Automatic Cell Classification Research Articles

Advancing Leukocyte Classification: A Cutting‐Edge Deep Learning Approach for AI‐Driven Clinical Diagnosis

ABSTRACTWhite blood cells (WBCs) are crucial components of the immune system, responsible for detecting and eliminating pathogens. Accurate detection and classification of WBCs are essential for various clinical diagnostics. This study aims to develop an AI framework for detecting and classifying WBCs from microscopic images using a customized YOLOv5 model with three key modifications. Firstly, the C3 module in YOLOv5's backbone is replaced with the innovative C3TR structure to enhance feature extraction and reduce background noise. Secondly, the BiFPN is integrated into the neck to improve feature localization and discrimination. Thirdly, an additional layer in the head enhances detection of small WBCs. Experiments on the BCCD dataset, comprising 352 microscopic blood smear images with leukocytes, demonstrated the framework's superiority over state‐of‐the‐art methods, achieving 99.4% accuracy. Furthermore, the model exhibits computational efficiency, operating over five times faster than existing YOLO models. These findings underscore the framework's promise in medical diagnostics, showcasing deep learning's supremacy in automated cell classification.

SlowMoMan: a web app for discovery of important features along user-drawn trajectories in 2D embeddings.

Nonlinear low-dimensional embeddings allow humans to visualize high-dimensional data, as is often seen in bioinformatics, where datasets may have tens of thousands of dimensions. However, relating the axes of a nonlinear embedding to the original dimensions is a nontrivial problem. In particular, humans may identify patterns or interesting subsections in the embedding, but cannot easily identify what those patterns correspond to in the original data. Thus, we present SlowMoMan (SLOW Motions on MANifolds), a web application which allows the user to draw a one-dimensional path onto a 2D embedding. Then, by back-projecting the manifold to the original, high-dimensional space, we sort the original features such that those most discriminative along the manifold are ranked highly. We show a number of pertinent use cases for our tool, including trajectory inference, spatial transcriptomics, and automatic cell classification. Software: https://yunwilliamyu.github.io/SlowMoMan/; Code: https://github.com/yunwilliamyu/SlowMoMan.

Blur-specific image quality assessment of microscopic hyperspectral images.

Hyperspectral (HS) imaging (HSI) expands the number of channels captured within the electromagnetic spectrum with respect to regular imaging. Thus, microscopic HSI can improve cancer diagnosis by automatic classification of cells. However, homogeneous focus is difficult to achieve in such images, being the aim of this work to automatically quantify their focus for further image correction. A HS image database for focus assessment was captured. Subjective scores of image focus were obtained from 24 subjects and then correlated to state-of-the-art methods. Maximum Local Variation, Fast Image Sharpness block-based Method and Local Phase Coherence algorithms provided the best correlation results. With respect to execution time, LPC was the fastest.

Analysing cerebrospinal fluid with explainable deep learning: From diagnostics to insights.

Analysis of cerebrospinal fluid (CSF) is essential for diagnostic workup of patients with neurological diseases and includes differential cell typing. The current gold standard is based on microscopic examination by specialised technicians and neuropathologists, which is time-consuming, labour-intensive and subjective. We, therefore, developed an image analysis approach based on expert annotations of 123,181 digitised CSF objects from 78 patients corresponding to 15 clinically relevant categories and trained a multiclass convolutional neural network (CNN). The CNN classified the 15 categories with high accuracy (mean AUC 97.3%). By using explainable artificial intelligence (XAI), we demonstrate that the CNN identified meaningful cellular substructures in CSF cells recapitulating human pattern recognition. Based on the evaluation of 511 cells selected from 12 different CSF samples, we validated the CNN by comparing it with seven board-certified neuropathologists blinded for clinical information. Inter-rater agreement between the CNN and the ground truth was non-inferior (Krippendorff's alpha 0.79) compared with the agreement of seven human raters and the ground truth (mean Krippendorff's alpha 0.72, range 0.56-0.81). The CNN assigned the correct diagnostic label (inflammatory, haemorrhagic or neoplastic) in 10 out of 11 clinical samples, compared with 7-11 out of 11 by human raters. Our approach provides the basis to overcome current limitations in automated cell classification for routine diagnostics and demonstrates how a visual explanation framework can connect machine decision-making with cell properties and thus provide a novel versatile and quantitative method for investigating CSF manifestations of various neurological diseases.

