Decoding tissue complexity: multiscale mapping of chemistry–structure–function relationships through advanced visualization technologies

Abstract

Comprehensively acquiring biological tissue information is pivotal for advancing our understanding of biological systems, elucidating disease mechanisms, and developing innovative clinical strategies. Biological tissues, as nature's archetypal biomaterials, exhibit multiscale structural and functional complexity that provides critical principles for synthetic biomaterials. Tissues/organs integrate molecular, biomechanical, and hierarchical architectural features across scales, offering a blueprint for engineering functional materials capable of mimicking or interfacing with living systems. Biological visualization technologies have emerged as indispensable tools for decoding tissue complexity, leveraging their unique technical advantages and multidimensional analytical capabilities to bridge the gap between macroscopic observations and molecular insights. The integration of cutting-edge technologies such as artificial intelligence (AI), augmented reality, and deep learning is revolutionizing the field and enabling real-time, high-resolution, and predictive analyses that transcend the limitations of traditional imaging modalities. This review systematically explores the principles, applications, and limitations of state-of-the-art biological visualization technologies, with a particular emphasis on the transformative advancements in AI-driven image analysis, multidimensional imaging and reconstruction, and multimodal data integration. By analyzing these technological trends, we envision a future where biological visualization evolves towards greater intelligence, multidimensionality, and multiscale precision, offering unprecedented theoretical and methodological support for deciphering tissue complexity and further advancing biomaterials development. These advancements promise to accelerate breakthroughs in precision medicine, tissue engineering, and therapeutic development, ultimately reshaping the landscape of biomedical research and clinical practice.

---

Zhiyuan Zhao

Zhiyuan Zhao received her undergraduate certificate from Shandong Second Medical University in China in the summer of 2023. After graduation, she joined the biomanufacturing and regenerative medicine research group at Chongqing University. Her current research focuses on 3D reconstruction of biological tissue information.

Haijun Cui

Haijun Cui is currently an associate professor at the College of Bioengineering, Chongqing University. He received his PhD in 2018 under the supervision of Prof. Shutao Wang from the Technical Institute of Physics and Chemistry, Chinese Academy of Sciences (TIPCCAS). From 2018 to 2021, he conducted postdoctoral research at Karlsruhe Institute of Technology (KIT), Germany (supervisor: Prof. Pavel A. Levkin). His research interests include the development of biomimetic biomaterials and the construction of cell spheroid-based cancer models.

Haitao Cui

Haitao Cui is a professor at the College of Bioengineering, Chongqing University. He did graduate research at Changchun Institute of Applied Chemistry, Chinese Academy of Sciences, and earned his PhD from the University of Chinese Academy of Sciences in 2015. He completed postdoc training and then worked as a Research Scientist at The George Washington University before joining Chongqing University. His research focuses on biomaterials and biomanufacturing technologies for tissue regeneration and disease studies.

---

1. Introduction

The multidimensional characterization of biological tissues/organs represents a cornerstone in modern biomedical research, driving advances in precision medicine, tissue engineering, and therapeutic development. As nature's optimized biomaterials, native tissues provide a gold standard for designing synthetic counterparts—their hierarchical architectures, mechanical gradients, and dynamic cell–matrix interactions offer critical blueprints for engineering bioactive scaffolds and adaptive implants. Understanding biological systems, elucidating disease mechanisms, and developing effective clinical strategies require comprehensive tissue data, integrating both structural and molecular information.^1–4 Tissues/organs exhibit remarkable complexity, with diverse structures and functions operating across multiple scales. To fully comprehend these complexities, it is essential to integrate both structural and molecular data, capturing not only the architecture of tissues but also the molecular interactions that drive their functions.^3,4 This integrative paradigm has become particularly critical in addressing complex pathological conditions such as cancer metastasis and neurodegenerative diseases, where structural deformations and molecular dysregulations coexist.

Imaging technologies have revolutionized our ability to decode tissue complexity, bridging macroscopic observations with molecular insights to characterize the morphology and three-dimensional (3D) configurations of biological tissues/organs across multiple scales, from macro- to nano-level resolutions.^5–7 Imaging modalities enable researchers and clinicians to obtain comprehensive structural and morphological data on biological tissues/organs, thereby facilitating deeper insights into their characteristics, molecular profiles, and functional dynamics.⁸

Traditional modalities, such as computed tomography (CT),⁹ magnetic resonance imaging (MRI),¹⁰ and ultrasound,¹¹ are extensively employed in clinical practice for visualizing large-scale tissue structures in living organisms. Notably, MRI demonstrates particular efficacy in imaging soft tissues (e.g., the brain, muscles, and the heart) through its superior contrast, while CT excels in providing detailed views of dense tissues (e.g., bone) via X-ray attenuation measurements. Ultrasound is predominantly utilized for real-time imaging applications, particularly in fetal monitoring, organ imaging, and biopsy procedures. The complementary nature of these techniques has revolutionized diagnostic precision, allowing the noninvasive monitoring of tissue architecture, disease progression, and injury.

To address the limitations of macroscopic imaging, researchers have developed high-resolution methods for cellular and molecular analyses. Histological staining techniques, particularly immunohistochemistry (IHC), are widely employed to visualize the distribution and localization of specific proteins within tissue samples.¹² Advances in multiplexed IHC now enable the simultaneous detection of multiple biomarkers in a single tissue section, dramatically enhancing our ability to map cellular microenvironments. Microscopy, including optical and electron microscopy (EM), facilitates tissue examination across various magnifications, with electron microscopy revealing subcellular structures at ultra-high-resolution, including organelles, intercellular connections, and membranes.¹³ Additionally, mechanical testing methodologies,¹⁴ such as tensile and compressive tests, characterize the mechanical properties of tissues, simulating the complex mechanical loads, i.e., friction, rotation, and shear, that tissues experience in their native environments.¹⁵ These approaches have spurred the growth of mechanobiology, uncovering how physical forces regulate cellular behavior and revolutionizing the rational design of biomaterials, e.g., engineering tissue-specific stiffness gradients in implants, and mapping hemodynamic shear stress profiles to guide the development of vascular grafts.

