An important step to translate single-cell measurement into clinical practice: Stereoscopic cells

Clinical single-cell measurements are gaining increasing attention of clinicians and researchers to conventional parameters in clinical biochemistry of haematology. However, significant challenges need to be overcome before these technologies can be fully integrated into clinical practice.¹ With the rapid development of biotechnology and methodology, single-cell multi-omics and trans-omics analyses have become increasingly feasible, stable and reproducible. The comprehensive profiles generated at the single-cell level provide novel insights into molecular phenomes, functional states, microenvironmental configurations, disease mechanisms and potential therapeutic targets. This emerging field—termed clinical single-cell biomedicine—is opening new frontiers in precision medicine.² In particular, spatial transcriptome furthermore enables the spatiotemporal cellular and molecular mapping, which brings the development of functional histopathology, by defining the precise locations and interactions of target cells, and integrating clinical image phenomes.^{3, 4} A spatial enhanced resolution omics-sequencing, Stereo-Seq (Figure 1B), utilises DNA nanoball-patterned arrays for high-resolution in situ RNA capture to demonstrate tissue/organ spatiotemporal transcriptomes at a single-cell level.⁵ This technology enables multi-dimensional, single-cell–level visualisation of transcriptomic data across tissues and organs. By integrating continuous transcriptomic images, stereoscopic molecular maps of cells can be reconstructed, providing detailed insights into cellular heterogeneity, lineage trajectories and functional states.^{5, 6} Such stereoscopic and temporal resolution of transcriptomics enables the identification of spatiotemporal variations, classifications, architectonics and functional communications of cells, all of which can be translated into clinical strategies. The application of single-cell multi-omics, combined with multi-dimensional genomic data, to human pathology has been proposed as a transformative milestone in stereoscopic diagnosis and therapeutic strategies.⁷

Stereo-Cell is a breakthrough spatially enhanced-resolution single-cell sequencing technology utilising high-density DNA nanoball-patterned arrays (Figure 1A). It enables the capture of intact individual cells or microstructures, and facilitates the profiling of stereoscopic, spatiotemporal multi-omics at true single-cell resolution.⁸ Unlike Stereo-Seq (Figure 1B), Stereo-Cell is designed to acquire stereoscopic multi-omic information from entire, structurally preserved cells, including complete morphological characteristics, surface protein profiles and intracellular molecular landscapes (Figure 1A). In contrast to conventional instrument-based cell separation methods, the DNA nanoball-oriented chips used in Stereo-Cell are capable of capturing all categories of cells with a wider range of input cell numbers. They demonstrate strong capability with other methodologies and can accommodate diverse target cell sizes and morphologies, from extracellular vesicles to large, complex cells. This includes cell types that are typically challenging to be captured, such as long multinucleated cardiac muscle cells, neurons with extended neurites, telocytes with long telopodes and skeletal myofibre. Different from single-cell RNA sequencing (scRNA-seq), Stereo-Cell, along with its methodological extensions, provides detailed information on cell morphological phenomes, protein patterns and membrane functions as integral aspects of cell identity, which are more readily adopted and defined within the knowledge framework of clinical biochemistry of haematology.¹ Those individual cells captured on the chip can be stained, re-traced and re-visualised to reveal their morphometric features, organelle structure and localisation, and multi-omic profiles. This capability paves the way for the conceptualisation of clinical artificial intelligent single cells (caiSCs), achieved by integrating clinical phenomes and multiple foundation models and datasets.^{9, 10} The whole-cell models created by caiSCs can simulate disease progression, immune responses or drug effects at the cellular level. The Stereo-Cell chip can serve as a potential culture medium chip, providing validation schemes for the whole-cell models developed by caiSCs. The validation results can further offer feedback to refine and improve the whole-cell models.

Stereo-Cell represents a crucial advancement in translating single-cell measurements into clinical practice, since it enables the capture of isolated and individually separated single cells, including those derived from blood and body fluids such as cerebrospinal fluid, bronchoalveolar lavage, pleural effusion, ascites or urine (Figure 1A). Stereo-Cell's unbiased capture capability enhances the detection of rare cell types and molecules, with the detection of circulating tumour cells (CTCs) and exosomes in the blood (Figure 1). Furthermore, Stereo-Cell can capture pre-existing cell islands in body fluids, such as erythroid islands in the bone marrow and CTC clusters, thereby facilitating the analysis of cellular interactions within the fluids. The concept and team of Stereo-Cell recently emerged require broader recognition and acceptance by experts in biology, systems biology, computational biology and related disciplines. As a novel approach, Stereo-Cell offers a comprehensive means of investigating the stereoscopic whole cell, as mentioned during the discovery and development of Stereo-Seq with the high-resolution imaging and sequencing.^{5, 6} Stereo-Seq generates high-resolution images of stereoscopic cells within tissues or organs, capturing detailed information on cellular connections and interactions in the microenvironment (Figure 1B). Together, Stereo-Cell and Stereo-Seq usher in a new ‘Stereo’ era of single-cell measurements a cutting-edge computational and biological framework designed to produce high-confidence, spatially resolved and intact single-cell data. This framework offers a comprehensive and systematic solution for processing and analysing multi-dimensional spatial data. With the increasing knowledge on Stereo-science, a new discipline of biomedicine named ‘Stereo-biomedicine’ will arise.

