD-ON-MEDI4736-1108. Interventional study (ongoing but not currently recruiting). Actual study start date: August 29, 2012. Primary completion date: June 23, 2017 (final data collection date for primary outcome measure).

Deep Learning Automates the Quantitative Analysis of Individual Cells in Live-Cell Imaging Experiments

Live-cell imaging has opened an exciting window into the role cellular heterogeneity plays in dynamic, living systems. A major critical challenge for this class of experiments is the problem of image segmentation, or determining which parts of a microscope image correspond to which individual cells. Current approaches require many hours of manual curation and depend on approaches that are difficult to share between labs. They are also unable to robustly segment the cytoplasms of mammalian cells. Here, we show that deep convolutional neural networks, a supervised machine learning method, can solve this challenge for multiple cell types across the domains of life. We demonstrate that this approach can robustly segment fluorescent images of cell nuclei as well as phase images of the cytoplasms of individual bacterial and mammalian cells from phase contrast images without the need for a fluorescent cytoplasmic marker. These networks also enable the simultaneous segmentation and identification of different mammalian cell types grown in co-culture. A quantitative comparison with prior methods demonstrates that convolutional neural networks have improved accuracy and lead to a significant reduction in curation time. We relay our experience in designing and optimizing deep convolutional neural networks for this task and outline several design rules that we found led to robust performance. We conclude that deep convolutional neural networks are an accurate method that require less curation time, are generalizable to a multiplicity of cell types, from bacteria to mammalian cells, and expand live-cell imaging capabilities to include multi-cell type systems.

Robust Nucleus/Cell Detection and Segmentation in Digital Pathology and Microscopy Images: A Comprehensive Review

Digital pathology and microscopy image analysis is widely used for comprehensive studies of cell morphology or tissue structure. Manual assessment is labor intensive and prone to interobserver variations. Computer-aided methods, which can significantly improve the objectivity and reproducibility, have attracted a great deal of interest in recent literature. Among the pipeline of building a computer-aided diagnosis system, nucleus or cell detection and segmentation play a very important role to describe the molecular morphological information. In the past few decades, many efforts have been devoted to automated nucleus/cell detection and segmentation. In this review, we provide a comprehensive summary of the recent state-of-the-art nucleus/cell segmentation approaches on different types of microscopy images including bright-field, phase-contrast, differential interference contrast, fluorescence, and electron microscopies. In addition, we discuss the challenges for the current methods and the potential future work of nucleus/cell detection and segmentation.

Automatic cell detection in bright-field microscopy for microbeam irradiation studies

Automatic cell detection in bright-field illumination microscopy is challenging due to cells' inherent optical properties. Applications including individual cell microbeam irradiation demand minimisation of additional cell stressing factors, so contrast-enhancing fluorescence microscopy should be avoided. Additionally, the use of optically non-homogeneous substrates amplifies the problem. This research focuses on the design of a method for automatic cell detection on polypropylene substrate, suitable for microbeam irradiation. In order to fulfil the relative requirements, the Harris corner detector was employed to detect apparent cellular features. These features-corners were clustered based on a dual-clustering technique according to the density of their distribution across the image. Weighted centroids were extracted from the clusters of corners and constituted the targets for irradiation. The proposed method identified more than 88% of the 1,738 V79 Chinese hamster cells examined. Moreover, a processing time of 2.6 s per image fulfilled the requirements for a near real-time cell detection-irradiation system.

Influence of the phase effect on gradient-based and statistics-based focus measures in bright field microscopy

Autofocusing is essential to high throughput microscopy and live cell imaging and requires reliable focus measures. Phase objects such as separated single Chinese hamster ovary cells are almost invisible at the optical focus position in bright field microscopy images. Because of the phase effect, defocused images of phase objects have more contrast. In this paper, we show that widely used focus measures exhibit an untypical behaviour for such images. In the case of homogeneous cells, that is, when most cells tend to lie in the same focal plane, both gradient-based and statistics-based focus measures tend to have a local minimum instead of a global maximum at the optical focus position. On the other hand, if images show inhomogeneous cells, gradient-based focus measures tend to yield typical focus curves, whereas statistics-based focus measures deliver curves similar to the case of homogeneous cells. These results were interpreted using the equation describing the phase effect and patch-wise analysis of the focus curves. Bioprocess engineering experts are also influenced by the phase effect. Forty-four focus positions selected by them led to the conclusion that they prefer to look at defocused images instead of those at the optical focus.