PrePrint Jan 2026 De Blander, et al.
Simple cell geometry metrics reveal cancer cell diversity and plasticity and guide combination therapy
Using Deepcell’s REM-I platform, the authors show that high-resolution, label-free imaging of live single cells can reveal meaningful cancer cell states directly from morphology. By profiling melanoma cells with REM-I and linking morphological features to gene-expression programs, they demonstrate that cell shape alone captures lineage identity and functional heterogeneity, enabling identification of distinct tumor subpopulations without staining or molecular perturbation. The study highlights REM-i as a powerful tool for uncovering biologically relevant cell states and transitions through morphology-based, AI-driven single-cell analysis.
PrePrint Jan 2026 Rupp, et al.
A Deep Learning–Enabled Single-Cell Morpholomic Atlas of Nasal Swabs Distinguishes Chronic Inflammation from Sinonasal Malignancy
Using Deepcell’s REM-I platform, this study builds a high-resolution, AI-driven single-cell morphologic atlas from nasal swab samples and shows that label-free cellular morphology can distinguish chronic inflammatory conditions from sinonasal malignancy. By imaging thousands of individual cells and extracting deep morphological features, the authors demonstrate that REM-I captures disease-specific cellular states and tissue composition directly from morphology, enabling accurate separation of malignant and inflammatory samples without molecular staining. The work highlights REM-I’s potential for non-invasive diagnostics and disease monitoring through rapid, morphology-based single-cell analysis of clinical specimens.
PrePrint Jan 2025 Ramathal, et al.
Deep learning-driven morphology analysis enables label-free classification of therapeutic agent-naive versus resistant cancer cells
Using Deepcell’s REM-I platform and AI-driven morphology analysis, this study shows that label-free single-cell imaging can predict functional phenotypes such as drug resistance directly from cell morphology. By training deep-learning classifiers on high-resolution images of multiple cancer cell lines, the authors demonstrate that subtle morphological signatures captured by REM-i distinguish drug-naïve from resistant states and encode complex phenotype information without molecular labels. The work highlights REM-I as a powerful approach for identifying functional cellular responses and resistance phenotypes through high-dimensional, morphology-based single-cell analysis.
Trends in Cell Biology Aug 2024 Kuhn, et al.
Accessible high-speed image-activated cell sorting
This review describes the rise of image-activated cell sorting, where high-speed microscopy is combined with microfluidics and AI to isolate cells based on visual phenotypes rather than molecular labels. It highlights how next-generation platforms, such as REM-I enable real-time identification and sorting of functional cell states from morphology alone, expanding single-cell discovery, functional genomics, and therapeutic cell selection beyond traditional marker-based methods.
PrePrint Jul 2024 Lattmann, et al.
Label-Free Melanoma Phenotype Classification Using Artificial Intelligence-Based Morphological Profiling
Using Deepcell’s REM-I platform and AI-driven morphology analysis, this study demonstrates that single-cell morphology alone can accurately classify melanoma cell phenotypes and functional states. By imaging live cells label-free and training a deep-learning–based melanoma phenotype classifier on REM-I morphological features, the authors show that subtle visual signatures capture lineage and state information that traditionally requires molecular profiling. The work highlights how REM-I enables rapid, non-perturbative identification of tumor cell phenotypes and heterogeneity directly from morphology, supporting morphology-based functional profiling and sorting of clinically relevant cancer cell states.
Biosensors Mar 2025 Du, et al.
Recent Developments (After 2020) in Flow Cytometry Worldwide and Within China
This is a review article summarizing recent developments in flow cytometry instrumentation and techniques, including spectral, mass, imaging, nano, and label-free approaches—that have advanced single-cell analysis for research and clinical applications worldwide and within China. It compares the features and trends of state-of-the-art cytometers, highlights cutting-edge technical improvements, and outlines future directions for the field, emphasizing how high-dimensional, label-free methods are expanding the capabilities of single-cell phenotyping.
Toxics Mar 2025 Jităreanu, et al.
The Evolution of In Vitro Toxicity Assessment Methods for Oral Cavity Tissues—From 2D Cell Cultures to Organ-on-a-Chip
This review highlights the growing role of advanced imaging and AI-based morphology profiling in modern toxicity testing. It notes that platforms like Deepcell’s REM-I enable label-free analysis and sorting of cells based on subtle morphological responses to stress or toxic exposure, positioning morphology-driven single-cell analysis as a powerful tool for next-generation in vitro toxicology.
Lab Chip Jan 2025 Menon, et al.