The convergence of computational technologies with imaging has ushered in a new era of intelligent tissue analysis. Artificial intelligence (AI), augmented reality (AR), and deep learning have revolutionized biological data interpretation.¹⁶ For instance, generative adversarial networks (GANs) are being employed to predict 3D tissue architectures from 2D histological sections, bridging the gap between microscopy and volumetric imaging. These emerging technologies enhance image analysis, pattern recognition, and predictive modeling, thereby facilitating more accurate and dynamic representations of biological tissues/organs.^17,18 The shift from conventional two-dimensional (2D) to 3D imaging techniques, such as 3D microscopy and volumetric imaging, has also enabled the reconstruction of tissue structures in more realistic and detailed ways, further improving our understanding of their spatial organization.^19–21In vivo imaging systems allow real-time observation of dynamic processes such as immune cell migration and tumor progression.

This review highlights current technologies and future directions in biological visualization. While traditional imaging techniques, such as CT, MRI, ultrasound, and histology, have been extensively reviewed,^12,22–24 our focus is on the latest innovations and the convergence of cutting-edge technologies. Moreover, 3D imaging reconstruction with AR enables real-time, immersive visualization of biological tissues, integrating high-fidelity anatomical models with multimodal data for enhanced analysis, revolutionizing personalized surgical planning, medical training, and patient education, while the strategic integration of multimodal imaging platforms with AI-driven analytics represents a paradigm shift, enabling predictive modeling of disease progression and therapeutic responses. We reveal that the next frontier in biological visualization lies in the development of multiscale, multi-omics integration platforms capable of correlating genomic, proteomic, and biomechanical data within spatially resolved tissue architectures. We anticipate that biological visualization technologies will continue to evolve towards greater intelligence and multidimensionality, thereby providing robust theoretical support for the comprehensive interpretation of biological tissues/organs and enabling the reverse engineering of biomaterials that mimic tissue responses (Fig. 1). These advancements promise to accelerate the translation of basic research findings into clinical applications, ultimately empowering personalized medicine and revolutionizing healthcare delivery.

Fig. 1 Overview of biological visualization techniques. The flowchart illustrates the systematic progression from addressing the fundamental human need to comprehend biological tissues to achieving a comprehensive characterization of their intricate molecule, structure, and function.

2. Advances in conventional imaging modalities for tissue visualization

Traditional imaging modalities, including CT, MRI, ultrasound, histological staining, and immunostaining, form the cornerstone of clinical diagnostics and biomedical research. CT employs X-ray technology to deliver high-resolution visualization of dense tissues such as bone, while MRI harnesses magnetic fields and radiofrequency pulses to generate superior soft tissue contrast, particularly for neurological and musculoskeletal imaging. In contrast, ultrasound operates via non-ionizing sound waves, offering real-time dynamic imaging that has become indispensable in obstetrics and cardiology. At the cellular level, histological staining techniques remain foundational for elucidating tissue architecture, while immunostaining enables precise localization of disease-specific biomarkers, revolutionizing the diagnostic paradigms in oncology, infection diseases, and autoimmune disorders. Recent technological evolution, exemplified by digital pathology platforms, AI-driven virtual staining, and multiplexed immunofluorescence (mIF), has elevated these classical methods to new heights – accelerating diagnostic workflows, enhancing analytical precision, and unmasking critical insights into tissue heterogeneity and spatial molecular organization. Table 1 systematically compares the technical specifications, clinical applications, and advancements of these modalities.

Table 1 Overview of traditional imaging techniques for biological tissue analysis

Imaging technique	Principle	Applications	Resolution (size level)	Scope of use	Advantages	Limitations	Cost	Time efficiency
CT	X-ray attenuation to create cross-sectional images	Neurology, oncology, cardiology, trauma, orthopedic diagnostics	∼1 mm (tissue level)	Internal organ imaging, e.g., bone	High resolution for dense tissues; fast acquisition	Ionizing radiation; limited soft tissue contrast	Moderate	Fast
MRI	Magnetic fields and radiofrequency pulses to produce detailed images	Neurology, musculoskeletal, cardiovascular diagnostics	0.5–1 mm (soft tissue level)	Brain, spinal cord, muscles, and other soft tissues	Non-invasive; excellent soft tissue contrast; no ionizing radiation	High cost; time-consuming; limited utility for dense tissues	High	Moderate
Ultrasound imaging	High-frequency sound waves to create real-time images	Obstetrics, cardiology, musculoskeletal diagnostics	∼1–5 mm (organ/tissue level)	Real-time diagnostics; pregnancy monitoring	Non-invasive; portable; cost-effective; no radiation	Limited penetration depth; operator-dependent image quality	Low	Very fast (real-time)
Histological staining	Chemical dyes to highlight tissue structures	Pathology, histology, biomedical research	Micrometer to nanometer range	Tissue biopsy examination	Detailed visualization on tissue architecture; customizable staining protocols	Time-consuming; tissue preparation and processing	Low	Slow (hours to days)
Immunostaining	Antibody–antigen binding with chromogenic/fluorescent labels	Cancer diagnosis, autoimmune diseases, infectious disease research	Micrometer level (protein localization)	Tissue biopsy; molecular pathology	High specificity for target protein; molecular-level analysis	Time-consuming; expensive reagents; potential for non-specific binding and background noise	Moderate	Slow (hours to days)

---

2.1. CT: from cross-sectional imaging to AI-driven innovations

CT is a widely used imaging modality in radiology that combines X-ray technology with advanced processing to generate detailed cross-sectional body images.^25,26 Unlike conventional radiography with its tissue overlap and limited contrast, CT enables high-clarity visualization of internal structures noninvasively.²⁷ Since Cormack's pioneering work in 1964,²⁸ CT technology has evolved from early computer-assisted tomography to advanced iterations like spiral CT,²⁹ multidetector-row CT,^30,31 dual-source CT,³² and photon-counting CT (PCCT), each enhancing speed, resolution, and diagnostic capabilities.

