A weakly supervised deep learning approach for label-free imaging flow-cytometry-based blood diagnostics

Information-Distilled Generative Label-Free Morphological Profiling Encodes Cellular Heterogeneity

Image-based cytometry faces challenges due to technical variations arising from different experimental batches and conditions, such as differences in instrument configurations or image acquisition protocols, impeding genuine biological interpretation of cell morphology. Existing solutions, often necessitating extensive pre-existing data knowledge or control samples across batches, have proved limited, especially with complex cell image data. To overcome this, "Cyto-Morphology Adversarial Distillation" (CytoMAD), a self-supervised multi-task learning strategy that distills biologically relevant cellular morphological information from batch variations, is introduced to enable integrated analysis across multiple data batches without complex data assumptions or extensive manual annotation. Unique to CytoMAD is its "morphology distillation", symbiotically paired with deep-learning image-contrast translation-offering additional interpretable insights into label-free cell morphology. The versatile efficacy of CytoMAD is demonstrated in augmenting the power of biophysical imaging cytometry. It allows integrated label-free classification of human lung cancer cell types and accurately recapitulates their progressive drug responses, even when trained without the drug concentration information. CytoMAD  also allows joint analysis of tumor biophysical cellular heterogeneity, linked to epithelial-mesenchymal plasticity, that standard fluorescence markers overlook. CytoMAD can substantiate the wide adoption of biophysical cytometry for cost-effective diagnosis and screening.

Prediction of antibody production performance change in Chinese hamster ovary cells using morphological profiling

Monoclonal antibodies (mAbs) represent a significant segment of biopharmaceuticals, with the market for mAb therapeutics expected to reach $200 billion in 2021. Chinese Hamster Ovary (CHO) cells are the industry standard for large-scale mAb production owing to their adaptability and genetic engineering capabilities. However, maintaining consistent product quality is challenging, primarily because of the inherent genetic instability of CHO cells. In this study, we address the need for advanced technologies for quality monitoring of host cells in biopharmaceuticals. We highlight the limitations of traditional cell assessment techniques such as flow cytometry and propose a noninvasive, label-free image-based analysis method. By utilizing advanced image processing and machine learning, this technique aims to non-invasively and quantitatively evaluate subtle quality changes in suspension cells. The research aims to investigate the use of morphological analysis for identifying subtle alterations in mAb productivity of CHO cells, employing cells stimulated by compounds as a model for this study. Our results show that the mAb productivity of CHO cells (day 8) can be predicted only from their early morphological profile (day 3). Our study also discusses the importance of strategic methods for forecasting host cell mAb productivity using morphological profiles, as inferred from our machine learning models specialized in predictive score prediction and anomaly prediction.

MHC-I upregulation safeguards neoplastic T cells in the skin against NK cell-mediated eradication in mycosis fungoides

Cancer-associated immune dysfunction is a major challenge for effective therapies. The emergence of antibodies targeting tumor cell-surface antigens led to advancements in the treatment of hematopoietic malignancies, particularly blood cancers. Yet their impact is constrained against tumors of hematopoietic origin manifesting in the skin. In this study, we employ a clonality-supervised deep learning methodology to dissect key pathological features implicated in mycosis fungoides, the most common cutaneous T-cell lymphoma. Our investigations unveil the prominence of the IL-32β-major histocompatibility complex (MHC)-I axis as a critical determinant in tumor T-cell immune evasion within the skin microenvironment. In patients' skin, we find MHC-I to detrimentally impact the functionality of natural killer (NK) cells, diminishing antibody-dependent cellular cytotoxicity and promoting resistance of tumor skin T-cells to cell-surface targeting therapies. Through murine experiments in female mice, we demonstrate that disruption of the MHC-I interaction with NK cell inhibitory Ly49 receptors restores NK cell anti-tumor activity and targeted T-cell lymphoma elimination in vivo. These findings underscore the significance of attenuating the MHC-I-dependent immunosuppressive networks within skin tumors. Overall, our study introduces a strategy to reinvigorate NK cell-mediated anti-tumor responses to overcome treatment resistance to existing cell-surface targeted therapies for skin lymphoma.