DEEP LEARNİNG ARCHİTECHTURE FOR CLASSİFİCATİON OF BREAST CANCER CELLS İN FLUORESCENCE MİCROSCOPY IMAGES

Biological cell classification plays a significant role in the field of biomedical research. Cell classification is useful in different biomedical applications like identification of a normal and abnormal cell, cancer cell recognition, behovioural studies of cells to different drugs etc. Automated cell classification techniquies would assist the radiologist for the disease diagnoses and to grasp the severity of the disease based on the intricate intracellular structures of the cells. In this work a deep learning architechutre based on EfficientNet is designed for automatic classification of human breast cancer cells from fluorescence microscopy images. More specifically transfer learning is employed to take the advantage of the pretrained model and further improvising the performance of the network by fine tuning several of last layers for learning the specific classification task. The proposed deep learning architechture is evaluated on human breast cancer cells, which gave 98.15% accuracy, precision, recall and F1 score. Comparitive analysis of the proposed architechture with the standard architechures is also performed to assert the efficacy of our model.

Automated bone cell classification for confocal laser scanning microscopy volumes

Manual cell classification in microscopy images is a time-consuming process that heavily relies on the subjective perception of the investigator. Identifying bone cells introduces additional difficulties with irregular geometries, and in some culture conditions, the presence of bone mineral. As fluorescence-based lineage tracing becomes more common, classifying cell types based upon cell color can further increase subjectivity. Our goal is to develop and validate a fully automated cell classification algorithm that can (i) objectively identify cells in flattened volumetric image stacks from three-dimensional (3D) bone cell cultures and (ii) classify the cells (osteoblast-lineage) based on the color of their cell bodies. The algorithm used here was developed in MATLAB 2019a and validated by comparing code outputs to manual labeling for eleven images. The precision, recall, and F1 scores were higher than 0.75 for all cell classifications, with the majority being greater than 0.80. No significant differences were found between the manually labelled and automated cell counts or cell classifications. Analysis time for a single image averaged seventeen seconds compared to more than ten minutes for manual labeling. This demonstrates that the program offers a fast, repeatable, and accurate way to classify bone cells by fluorescence in confocal microscopy image data sets. This process can be expanded to improve investigation of other pre-clinical models and histological sections of pathological tissues where color or fluorescence-based differences are used for cell identification.

Normalization of the Immune Microenvironment during Lenalidomide Maintenance Is Associated with Sustained MRD Negativity in Patients with Multiple Myeloma

Normalization of the Immune Microenvironment during Lenalidomide Maintenance Is Associated with Sustained MRD Negativity in Patients with Multiple Myeloma

Machine Learning-Based Automated Selection of Regions for Analysis on Bone Marrow Aspirate Smears