PCCT represents a paradigm shift in CT technology, distinguishing itself from traditional energy-integrating detectors by directly converting X-ray photons into electrical signals, and sorting them into multiple energy bins (Fig. 2A and B).²³ This innovative approach enables enhanced image quality and higher spatial resolution,³³ as demonstrated in non-contrast imaging of temporal bones, where PCCT has achieved remarkable visualization of intricate inner ear structures in both cadaveric specimens and human subjects (Fig. 2C and D).^34,35 The incorporation of tin filtration, though not exclusive to PCCT, has further reduced radiation doses by up to 85% in temporal bone and sinus imaging.³⁶ In musculoskeletal applications, PCCT has demonstrated lower image noise, higher bone signal-to-noise ratios, and sharper edge delineation compared to conventional CT, as evidenced by studies on human cadaveric specimens^37,38 and clinical patients.^38,39 With spatial resolutions as fine as 125 μm, PCCT is particularly advantageous for imaging small anatomical regions such as extremities.³⁹ In coronary CT angiography, PCCT outperforms traditional CT by minimizing metallic artifacts, improving the depiction of calcifications, stents, and coronary lumens, and enhancing the detection of low-contrast structures such as non-calcified plaques.^40–42

Fig. 2 Applications of conventional imaging techniques in disease diagnosis. (A) Schematic illustration of a scintillation energy integrating detector and (B) an energy-resolving photon-counting CT (PCCT) (reproduced from ref. 23 with permission from the Radiological Society of North America, copyright 2023). MRI of the kidney tissue, showing its ability to evaluate both tissue structure (C) and function (D) under normal and pathological conditions (reproduced from ref. 43 with permission from Elsevier, copyright 2022). Comparison between a traditional ultrasound device (E) and a bioadhesive ultrasound (BAUS) device (F). (G) BAUS imaging demonstrating reliable long-term monitoring of cardiac motion (reproduced from ref. 44 with permission from Science, copyright 2022).

The integration of AI into CT imaging is rapidly transforming the field, automating critical processes such as data acquisition, image segmentation, and quantitative analysis. For instance, AI algorithms are revolutionizing coronary artery CT by enabling precise artery segmentation and quantifying coronary flow reserve.²³ Additionally, AI-assisted material decomposition facilitates the automatic generation of quantitative maps, significantly improving the detection and characterization of pathological conditions. By streamlining image interpretation and automating routine tasks, AI can enhance the efficiency and diagnostic accuracy of CT imaging.

Despite these remarkable advancements, challenges remain. While CT technology has achieved higher-resolution images and reduced radiation exposure, the persistent use of ionizing radiation continues to raise concerns, particularly in cases requiring frequent scans. Nevertheless, the ongoing evolution of CT, driven by innovations such as PCCT and AI integration, underscores its indispensable role in modern diagnostic imaging.

2.2. MRI: from soft tissue visualization to AI-driven precision diagnostics

MRI is a non-invasive imaging technique that uses magnetic fields⁴³ and radiofrequency pulses to produce high-resolution images without ionizing radiation.²⁴ Its exceptional soft tissue contrast and ability to combine structural and functional data make it indispensable in neurology, cardiology, and musculoskeletal imaging. The fundamental principle of MRI involves placing the patient within a strong external magnetic field, which aligns the nuclei of many atoms (i.e., hydrogen) in the body. Subsequent application of RF pulses disrupts this alignment, and the energy released during realignment is detected and processed by a computer to generate detailed MR images.⁴⁵ The origins of MRI date back to the 1940s, with significant advancements in the 1970s transitioning it from a laboratory technique to a clinical powerhouse. By the 1990s, the advent of high-field MRI systems markedly improved spatial resolution and image quality, revolutionizing disease detection and diagnosis.

In recent decades, MRI technology has undergone remarkable advancements. The 2000s witnessed the emergence of functional MRI (fMRI), a groundbreaking tool for mapping brain activity by detecting changes in blood oxygenation associated with neural function. More recently, multiparametric MRI (mpMRI)⁴⁶ has further expanded the capabilities of MRI by enabling the simultaneous acquisition of multiple quantitative metrics, such as tissue microstructure, oxygenation levels, and blood flow, within a single scan session (Fig. 2E and F).⁴³ Cardiac magnetic resonance (CMR) exemplifies the versatility of modern MRI, offering comprehensive, radiation-free imaging of the heart. CMR provides critical insights into cardiac anatomy, wall motion, valve function, blood flow, myocardial perfusion, cardiac metabolism, and coronary artery health, making it an invaluable tool for diagnosing and managing cardiovascular diseases.⁴⁷ As CMR technology continues to evolve, its role in early diagnosis and personalized treatment strategies is expected to grow significantly.

The integration of AI into MRI has catalyzed a paradigm shift in medical imaging, revolutionizing image processing capabilities and significantly enhancing diagnostic precision. AI algorithms excel at processing and analyzing the vast datasets generated by MRI, identifying key morphological and kinetic features of diseases, and assisting radiologists in differentiating benign from malignant lesions. For instance, Jiang et al. demonstrated that AI-assisted interpretation of dynamic contrast-enhanced breast MRI significantly improves the efficiency and accuracy of distinguishing cancerous from noncancerous lesions.⁴⁸ Beyond image analysis, AI is optimizing MRI scanning techniques, such as non-Cartesian k-space sampling trajectories and advanced image reconstruction methods,⁴⁹ while automating scan prescription processes.^50,51 These innovations are streamlining clinical workflows, reducing scan times, and enhancing diagnostic precision.^52,53

Despite its unparalleled advantages in soft tissue imaging and radiation-free operation, MRI faces several challenges, including prolonged scan times, high operational costs, and patient-related issues such as claustrophobia and movement artifacts. However, ongoing advancements in AI, hardware design, and imaging protocols are addressing these limitations, promising faster, more accurate, and more patient-friendly MRI systems. As these innovations continue to unfold, MRI is poised to remain at the forefront of diagnostic imaging, driving advancements in precision medicine and personalized healthcare.