MSCProfiler: a single cell image processing workflow to investigate mesenchymal stem cell heterogeneity

Single cell cytometry has demonstrated plausible immuno-heterogeneity of mesenchymal stem cells (MSCs) owing to their multivariate stromal origin. To contribute successfully to next-generation stem cell therapeutics, a deeper understanding of their cellular morphology and immunophenotype is important. In this study, the authors describe MSCProfiler, an image analysis pipeline developed using CellProfiler software. This workflow can extract geometrical and texture features such as shape, size, eccentricity and entropy, along with intensity values of the surface markers from multiple single cell images obtained using imaging flow cytometry. This screening pipeline can be used to analyze geometrical and texture features of all types of MSCs across different passages hallmarked by enhanced feature extraction potential from brightfield and fluorescent images of the cells.

Smartphone Imaging Flow Cytometry for High-Throughput Single-Cell Analysis

We present a portable imaging flow cytometer comprising a smartphone, a small-footprint optical framework, and a PDMS-based microfluidic device. Flow cytometric analysis is performed in a sheathless manner via elasto-inertial focusing with a custom-written Android program, integrating a graphical user interface (GUI) that provides a high degree of user control over image acquisition. The proposed system offers two different operational modes. First, "post-processing" mode enables particle/cell sizing at throughputs of up to 67 000 particles/s. Alternatively, "real-time" mode allows for integrated cell/particle classification with machine learning at throughputs of 100 particles/s. To showcase the efficacy of our platform, polystyrene particles are accurately enumerated within heterogeneous populations using the post-processing mode. In real-time mode, an open-source machine learning algorithm is deployed within a custom-developed Android application to classify samples containing cells of similar size but with different morphologies. The flow cytometer can extract high-resolution bright-field images with a spatial resolution <700 nm using the developed machine learning-based algorithm, achieving classification accuracies of 97% and 93% for Jurkat and EL4 cells, respectively. Our results confirm that the smartphone imaging flow cytometer (sIFC) is capable of both enumerating single particles in flow and identifying morphological features with high resolution and minimal hardware.

DeepIFC: Virtual fluorescent labeling of blood cells in imaging flow cytometry data with deep learning

Imaging flow cytometry (IFC) combines flow cytometry with microscopy, allowing rapid characterization of cellular and molecular properties via high-throughput single-cell fluorescent imaging. However, fluorescent labeling is costly and time-consuming. We present a computational method called DeepIFC based on the Inception U-Net neural network architecture, able to generate fluorescent marker images and learn morphological features from IFC brightfield and darkfield images. Furthermore, the DeepIFC workflow identifies cell types from the generated fluorescent images and visualizes the single-cell features generated in a 2D space. We demonstrate that rarer cell types are predicted well when a balanced data set is used to train the model, and the model is able to recognize red blood cells not seen during model training as a distinct entity. In summary, DeepIFC allows accurate cell reconstruction, typing and recognition of unseen cell types from brightfield and darkfield images via virtual fluorescent labeling.

Computer vision meets microfluidics: a label-free method for high-throughput cell analysis

In this paper, we review the integration of microfluidic chips and computer vision, which has great potential to advance research in the life sciences and biology, particularly in the analysis of cell imaging data. Microfluidic chips enable the generation of large amounts of visual data at the single-cell level, while computer vision techniques can rapidly process and analyze these data to extract valuable information about cellular health and function. One of the key advantages of this integrative approach is that it allows for noninvasive and low-damage cellular characterization, which is important for studying delicate or fragile microbial cells. The use of microfluidic chips provides a highly controlled environment for cell growth and manipulation, minimizes experimental variability and improves the accuracy of data analysis. Computer vision can be used to recognize and analyze target species within heterogeneous microbial populations, which is important for understanding the physiological status of cells in complex biological systems. As hardware and artificial intelligence algorithms continue to improve, computer vision is expected to become an increasingly powerful tool for in situ cell analysis. The use of microelectromechanical devices in combination with microfluidic chips and computer vision could enable the development of label-free, automatic, low-cost, and fast cellular information recognition and the high-throughput analysis of cellular responses to different compounds, for broad applications in fields such as drug discovery, diagnostics, and personalized medicine.