Abstract Manual microscopic examination of bone marrow aspirate (BMA) smears and counting of cell populations remains the standard of practice for accurate assessment of benign and neoplastic bone marrow disorders. While automated cell classification software using machine learning models has been developed and applied to BMAs, current systems nonetheless require manual identification of optimal regions within the slide that are rich in marrow hematopoietic cells. To address this issue, we have developed a machine learning-based platform for automated identification of optimal regions in whole-slide images of BMA smears. A training dataset was developed by manual annotation of 53 BMA slides across biopsy diagnoses including unremarkable trilineal hematopoiesis, acute leukemia, and plasma cell neoplasms, as well as across differences in total cellularity represented by a spectrum of marrow nucleated cell content and white blood cell counts. 10,537 regions among these 53 slides were manually annotated as either “optimal” (regions near aspirate particles with high proportions of marrow nucleated cells), “particle” (aspirate particles), or “hemodilute” (blood-rich regions with high proportions of red blood cells). Training of a neural network-based classifier on 10x magnification slides with region cropping and image augmentation resulted in a classifier with substantial accuracy on new testing-set BMA slides (one-vs-rest AUROC > 0.999 across 10 training/testing splits for all 3 region classes), with very few particle and hemodilute regions being classified as optimal (particle: 0.83%, hemodilute: 0.39%). Additionally, this classifier accurately classifies BMA regions on slides from hematological disorders not represented in the training data, including Burkitt lymphoma (AUROC > 0.999 across region classes), chronic myeloid leukemia (AUROC > 0.999 across region classes), and diffuse large B-cell lymphoma (AUROC = 1 across region classes), demonstrating the broad applicability of our approach. To assess the performance of our classifier on whole-slide images, tiles from 10x magnification slides were manually annotated by three participants with notable concordance (Krippendorff’s alpha = 0.424); substantial agreement was found between manual annotations and model predictions within whole-slide images (optimal AUROC = 0.958, particle AUROC = 1.0, hemodilute AUROC = 0.947). Based on these promising results, this machine learning-based region classification model is being connected to a previously-developed bone marrow cell classifier to fully automate differential cell counting in whole-slide images. The development of this novel automated pipeline has potential to streamline the diagnostic process for hematological disorders while enhancing accuracy and replicability, as well as decreasing diagnostic turnaround time for improving patient care.

Machine learning for cell classification and neighborhood analysis in glioma tissue.

Multiplexed and spatially resolved single-cell analyses that intend to study tissue heterogeneity and cell organization invariably face as a first step the challenge of cell classification. Accuracy and reproducibility are important for the downstream process of counting cells, quantifying cell-cell interactions, and extracting information on disease-specific localized cell niches. Novel staining techniques make it possible to visualize and quantify large numbers of cell-specific molecular markers in parallel. However, due to variations in sample handling and artifacts from staining and scanning, cells of the same type may present different marker profiles both within and across samples. We address multiplexed immunofluorescence data from tissue microarrays of low-grade gliomas and present a methodology using two different machine learning architectures and features insensitive to illumination to perform cell classification. The fully automated cell classification provides a measure of confidence for the decision and requires a comparably small annotated data set for training, which can be created using freely available tools. Using the proposed method, we reached an accuracy of 83.1% on cell classification without the need for standardization of samples. Using our confidence measure, cells with low-confidence classifications could be excluded, pushing the classification accuracy to 94.5%. Next, we used the cell classification results to search for cell niches with an unsupervised learning approach based on graph neural networks. We show that the approach can re-detect specialized tissue niches in previously published data, and that our proposed cell classification leads to niche definitions that may be relevant for sub-groups of glioma, if applied to larger data sets.

Learning deep features for dead and living breast cancer cell classification without staining

Automated cell classification in cancer biology is a challenging topic in computer vision and machine learning research. Breast cancer is the most common malignancy in women that usually involves phenotypically diverse populations of breast cancer cells and an heterogeneous stroma. In recent years, automated microscopy technologies are allowing the study of live cells over extended periods of time, simplifying the task of compiling large image databases. For instance, there have been several studies oriented towards building machine learning systems capable of automatically classifying images of different cell types (i.e. motor neurons, stem cells). In this work we were interested in classifying breast cancer cells as live or dead, based on a set of automatically retrieved morphological characteristics using image processing techniques. Our hypothesis is that live-dead classification can be performed without any staining and using only bright-field images as input. We tackled this problem using the JIMT-1 breast cancer cell line that grows as an adherent monolayer. First, a vast image set composed by JIMT-1 human breast cancer cells that had been exposed to a chemotherapeutic drug treatment (doxorubicin and paclitaxel) or vehicle control was compiled. Next, several classifiers were trained based on well-known convolutional neural networks (CNN) backbones to perform supervised classification using labels obtained from fluorescence microscopy images associated with each bright-field image. Model performances were evaluated and compared on a large number of bright-field images. The best model reached an AUC = 0.941 for classifying breast cancer cells without treatment. Furthermore, it reached AUC = 0.978 when classifying breast cancer cells under drug treatment. Our results highlight the potential of machine learning and computational image analysis to build new diagnosis tools that benefit the biomedical field by reducing cost, time, and stimulating work reproducibility. More importantly, we analyzed the way our classifiers clusterize bright-field images in the learned high-dimensional embedding and linked these groups to salient visual characteristics in live-dead cell biology observed by trained experts.