2.3. Ultrasound imaging: from real-time diagnostics to wearable and AI integration

Ultrasound imaging uses high-frequency sound waves⁵⁴ and a transducer to capture tissue echoes, which are converted into real-time, radiation-free images based on variations in echo patterns.⁵⁵

Since its introduction over 40 years ago, ultrasound has become prominent in obstetrics, cardiology, and musculoskeletal imaging. The development of grayscale ultrasonography in the 1970s^56–58 and the advent of the color Doppler flow-mapping system in 1985⁵⁹ marked significant milestones. Currently, high-end ultrasonic systems integrate continuous-wave (CW) and pulsed-wave (PW) flow measurements, further enhancing their diagnostic utility.

Recent advancements in ultrasound technology have focused on miniaturization and portability, where pocket-sized scanners and wireless systems now enable real-time imaging in various settings, from emergency rooms to remote locations. A particularly promising development is the emergency of wearable ultrasound devices to discern human physiology; for example, cardiac ultrasound imagers are capable of monitoring heart function in real time through deep learning algorithms to track parameters like stroke volume and cardiac output (Fig. 2G).^44,60,61 Other breakthroughs include wearable ultrasound patches for bladder volume monitoring and flexible probes for deep tissue signal tracking,^44,62,63 paving the way for continuous, non-invasive physiological monitoring.

The integration of AI into ultrasound imaging is revolutionizing the field by improving image quality, automating interpretation, and enhancing diagnostic accuracy. AI algorithms (machine learning or deep learning) are being employed to detect acoustic shadows, reduce image noise (despeckling),^64,65 and enhance image clarity.⁶⁶ These advancements are enabling smarter, more portable devices that continuously learn and adapt, improving diagnostic capabilities over time.²² Empowered by deep learning, AI-driven systems are facilitating automated image analysis, reducing the reliance on operator expertise, and making ultrasound more accessible in resource-limited settings.⁶⁷

Generally, ultrasound has evolved into a versatile, cost-effective imaging modality with broad clinical applications. Recent advancements in portability, wearable technology, and AI integration are poised to further expand its utility, making it a powerful tool for real-time diagnostics and continuous monitoring. However, challenges remain, including limitations in imaging depth and the need for large datasets to train AI algorithms. Despite these hurdles, the ongoing convergence of ultrasound with AI and wearable technologies promises to revolutionize medical imaging, offering improved diagnostics and greater accessibility worldwide.

2.4. Histological staining: from traditional techniques to digital innovations

Histological staining is universally recognized as the gold standard for tissue examination in clinical pathology and life science research. By employing chromatic dyes or fluorescence labels, this technique enables the visualization of tissue and cellular structures, providing critical insights into tissue morphology under microscopic examination.^12,68,69 Over the past century, histological staining has become an indispensable tool,⁷⁰ offering foundational knowledge for disease diagnosis and biological research.

Traditional staining methods, such as hematoxylin and eosin (H&E), remain the most widely used in histopathology due to their ability to contrast nuclei with extracellular matrices (ECM).⁷¹ Specialized stains such as Masson's trichrome (MT)⁷² and periodic acid-Schiff (PAS)⁷³ are used to highlight specific tissue components, i.e., collagen fibers and glycoproteins, proving particularly valuable in cardiac and kidney pathology.

Recent advancements have introduced innovative alternatives to conventional histochemical staining, including nonlinear microscopy,⁷⁴ Raman scattering,⁷⁵ programmable supercontinuum pulses,⁷⁶ and reflectance confocal microscopy.⁷⁷ Besides, digital pathology^78,79 has emerged as a transformative approach, leveraging automated high-throughput slide scanners, digital image viewers, and deep learning algorithms to generate virtual histological stains. These innovations enable multiple staining types to be digitally applied to a single tissue image without the need for physical dyes.^80–84 For example, Zhang et al. presented a new deep-learning-based framework that generates virtually stained images from label-free tissue samples. This approach allows multiple stains to be digitally applied to the same tissue section, which is unachievable with traditional staining methods (Fig. 3A).⁸⁵ Similarly, Haan et al. demonstrated the use of supervised learning to transform H&E-stained images into specialized stains (e.g., MT, PAS, and Jones silver stain) for kidney biopsies, improving diagnostic accuracy comparable to traditional staining, while significantly reducing time and cost.⁸⁶ Deep learning techniques created new opportunities for digitally generating histological stains, providing rapid, cost-effective, and accurate alternatives to standard chemical staining methods.

Fig. 3 Emerging staining techniques developed for tissue observation. (A) Comparison of traditional chemical staining of kidney tissue with virtual staining, exemplified by H&E and MT staining (reproduced from ref. 85 with permission from Nature, copyright 2020). Comparative multiplex immunofluorescence of immune cell infiltration in paired gastric cancer tissue specimens before (B) and after (C) neoadjuvant chemotherapy (NAC) (reproduced from ref. 87 with permission from BMJ Journals, copyright 2022). (D) Thermally accelerated deep immunostaining applied to brain tissue (reproduced from ref. 88 with permission from Nature, copyright 2022). Scale bars, 500 μm.

While histological staining techniques have been integral to tissue examination for over a century, the integration of digital pathology and machine learning is revolutionizing the field. These advancements offer faster, more efficient, and cost-effective alternatives, paving the way for enhanced diagnostic and research capabilities.

2.5. Immunostaining: from protein detection to multiplex and deep tissue analysis

Immunostaining is a specialized technique for identifying specific proteins in tissue samples by utilizing antibodies that bind to target antigens, followed by visualization through chromogenic or fluorescent labels. The two primary methods are IHC and immunofluorescence (IF),⁸⁹ each offering unique advantages for clinical diagnostics and research.

IHC uses enzyme-linked detection systems to visualize antigen–antibody interactions under light microscopy. It is widely used in diagnosing cancers, infectious diseases, and skin tumors. In dermatologic research, IHC identifies tissue-bound autoantibodies, pathogenic RNA in fresh or paraffin-embedded tissues, DNA aberrations, and specific antigens.⁹⁰ IF, on the other hand, relies on fluorescence to detect autoantibody–antigen complexes, making it particularly valuable in dermatology and autoimmune disease diagnosis. Techniques including direct immunofluorescence (DIF), indirect immunofluorescence (IIF), and salt-split skin are employed based on the clinical scenario. For instance, DIF is ideal for detecting tissue-bound autoantibodies, while IIF is used for identifying circulating autoantibodies.⁹¹ In DIF, the primary antibody is directly conjugated to a fluorescent label, whereas IIF uses a labeled secondary antibody to amplify the signal, leading to increased sensitivity.⁹² However, DIF often requires higher antibody concentrations to obtain sufficient signals due to limited label options, which can increase costs and complexity.