Real-time fluorescence imaging flow cytometry enabled by motion deblurring and deep learning algorithms

Fluorescence imaging flow cytometry (IFC) has been demonstrated as a crucial biomedical technique for analyzing specific cell subpopulations from heterogeneous cellular populations. However, the high-speed flow of fluorescent cells leads to motion blur in cell images, making it challenging to identify cell types from the raw images. In this study, we present a real-time single-cell imaging and classification system based on a fluorescence microscope and deep learning algorithm, which is able to directly identify cell types from motion-blur images. To obtain annotated datasets of blurred images for deep learning model training, we developed a motion deblurring algorithm for the reconstruction of blur-free images. To demonstrate the ability of this system, deblurred images of HeLa cells with various fluorescent labels and HeLa cells at different cell cycle stages were acquired. The trained ResNet achieved a high accuracy of 96.6% for single-cell classification of HeLa cells in three different mitotic stages, with a short processing time of only 2 ms. This technology provides a simple way to realize single-cell fluorescence IFC and real-time cell classification, offering significant potential in various biological and medical applications.

Label-free macrophage phenotype classification using machine learning methods

Macrophages are heterogeneous innate immune cells that are functionally shaped by their surrounding microenvironment. Diverse macrophage populations have multifaceted differences related to their morphology, metabolism, expressed markers, and functions, where the identification of the different phenotypes is of an utmost importance in modelling immune response. While expressed markers are the most used signature to classify phenotypes, multiple reports indicate that macrophage morphology and autofluorescence are also valuable clues that can be used in the identification process. In this work, we investigated macrophage autofluorescence as a distinct feature for classifying six different macrophage phenotypes, namely: M0, M1, M2a, M2b, M2c, and M2d. The identification was based on extracted signals from multi-channel/multi-wavelength flow cytometer. To achieve the identification, we constructed a dataset containing 152,438 cell events each having a response vector of 45 optical signals fingerprint. Based on this dataset, we applied different supervised machine learning methods to detect phenotype specific fingerprint from the response vector, where the fully connected neural network architecture provided the highest classification accuracy of 75.8% for the six phenotypes compared simultaneously. Furthermore, by restricting the number of phenotypes in the experiment, the proposed framework produces higher classification accuracies, averaging 92.0%, 91.9%, 84.2%, and 80.4% for a pool of two, three, four, five phenotypes, respectively. These results indicate the potential of the intrinsic autofluorescence for classifying macrophage phenotypes, with the proposed method being quick, simple, and cost-effective way to accelerate the discovery of macrophage phenotypical diversity.

Dimensionality reduction by UMAP for visualizing and aiding in classification of imaging flow cytometry data

Recent advances in imaging flow cytometry (IFC) have revolutionized high-throughput multiparameter analyses at single-cell resolution. Although enabling the discovery of population heterogeneities and the detection of rare events, IFC generates hyperdimensional datasets that demand innovative analytical approaches. Current methods work in a supervised manner, utilize only limited information content, or require large annotated reference datasets. Dimensionality reduction algorithms, including uniform manifold approximation and projection (UMAP), have been successfully applied to analyze the large number of parameters generated in various high-throughput techniques.

Here, we apply a workflow incorporating UMAP to analyze different IFC datasets. We demonstrate that it out-competes other popular dimensionality reduction methods in speed and accuracy. Moreover, it enables fast visualization, clustering, and tagging of unannotated objects in large-scale experiments. We anticipate that our workflow will be a robust method to address complex IFC datasets, either alone or as an upstream addition to the deep learning approaches.

•

UMAP dimensionality reduction provides fast and accurate method of IFC data analysis

•

UMAP yields improved object clustering and tagging of the multispectral IFC data

•

PCA decomposition allows multispectral signals merging for direct image embedding

Imaging approaches for monitoring three-dimensional cell and tissue culture systems

The past decade has seen an increasing demand for more complex, reproducible and physiologically relevant tissue cultures that can mimic the structural and biological features of living tissues. Monitoring the viability, development and responses of such tissues in real-time are challenging due to the complexities of cell culture physical characteristics and the environments in which these cultures need to be maintained in. Significant developments in optics, such as optical manipulation, improved detection and data analysis, have made optical imaging a preferred choice for many three-dimensional (3D) cell culture monitoring applications. The aim of this review is to discuss the challenges associated with imaging and monitoring 3D tissues and cell culture, and highlight topical label-free imaging tools that enable bioengineers and biophysicists to non-invasively characterise engineered living tissues.

Practical machine learning for disease diagnosis

Deep learning neural networks are a powerful tool in the analytical toolbox of modern microscopy, but they come with an exacting requirement for accurately annotated, ground truth cell images. Otesteanu et al. (2021) elegantly streamline this process, implementing network training by using patient-level rather than cell-level disease classification.