FOntCell: Fusion of Ontologies of Cells.

High-throughput cell-data technologies such as single-cell RNA-seq create a demand for algorithms for automatic cell classification and characterization. There exist several cell classification ontologies with complementary information. However, one needs to merge them to synergistically combine their information. The main difficulty in merging is to match the ontologies since they use different naming conventions. Therefore, we developed an algorithm that merges ontologies by integrating the name matching between class label names with the structure mapping between the ontology elements based on graph convolution. Since the structure mapping is a time consuming process, we designed two methods to perform the graph convolution: vectorial structure matching and constraint-based structure matching. To perform the vectorial structure matching, we designed a general method to calculate the similarities between vectors of different lengths for different metrics. Additionally, we adapted the slower Blondel method to work for structure matching. We implemented our algorithms into FOntCell, a software module in Python for efficient automatic parallel-computed merging/fusion of ontologies in the same or similar knowledge domains. FOntCell can unify dispersed knowledge from one domain into a unique ontology in OWL format and iteratively reuse it to continuously adapt ontologies with new data endlessly produced by data-driven classification methods, such as of the Human Cell Atlas. To navigate easily across the merged ontologies, it generates HTML files with tabulated and graphic summaries, and interactive circular Directed Acyclic Graphs. We used FOntCell to merge the CELDA, LifeMap and LungMAP Human Anatomy cell ontologies into a comprehensive cell ontology. We compared FOntCell with tools used for the alignment of mouse and human anatomy ontologies task proposed by the Ontology Alignment Evaluation Initiative (OAEI) and found that the Fβ alignment accuracies of FOntCell are above the geometric mean of the other tools; more importantly, it outperforms significantly the best OAEI tools in cell ontology alignment in terms of Fβ alignment accuracies.

Cytoplasmic control of intranuclear polarity by human cytomegalovirus.

Despite its size and rigidity, the nucleus is moved or reorganized by cytoskeletal filaments under various conditions1–11. Moreover, while chromatin organize into non-random domains12, extensive heterogeneity at the single-cell level13 means that precisely how and why nuclei reorganize remains an area of intense investigation. Here, we developed convolutional neural network (CNN)-based automated cell classification and analysis pipelines which revealed the extent to which human cytomegalovirus (HCMV) generates nuclear polarity through a virus-assembled microtubule-organizing center (MTOC). Tubulin acetylation enables microtubules emanating from this MTOC to rotate the nucleus via cytoplasmically exposed dynein-binding domains in the outer nuclear membrane protein, Nesprin-2G, which polarizes the inner nuclear membrane protein, SUN1. This in turn creates intranuclear polarity in Emerin to control nuclear actin filaments that spatially segregate viral DNA from inactive histones and host DNA, maximizing virus replication. Our findings uncover the extent to which viruses can control the nucleus from the cytoplasm.

Developing and Preliminary Validating an Automatic Cell Classification System for Bone Marrow Smears: a Pilot Study

Bone marrow smear examination is an indispensable diagnostic tool in the evaluation of hematological diseases, but the process of manual differential count is labor extensive. In this study, we developed an automatic system with integrated scanning hardware and machine learning-based software to perform differential cell count on bone marrow smears to assist diagnosis. The initial development of the artificial neural network was based on 3000 marrow smear samples retrospectively archived from Sir Run Run Shaw Hospital affiliated to Zhejiang University School of Medicine between June 2016 and December 2018. The preliminary field validating test of the system was based on 124 marrow smears newly collected from the Second Affiliated Hospital of Harbin Medical University between April 2019 and November 2019. The study was performed in parallel of machine automatic recognition with conventional manual differential count by pathologists using the microscope. We selected representative 600,000 marrow cell images as training set of the algorithm, followed by random captured 30,867 cell images for validation. In validation, the overall accuracy of automatic cell classification was 90.1% (95% CI, 89.8–90.5%). In a preliminary field validating test, the reliability coefficient (ICC) of cell series proportion between the two analysis methods were high (ICC ≥ 0.883, P < 0.0001) and the results by the two analysis methods were consistent for granulocytes and erythrocytes. The system was effective in cell classification and differential cell count on marrow smears. It provides a useful digital tool in the screening and evaluation of various hematological disorders.