Recent innovations, such as mIF, enable the simultaneous detection of multiple biomarkers within a single tissue sample, offering unprecedented insights into tissue heterogeneity and cellular interactions (Fig. 3B and C).⁸⁷ Complementary techniques, including tyramide signal amplification (TSA) and multispectral imaging, further enhance mIF by improving signal detection accuracy and enabling precise quantification of target proteins.^93,94

Despite their widespread adoption, traditional immunostaining methods face inherent limitations, particularly in tissue depth penetration, due to the constraints of light-based imaging. To address this challenge, novel developments, such as SPEARs (chemically engineered antibodies), have emerged, facilitating deeper tissue penetration and a more comprehensive analysis of complex biological structures (Fig. 3D).⁸⁸ Continuous advancements in multiplexing, signal amplification, and deep tissue imaging are improving the sensitivity, precision, and applicability of these techniques, solidifying their role as powerful tools for unraveling tissue biology and disease mechanisms.

3. Advanced imaging techniques for detailed tissue analysis

Advanced imaging techniques have revolutionized the detailed analysis of biological tissues, providing high-resolution insights into their structural and functional properties across multiple scales. These technologies enable researchers to explore tissues at unprecedented levels of detail, from subcellular structures to complex 3D architectures. The following sections delve into key advanced imaging technologies, including electron microscopy, confocal and multiphoton microscopy, and mechanical testing. Each technique offers unique advantages and applications, collectively contributing to a comprehensive understanding of tissue biology (Table 2).

Table 2 Overview of advanced imaging and mechanical testing techniques for biological tissue analysis

Technique	Principle	Applications	Resolution (size level)	Scope of use	Advantages	Limitations	Cost	Time efficiency
EM	Uses electron beams to create high-resolution images	Subcellular structures, organelles, membranes	50× to 10^−6× (nanoscale)	Biomaterials, cellular structures	Ultra-high resolution, detailed structural information	Requires vacuum, complex sample preparation	High	Moderate
Confocal microscopy	Uses laser scanning to reduce out-of-focus light	3D imaging of cells and tissues	Micrometer level	Biological samples, cellular imaging	High-quality 3D images, non-invasive optical slicing	Limited depth penetration, expensive equipment	High	Moderate
MM	Uses nonlinear optical phenomena to generate signals	Real-time imaging of live tissues, deep tissue imaging	Micrometer level	Biological samples, deep tissue imaging	High-resolution 3D images, less phototoxicity	Requires high laser power, expensive equipment	High	Moderate
Tensile testing	Applies controlled forces to measure mechanical properties	Stress–strain analysis, tissue elasticity	Millimeter to centimeter level	Muscles, tendons, ligaments, skin	Quantitative data on mechanical properties, predictive modeling	Destructive testing	Low	Fast
Compression testing	Compresses tissue samples to simulate physiological stress conditions	Elastic modulus, ultimate strength measurement	Millimeter to centimeter level	Bone, muscle, blood vessels	Simple, quick measurement of mechanical properties	Destructive testing	Low	Fast
Creep and stress relaxation	Measures gradual deformation and stress reduction under constant load	Viscoelastic property analysis	Millimeter to centimeter level	Various biological tissues	Insights into tissue viscoelastic behavior, constitutive model development	Time-consuming, requires specialized equipment	Moderate	Slow
DMA	Applies oscillating stress to measure viscoelastic behavior	Structure–function relationship analysis, mechanical integrity assessment	Millimeter to centimeter level	Tendons, heart valves, various tissues	Detailed viscoelastic parameters, comprehensive mechanical maps, different patterns	Requires specialized equipment, complex data analysis	High	Moderate

---

3.1. EM: unveiling tissue ultrastructure at nanoscale resolution

While histological staining, immunostaining, or tissue clearing relies on optical microscopy, their resolution and imaging detail are inherently limited. To overcome these constraints, EM has emerged as a powerful technique, offering nanometer-scale resolution. EM is broadly categorized into two main types: scanning electron microscopy (SEM) and traditional transmission electron microscopy (TEM).

SEM is widely used for characterizing biomaterials, providing magnifications ranging from 10× to over 300 000×.⁹⁵ SEM operates by focusing a finely tuned electron beam onto the sample surface, generating signals that provide detailed information about the surface topography, composition, and electrical properties.⁹⁶ This technique is particularly valuable for visualizing the 3D architecture of tissues at high resolution (Fig. 4A–C).⁹⁷ The integration of focused ion beam (FIB) technology with SEM has further expanded its capabilities, enabling 3D structural analysis through serial sectioning and imaging.⁹⁸ This combined approach not only provides detailed internal microstructure information but also facilitates chemical composition analysis via energy-dispersive X-ray spectroscopy (EDS), offering insights into phase interactions and buried interfaces within samples.

Fig. 4 EM and confocal microscopy for detailed tissue analysis. Scanning electron microscopy images of subcutaneous segments of an infected epidural catheter. The outer surface of the catheter with (A) low, (B) medium, or (C) high magnification (reproduced from ref. 97 with permission from BMJ Journals, copyright 2020). (D) TEM analysis of time-sequence images of biomineralization stages (reproduced from ref. 99 with permission from ACS Publications, copyright 2020). (E) Schematic illustration of a galvanometer scanning confocal design. (F) Images of an adult female zebrafish's brain (top) and gills (bottom) obtained using an IR-LSM instrument (reproduced from ref. 100 with permission from Nature, copyright 2023).