Stereo-Seq combines a series of tissue section images into a single, high-resolution stereoscopic image, minimising stitching errors, improving overall image quality and signal-to-noise ratio, ensuring precise spatial alignment and accurate segmentation, and enabling the visualisation of cellular interactions and gene expression patterns within the tissue context.¹¹ Stereo-Cell, on the other hand, incorporates immunofluorescent staining and imaging to detect both homotypic and heterotypic doublets, ensuring accurate transcript assignment to each intact cell and the detailed characterisation of stereoscopic, whole-cell features. It enables spatial visualisation of superimposed nuclei and the distribution of captured transcripts, providing reliable, accurate and flexible multi-omic profiles at the level of individual whole cells.⁸ Stereo-Seq describes stereoscopic cells with detailed gene expression maps at single-cell level in large tissue sections, providing insights into tissue architecture and cellular heterogeneity. Stereo-Cell enriches the concept of stereoscopic cell imaging by enabling in-depth three-dimensional (3D) visualisation of cells. This includes the analysis of biological structures and behaviours, the spatiotemporal arrangement and distribution of biomolecules (e.g., DNA, RNA, proteins or metabolites), as well as their interactions and functions in large-scale chips.

Stereo-Cell has addressed several critical scientific and technological limitations of scRNA-seq and Stereo-Seq by offering a more intuitive data and clinically interpretable results that are more easily accepted and understood by biologists and clinicians. One of the key challenges in single-cell analysis lies in the definition, validation and application criteria of single-cell identity marker gene panels (ciMGPs), which are commonly employed to delineate cell populations, subtypes and functional states. Persistent concerns have been raised about whether these ciMGPs are truly specific to cell subtypes or states, and whether they adequately reflect the underlying biological characteristics, despite the development of automated computational annotation tools aimed at expediting identification and improving accuracy and specificity in cell type classification.¹² To minimise redundancy in the usages of ciMGPs usage across multiple subtypes or functional states, the overlap expression rate has been introduced as a metric to evaluate their specificity, categorising them as cell-specific, cell-associated or cell-reference genes.¹³ Distinctively, Stereo-Cell enables the identification of cell types and subtypes/states through the integrated analysis of ciMGPs, cell morphometrics and cell-specific protein markers, two of which have already been widely applied in clinical practice for decades.

There is growing evidence supporting the potential translation of Stereo-Cell into clinical biochemistry of haematology and of Stereo-Seq into clinical molecular pathology. Nonetheless, several key challenges remain to be solved, including prolonged turnaround times, high-costs, variable capture efficiency and stability, technical complexity and limited repeatability, as well as the need for integrated and comprehensive data analyses. One of the most important issues in translation of Stereo-Cell and Stereo-Seq to clinical practice lies in standardising the entire workflow, from initial clinical sampling to the final report generation. Despite these challenges, Stereo-Cell's capability for rare cell captures and simultaneous provision of image and panoramic molecular information highlights its significant potential in clinical applications. The integration of artificial intelligence algorithms with image processing and molecular analysis techniques will enable the development of an automated analysis and diagnosis model, enhancing automation, accuracy and efficiency for clinical use. Stereoscopic cells, equipped with 3D molecular architectures and multi-omic profiles, are expected to visualise cellular processes and stereotypical behaviours within the extracellular matrix to elucidate underlying mechanisms. Furthermore, they are also expected to determine the stereochemistry of drug candidates and their binding affinity to intracellular targets for drug discovery, and track the spatiotemporal biochemical reactions and metabolic activities for precision therapy.

We believe that Stereo-Cell/Seq, supported by caiSCs, can provide rapid and comprehensive reporting of disease-driven indications, offer predictable patterns of stereoscopic cells in response to specific stimuli, and yield new insights into developmental biology and disease pathogenesis. Furthermore, this approach enables dynamic monitoring of disease progression, severity and prognosis. As Stereo-Cell and Stereo-Seq technologies continue to mature, stereoscopic cells with integrated multi-omic profiles and disease-specific phenomes may emerge as a new fundamental unit in Stereo-biomedicine, offering a powerful and indispensable tool for clinical and translational medicine.

Xiangdong Wang and Jia Fan took the full responsbilities for ideas, designs and writing. Xuanqi Liu and Wanxin Duan are responsible for the writing and correction.

The authors declare no conflicts of interest.

Not Applicable.

The authors have nothing to report.