Morphogo: An Automatic Bone Marrow Cell Classification System on Digital Images Analyzed by Artificial Intelligence

Introduction: The nucleated-cell differential count on the bone marrow aspirate smears is required for the clinical diagnosis of hematological malignancy. Manual bone marrow differential count is time consuming and lacks consistency. In this study, a novel artificial intelligence (AI)-based system was developed to perform cell automatic classification of bone marrow cells and determine its potential clinical applications. Materials and Methods: Bone marrow aspirate smears were collected from the Xinqiao Hospital of Army Medical University. First, an automated analysis system (Morphogo) scanned and generated whole digital images of bone marrow smears. Then, the nucleated marrow cells in the selected areas of the smears at a magnification of ×1,000 were analyzed by the software utilizing an AI-based platform. The cell classification results were further reviewed and confirmed independently by 2 experienced pathologists. The automatic cell classification performance of the system was evaluated using 3 categories: accuracy, sensitivity, and specificity. Correlation coefficients and linear regression equations between automatic cell classification by the AI-based system and concurrent manual differential count were calculated. Results: In 230 cases, the classification accuracy was above 85.7% for hematopoietic lineage cells. Averages of sensitivity and specificity of the system were found to be 69.4 and 97.2%, respectively. The differential cell percentage of the automated count based on 200–500 cell counts was correlated with differential cell percentage provided by the pathologists for granulocytes, erythrocytes, and lymphocytes (r ≥ 0.762, p < 0.001). Discussion/Conclusion: This pilot study confirmed that the Morphogo system is a reliable tool for automatic bone marrow cell differential count analysis and has potential for clinical applications. Current ongoing large-scale multicenter validation studies will provide more information to further confirm the clinical utility of the system.

A Smartphone-Based Cell Segmentation to Support Nasal Cytology

Rhinology studies the anatomy, physiology, and diseases affecting the nasal region—one of the most modern techniques to diagnose these diseases is nasal cytology, which involves microscopic analysis of the cells contained in the nasal mucosa. The standard clinical protocol regulates the compilation of the rhino-cytogram by observing, for each slide, at least 50 fields under an optical microscope to evaluate the cell population and search for cells important for diagnosis. The time and effort required for the specialist to analyze a slide are significant. In this paper, we present a smartphones-based system to support cell segmentation on images acquired directly from the microscope. Then, the specialist can analyze the cells and the other elements extracted directly or, alternatively, he can send them to Rhino-cyt, a server system recently presented in the literature, that also performs the automatic cell classification, giving back the final rhinocytogram. This way he significantly reduces the time for diagnosing. The system crops cells with sensitivity = 0.96, which is satisfactory because it shows that cells are not overlooked as false negatives are few, and therefore largely sufficient to support the specialist effectively. The use of traditional image processing techniques to preprocess the images also makes the process sustainable from the computational point of view for medium–low end architectures and is battery-efficient on a mobile phone.

Unsupervised organization of cervical cells using bright-field and single-shot digital holographic microscopy.

We report results on unsupervised organization of cervical cells using microscopy of Pap-smear samples in brightfield (3-channel color) as well as high-resolution quantitative phase imaging modalities. A number of morphological parameters are measured for each of the 1450 cell nuclei (from 10 woman subjects) imaged in this study. The principal component analysis (PCA) methodology applied to this data shows that the cell image clustering performance improves significantly when brightfield as well as phase information is utilized for PCA as compared to when brightfield-only information is used. The results point to the feasibility of an image-based tool that will be able to mark suspicious cells for further examination by the pathologist. More importantly, our results suggest that the information in quantitative phase images of cells that is typically not used in clinical practice is valuable for automated cell classification applications in general.

Automatic classification of cells in microscopic fecal images using convolutional neural networks.