TEM, on the other hand, generates direct diffraction images based on the electron distribution within the sample, revealing ultrastructural features such as dislocations, grain boundaries, and defects. Liquid TEM provides new insights into fundamental biomineralization processes and essential physiological and pathological processes for a wide range of organisms (Fig. 4D).⁹⁹ For instance, David et al. employed TEM to compare the membrane and ligamentum flavum in patients undergoing lumbar decompression/discectomy, uncovering alterations in histological and ultrastructural composition.¹⁰¹ With magnifications ranging from 50× to 106×, TEM has achieved its exceptional resolution by leveraging the extremely small effective wavelength of electrons.¹⁰² Generally, TEM is indispensable for studying subcellular structures and molecular interactions at nanometer resolution.

3.2. Confocal and multiphoton microscopy: high-resolution 3D imaging of biological tissues

Fluorescence microscopy enables the observation of molecular localization and dynamics within cells and tissues; however, its main limitation is imaging blurriness caused by out-of-focus light, which occurs when attempting to visualize a 3D structure in two dimensions.¹⁰³ To address this challenge, advanced optical microscopy techniques, such as laser scanning confocal microscopy (LSCM) and multiphoton microscopy (MM), have been developed. LSCM significantly reduces out-of-focus light by using a pinhole aperture to exclude defocused signals, thereby enhancing imaging clarity. This technique combines high lateral resolution with non-invasive optical sectioning, enabling the generation of high-quality 3D images of biological samples. MM, on the other hand, leverages nonlinear optical phenomena, where two or more photons simultaneously interact with a sample to produce a fluorescence signal. This technique not only achieves high-resolution 3D images but also enables hyperspectral imaging, providing detailed insights into tissue structure and function.¹⁰⁴ For example, Tomoki et al. utilized multiphoton microscopy and fluorescence labeling to visualize slender sperm cells within fixed seminiferous tubules without the need for physical sectioning, demonstrating its potential for real-time imaging of testicular tissue.¹⁰⁵ Xin et al. employed nanoprobes to precisely delineate tumor margins and provided intraoperative guidance through real-time fluorescence imaging, enabling complete tumor resection.^100,106 Kevin et al. designed a mid-infrared confocal laser scanning microscope (IR-LSM) for full-slide chemical imaging. The system achieves full-slide, speckle-free imaging within 3 minutes per discrete wavelength at 10× magnification (2 μm pixel⁻¹), and supports high-resolution imaging at 20× magnification (1 μm pixel⁻¹). Both provide spatial quality at theoretical limits while maintaining a high signal-to-noise ratio (>100 : 1) (Fig. 4E and F).¹⁰⁷ These advanced techniques have overcome the limitations of traditional fluorescence microscopy, offering researchers powerful tools for high-resolution 3D imaging of biological samples.

3.3. Mechanical testing and visualization: mapping tissue biomechanics with precision

Studying the mechanical properties of biological tissues, combined with advanced visualization techniques, provides critical insights into their functional performance under physiological and pathological conditions. For instance, the elasticity and toughness of heart tissue directly affect its pumping efficiency; thus, changes in mechanical properties can serve as biomarkers for cardiovascular diseases. Similarly, tumor tissue often exhibits distinct mechanical characteristics compared to healthy tissues. Mechanical testing, integrated with advanced visualization methods, such as digital image correlation (DIC),^108,109 high-speed imaging,¹¹⁰ and computational modeling,^111,112 enables researchers to map tissue deformation, strain distribution, and internal structural changes, offering a comprehensive understanding of tissue behavior and function. These approaches not only aid in early disease detection and diagnosis but also provide valuable insights for tissue engineering and regenerative medicine.

3.3.1. Tensile testing. Tensile testing is a commonly used method for studying the mechanical properties of biological tissues under stress.¹¹³ This method employs a tensile testing machine to apply controlled forces and displacements that simulate physiological stress conditions. Tissue samples, such as muscles,¹¹⁴ tendons,¹¹⁵ ligaments, and skin, are harvested from human or animal sources, with their size and shape tailored to specific research objectives. For example, Emily et al. analyzed the tensile stress distribution between thoracolumbar fascia and muscle tissues, providing insights into stress-shielding mechanisms in degenerative musculoskeletal diseases.¹¹⁶ Tensile testing generates quantitative data, such as stress–strain curves, stiffness, and strength, which are crucial for establishing constitutive models to predict tissue behavior under different stress conditions. Advanced visualization techniques, such as DIC, are often integrated with tensile testing to map strain distribution across tissue surfaces in real time. This combination allows researchers to visualize localized deformation and failure mechanisms through strain maps and stress distribution plots, providing a deeper understanding of tissue mechanics and improving the accuracy of predictive models. Additionally, computational tools can generate color-coded mechanical property maps (e.g., strain or stress cloud maps) from experimental data, offering intuitive visual representations of tissue behavior under load.

3.3.2. Compression testing. Compression testing is another universal approach to studying the mechanical properties of tissues, such as bone, muscle, and blood vessels.¹¹⁷ It quickly measures key characteristics such as apparent elastic modulus and ultimate strength¹¹⁸ using specialized equipment to apply controlled compressive forces that mimic physiological conditions. For example, Blake et al. employed compression testing to establish the relationship between the elastic modulus and strain rate in prostate tissue, enhancing our understanding of its response to dynamic loading.¹¹⁹ Similarly, Jacques et al. conducted compression experiments on subcutaneous tissue to model its elastic properties under physiological conditions.¹²⁰ These testing data provide crucial information for applications in tissue engineering and pathophysiology. Visualization techniques, such as high-speed cameras, can capture real-time tissue deformation and internal structural changes to reveal the dynamic response of tissues to rapid compression. Furthermore, similar to tensile testing, stress–strain cloud maps can also be generated from the compression data.