Microfluidics for morpholomics and spatial omics applications
This review outlines how advanced microfluidic and imaging platforms are driving the field of morpholomics, using single-cell morphology as a high-dimensional biological readout, and highlights technologies like Deepcell’s REM-I as enabling scalable, label-free cell profiling. By integrating precise microfluidic handling with AI-driven image analysis, these platforms capture functional cell states directly from morphology, supporting applications in drug discovery, disease monitoring, and biomanufacturing.
Nature Reviews Jul 2025 Ding, et al.
Image-activated cell sorting
This review article highlights image-activated cell sorting as a new paradigm for single-cell analysis, combining high-speed imaging, microfluidics, and AI to isolate cells based on morphology rather than labels. It positions platforms like Deepcell’s REM-I as enabling real-time, label-free identification and sorting of functional cell states, expanding capabilities for discovery, therapeutic cell selection, and high-content screening.
PLOS Computational Biology July 2024 Navarez, et al.
Using deep learning to decipher the impact of telomerase promoter mutations on the dynamic metastatic morpholome
Using high-resolution imaging and deep-learning analysis—including approaches compatible with Deepcell’s REM-I platform, this study shows that telomerase promoter mutations in melanoma produce subtle but measurable morphological signatures at the single-cell level. By profiling thousands of cells and training AI models on morphology alone, the authors demonstrate that these mutations drive distinct morphologic states linked to metastatic potential and cellular behavior, enabling classification of mutant vs. wild-type cells without molecular labeling. The work highlights how REM-I–style, AI-driven morphology profiling can uncover genotype-to-phenotype relationships and predict functional cancer traits directly from label-free single-cell imaging.
The Journal of Immunology May 2024 Balakrishnan, et al.
Self-supervised deep learning enables label-free, high-dimensional morphology profiling of immune cell types
Using Deepcell’s REM-I platform, this study shows that self-supervised deep learning on high-speed, label-free single-cell images can generate rich morpholomic profiles that distinguish major human immune cell types (e.g., T cells, B cells, monocytes, NK cells) from peripheral blood mononuclear cells. By embedding morphological features into a high-dimensional space, REM-i captures reproducible, quantitative descriptions of immune cell morphology that separate cell subsets in analysis and reflect consistent phenotypes across donors without antibody labeling.
Biomarkers, Immune Monitoring and Novel Technologies Nov 2023 DeGeus, et al.
76 Customizable polymer-based synthetic cells for imaging flow cytometry
Using Deepcell’s REM-I image-based flow cytometry platform, this study demonstrates consistent, high-resolution morphological profiling of synthetic cell systems, showing that label-free imaging can reproducibly characterize cell size, shape, and structural features at single-cell resolution. The work highlights REM-I’s ability to generate quantitative morphology data for engineered or therapeutic cell platforms, supporting morphology-driven characterization and quality assessment of complex cell populations.
Nature Communications Biology Sep 2023 Salek, et al.
COSMOS: a platform for real-time morphology-based, label-free cell sorting using deep learning
This is the inaugural paper describing Deepcell’s underlying technology. It demonstrates that AI-driven, high-resolution single-cell imaging can classify and sort live cells based purely on morphology, showing that deep-learning–derived morphological embeddings capture rich biological identity and enable purification of unlabeled viable cells with desired traits. The work highlights the power of REM-I–style platforms to perform real-time, label-free functional cell characterization and sorting directly from morphology, bridging imaging and downstream molecular analysis.
Modern Pathology Aug 2023 Mavropoulos, et al.
Artificial Intelligence-Driven Morphology-Based Enrichment of Malignant Cells from Body Fluid
This study highlights the power of AI-driven, image-based cell profiling to extract high-dimensional biological information directly from single-cell morphology. It shows how next-generation platforms such as Deepcell’s REM-I enable label-free identification and functional characterization of heterogeneous cell populations, using morphological signatures to uncover cell states and behaviors that traditional marker-based methods can miss, with applications across discovery, diagnostics, and cell therapy.
Journal of the American Society of Cytopathology Nov 2022 Kim, et al.
Molecular Study on Enriched Tumor Cells from Body Fluids via a Label-Free COSMOS Method – A Feasibility Testing of “SMART” (Single cell, Multiplex, AI-based, Real-Time) Cytology
This review describes the emergence of image-activated cell sorting technologies that combine high-speed microscopy, microfluidics, and machine learning to isolate cells based on visual phenotypes rather than molecular labels. It highlights how platforms like Deepcell’s REM-I enable real-time, label-free identification and sorting of functional cell states directly from morphology, expanding capabilities for single-cell discovery, therapeutic cell selection, and high-content screening.