The analysis of fecal-type components for clinical diagnosis is important. The main examination involves the counting of red blood cells (RBCs), white blood cells (WBCs), and molds under the microscopic. With the development of machine vision, some vision-based detection schemes have been proposed. However, these methods have a single target for detection, with low detection efficiency and low accuracy. We proposed an algorithm to identify the visible image of fecal composition based on intelligent deep learning. The algorithm mainly includes region proposal and candidate recognition. In the process of segmentation, we proposed a morphology extraction algorithm in a complex background. As for the candidate recognition, we proposed a new convolutional neural network (CNN) architecture based on Inception-v3 and principal component analysis (PCA). This method achieves high-average Precision of 90.7%, which is better than the other mainstream CNN models. Finally, the images within the rectangle marks were obtained. The total time for detection of an image was roughly 1200 ms. The algorithm proposed in the present paper can be integrated into an automatic fecal detection system.

Convolutional neural network for cell classification using microscope images of intracellular actin networks.

Automated cell classification is an important yet a challenging computer vision task with significant benefits to biomedicine. In recent years, there have been several studies attempted to build an artificial intelligence-based cell classifier using label-free cellular images obtained from an optical microscope. Although these studies showed promising results, such classifiers were not able to reflect the biological diversity of different types of cell. While in terms of malignant cell, it is well-known that intracellular actin filaments are altered substantially. This is thought to be closely related to the abnormal growth features of tumor cells, their ability to invade surrounding tissues and also to metastasize. Therefore, being able to classify different types of cell based on their biological behaviors using automated technique is more advantageous. This article reveals the difference in the actin cytoskeleton structures between breast normal and cancer cells, which may provide new information regarding malignant changes and be used as additional diagnostic marker. Since the features cannot be well detected by human eyes, we proposed the application of convolutional neural network (CNN) in cell classification based on actin-labeled fluorescence microscopy images. The CNN was evaluated on a large number of actin-labeled fluorescence microscopy images of one human normal breast epithelial cell line and two types of human breast cancer cell line with different levels of aggressiveness. The study revealed that the CNN performed better in the cell classification task compared to a human expert.

Inter-observer variance and the need for standardization in the morphological classification of myelodysplastic syndrome

In this era of genome medicine, the sub-classification of myeloid neoplasms, including myelodysplastic syndrome (MDS), is now supported by genetic testing in selected cases. However, as the initial suspicion and primary diagnosis of the disease still largely relies on morphological features and numbers of hematopoietic cells, the establishment of a uniform diagnostic basis, especially for cell morphology, is essential. In this study, we collected nearly 100,000 hematopoietic cell images from 499 peripheral blood smear specimens from patients with MDS and used these to evaluate the standardization of morphological classification by medical technologists. The observers in this study ranged between two to eleven for each image, and the images were classified according to MDS criteria through a web-based system. We found considerable inter-observer variance in the assessment of dysplastic features. Observers did not recognize cytoplasmic hypo–granularity unless almost all granules in neutrophils were absent. Pseudo Pelger–Huët anomalies were also often overlooked, except for cells with a very typical “pince-nez” appearance. Taken together, this study suggests a requirement for further standardization in terms of morphological cell classification, and a need for the development of automatic cell classification–supporting devices for the accurate diagnosis of MDS.

Dictionary-enhanced imaging cytometry

State-of-the-art high-throughput microscopes are now capable of recording image data at a phenomenal rate, imaging entire microscope slides in minutes. In this paper we investigate how a large image set can be used to perform automated cell classification and denoising. To this end, we acquire an image library consisting of over one quarter-million white blood cell (WBC) nuclei together with CD15/CD16 protein expression for each cell. We show that the WBC nucleus images alone can be used to replicate CD expression-based gating, even in the presence of significant imaging noise. We also demonstrate that accurate estimates of white blood cell images can be recovered from extremely noisy images by comparing with a reference dictionary. This has implications for dose-limited imaging when samples belong to a highly restricted class such as a well-studied cell type. Furthermore, large image libraries may endow microscopes with capabilities beyond their hardware specifications in terms of sensitivity and resolution. We call for researchers to crowd source large image libraries of common cell lines to explore this possibility.