3.3.3. Creep and stress relaxation. Creep and stress relaxation are critical phenomena for understanding the time-dependent mechanical behavior of biological tissues. Creep refers to the gradual increase in deformation under a constant load,¹²¹ typically divided into three stages: primary creep (decreasing strain rate), secondary creep (steady strain rate), and tertiary creep (accelerating strain rate leading to failure), while stress relaxation describes the gradual reduction in stress when a material is held at a constant strain, as the internal structure reorganizes to minimize resistance to deformation.¹²² These behaviors are essential for characterizing the viscoelastic properties of tissues, which are influenced by microstructural interactions between fibers and the ECM. For example, Afshin et al. investigated the kinematics of microstructural reorganization in the aortic valve, revealing how stress relaxation and creep behavior arise from fiber–ECM interactions at the microscale.¹²³ Similarly, William et al. studied the stress relaxation and creep deformation behavior of the human scalp, demonstrating a relaxation time of 10 270 s and a creep deformation threshold force of 105 N. These findings offer valuable data for developing and validating constitutive models of tissue deformation.¹²⁴ To visualize these time-dependent behaviors, we expect that researchers can employ strain–time curves and stress–time curves, which graphically depict creep and stress relaxation processes. Moreover, advanced computational tools can further generate dynamic mechanical graphs to illustrate spatial variations in strain and stress over time, offering intuitive insights into tissue adaptation and failure mechanisms.

3.3.4. Dynamic mechanical analysis (DMA). DMA is also a powerful technique for studying the viscoelastic properties of biological tissues by applying oscillating stress and measuring the resulting strain as a function of oscillatory frequency and temperature.¹²⁵ DMA provides crucial parameters of tissue biomechanics such as the storage modulus (elastic response), loss modulus (viscous response), and damping factor (energy dissipation), improving our understanding of structure–function relationships, mechanical integrity, and tissue response under dynamic loading conditions.¹²⁶ For instance, Jennifer et al. used DMA to compare natural and decellularized porcine superflexor tendons, revealing significant reductions in the dynamic modulus, storage modulus, and loss modulus for the decellularized tissue.¹²⁷ Similarly, Lejla et al. compared the biomechanical properties of porcine mitral leaflets and small intestinal submucosa ECM (SIS-ECM), demonstrating that the bilayer SIS-ECM is better suited as a heart valve repair material.¹²⁸

DMA experiments often generate frequency sweep curves and temperature sweep curves, which visually represent changes in viscoelastic properties under varying conditions. The experimental data are processed to calculate various mechanical moduli, such as axial and circumferential tensile moduli, as well as the compressive modulus. Mechanical property maps can be created from a comprehensive DMA dataset to illustrate spatial variations in storage and loss moduli, offering insights into the distinct mechanical distribution and propagation characteristics of the tissue across different regions. These visualization tools enable researchers to correlate dynamic mechanical behavior with tissue microstructure, enhancing the development of biomaterials and therapeutic strategies.

4. Emerging technologies in biological tissue visualization

The field of biological tissue visualization has undergone transformative advancements with the development of innovative technologies. These emerging techniques provide unprecedented insights into the structure and function of biological tissues, enabling more accurate and comprehensive analysis. From rendering tissues transparent to visualizing molecular structures at near-atomic resolution, these technologies are revolutionizing both clinical and research applications. They are paving the way for new research directions and clinical breakthroughs, offering powerful tools to unravel the complexities of biological tissues (Table 3).

Table 3 Overview of emerging techniques for biological tissue visualization

Imaging technique	Principle	Applications	Resolution (size level)	Scope of use	Advantages	Limitations	Cost	Time efficiency
Tissue clearing	Renders biological samples transparent by reducing light scattering and absorption	Whole-body and whole-organ imaging, long-distance nerve tracing, cancer metastasis identification	Micrometer to millimeter level	Intact biological specimens, entire organs, or organisms	Preserves native structure, avoids mechanical damage, high parallelization	A lengthy process, requires incubation	Moderate	Moderate
Light sheet microscopes and 3D imaging	Use light sheets to illuminate samples and capture 3D images	Pathological diagnosis, 3D morphological biomarker identification	Micrometer level	Tissue samples, clinical practice	Comprehensive 3D analysis, reduce sampling bias	Large datasets, require AI for data management	High	Moderate
Cryo-EM	Uses electron beams to image rapidly frozen samples at near-atomic resolution	Structural biology, protein and virus structure analysis	Near-atomic level	Macromolecular complexes, proteins, viruses	High resolution, preserves native structure	Requires specialized equipment, complex sample preparation	High	Moderate
Single-cell technologies	Analyze individual cells for gene expression, chromatin accessibility, and protein levels	Cellular atlases, disease mechanisms, personalized medicine	Single-cell level	Individual cells, complex biological systems	High resolution, capture cellular heterogeneity	High costs, complex data analysis	High	Moderate
AI-assisted tissue image analysis techniques	Utilize machine learning and deep learning for digital histology	Histological workflows, cancer research, brain connectome reconstruction	Micrometer to nanometer level	Histological samples, complex biological systems	Sustainable, efficient, cost-effective, high accuracy	Require specialized equipment, complex data analysis	High	High

---

4.1. Tissue clearing: from whole-organ transparency to 3D histological analysis

Tissue-clearing techniques render biological samples transparent, enabling deep imaging of large tissue volumes using light microscopy. By eliminating the need for time-consuming tissue sectioning, which can introduce artifacts, tissue clearing allows researchers to extract detailed information from intact biological specimens with a global perspective.^129,130 This technique enhances light penetration by reducing scattering and absorption, making it possible to visualize entire organs or even whole organisms. Biological tissues are composed of heterogeneous substances with varying optical properties, such as connective tissue fibers, organelles, lipid particles, and proteins, each with refractive indices. This heterogeneity causes light scattering and absorption, which are the primary reasons for tissue opacity. Light scattering occurs as incoming light is deflected multiple times while traveling through the tissue, leading to a decay in light intensity. Additionally, light absorption by endogenous substances, such as heme and melanin, further attenuates light transmission, hindering whole-body and whole-organ imaging. Tissue clearing addresses these challenges by physically or chemically manipulating the optical properties of tissues to reduce scattering and absorption, thereby facilitating volumetric imaging. The ideal tissue-clearing technique should achieve excellent transparency while preserving fluorescent proteins, native architecture, and key molecular information. For example, Zhao et al. developed Shanel, an innovative tissue-clearing method specifically designed for rigid human organs (Fig. 5A). Utilizing this novel approach, they successfully rendered adult human brain tissue transparent and performed 3D histological analysis through depth-resolved antibody and dye labeling (Fig. 5B). This technique enabled the visualization of structural details within intact human eyes at cellular resolution (Fig. 5C).¹³⁰

Fig. 5 Advances in emerging bioimaging technologies for tissue structure characterization. (A) Schematic diagram showing the enhanced efficiency and depth of tissue permeabilization achieved by the SHANEL method. (B) A fully transparent image of an intact human brain, demonstrating the effectiveness of tissue clearing. (C) Schematic diagram of tissue clearing results of various structures within the human eye (reproduced from ref. 128 with permission from Cell, copyright 2020). (D) 3D renderings of two-color immunostained mouse mammary glands, captured using a tiling light sheet microscope, showing both combined and separate color channels (reproduced from ref. 8 with permission from Cell Reports, copyright 2020). Scale bars, 1 mm.

Compared to traditional serial sectioning, tissue clearing offers several advantages. Although the process can be lengthier due to the requirement for incubation in various solutions and the diffusion of labels, it requires minimal hands-on time and no specialized skills. Clearing protocols that do not rely on specialized equipment allow high-throughput parallelization, which is not feasible with single vibratome or cryostat systems. Furthermore, tissue clearing avoids mechanical damage caused by sectioning and eliminates the computational burden of aligning, de-warping, and assembling thousands of 2D images into a 3D reconstruction. Analysis of intact, cleared samples has been shown to improve accuracy compared with quantitative stereology¹³¹ and traditional 2D histopathology.¹³² Tissue clearing also enables studies that are impossible with serial sectioning, such as long-distance nerve tracing¹³³ across entire organs or identification of cancer metastasis^95,134 throughout whole organisms. These capabilities make tissue clearing an invaluable tool for advancing our understanding of complex biological systems.

4.2. Light sheet microscopy and 3D imaging: revolutionizing pathological diagnosis

Traditional pathological diagnosis primarily relies on 2D observations of tissue slices,^8,135 which often fail to capture the full complexity and spatial organization of tissues.^136,137 The advent of 3D pathology techniques has revolutionized the field, enabling comprehensive, multidimensional analysis of tissue samples and providing more accurate diagnostic information for clinical practice. For instance, Chen et al. developed a tiling light-sheet microscopy system compatible with all tissue-clearing methods, enabling rapid multicolor 3D imaging at the micrometer scale. This innovative approach utilizes tiled light sheets to achieve enhanced spatial resolution and improved optical sectioning capability (Fig. 5D).⁸

A key challenge for future pathologists lies in effectively reviewing the vast datasets generated by 3D imaging and developing standardized methods to interpret the additional insights provided by 3D pathology.¹³⁸ AI is expected to play a pivotal role in addressing this challenge, assisting pathologists in selecting high-risk 2D sections or serving as a complement to fully automated decision support systems. Moreover, AI can be instrumental in identifying 3D morphological biomarkers that are invisible to traditional 2D methods.¹³⁹ Researchers envision the development of laboratory-developed tests (LDTs) based on 3D pathology, which will be non-destructive to valuable tissue samples – a significant advantage over many conventional molecular LDTs used in patient management.

4.3. Cryo-electron microscopy (cryo-EM): visualizing molecular structures at near-atomic resolution

Cryo-EM is a cutting-edge imaging technique that enables the visualization of biological molecules at near-atomic resolution.¹³⁸ This technique involves rapidly freezing samples below −150 °C and trapping them in vitreous ice to preserve their native structure. The frozen samples are then imaged using an EM, where a beam of electrons passes through the specimen, resulting in a series of 2D projections.¹⁴⁰ Advanced computational algorithms reconstruct a 3D model of the sample from these projections, providing detailed insights into molecular structures.¹⁴¹

Cryo-EM offers several advantages over traditional EM, including higher resolution and the preservation of native molecule structure (e.g., proteins, viruses, and other macromolecular complexes) without the need for staining or fixing, which can introduce artifacts.^141,142 Recent advancements, such as direct electron detectors and improved image processing algorithms, have significantly enhanced the resolution and throughput of cryo-EM, enabling researchers to gain detailed insights into the molecular mechanisms of biological processes.¹⁴³ For example, Dimitry et al. introduced Warp, a software platform that automates all preprocessing steps for cryo-EM data acquisition and enables real-time performance evaluation.¹⁴²

Despite its transformative potential, cryo-EM requires specialized equipment and expertise, and the sample preparation process can be complex. Nevertheless, cryo-EM has revolutionized the field of structural biology and continues to drive significant scientific discoveries.¹⁴⁴

4.4. Single-cell technologies: decoding cellular heterogeneity and function

Single-cell technologies have revolutionized the study of biological tissues by enabling the analysis of individual cells rather than bulk tissue samples. For example, Zhang et al. developed deep learning-driven adaptive optics (DLAO) that uses a deep neural network (DNN) to directly decode wavefront distortions from single-molecule emission patterns. Integrating Kalman-filtered dynamic control with deformable mirrors, the system corrects 28 aberration modes in real time with merely 3–20 mirror adjustments. This framework achieves near-ideal emission pattern recovery, preserving 3D super-resolution through 130 μm brain tissue.¹⁴⁵ These technologies allow researchers to investigate genes, proteins, and other cellular features at the single-cell level, providing unprecedented insights into cellular heterogeneity and function.^146,147 Commonly used methods include single-cell RNA sequencing (scRNA-seq) and single-cell ATAC-seq (scATAC-seq),¹⁴⁸ which profile gene expression and chromatin accessibility, respectively. These methods have been instrumental in constructing cellular atlases, such as the Human Cell Atlas project, which aims to catalog every cell type in the human body.¹⁴⁹

Recent advancements in single-cell technologies include spatial transcriptomics, which combines scRNA-seq with spatial information to map the spatial organization of cells within tissues. This approach allows researchers to study the spatial dynamics of gene expression and cell–cell interactions in their native tissue context.¹⁵⁰ Additionally, the state-of-the-art progress in single-cell omics has been extended to spatial metabolomics platforms. Spatial metabolomics, through its capacity for in situ mapping of metabolite distributions, has emerged as a catalytic driver of innovation in mass spectrometry, metabolomics, and spatial omics. By elucidating the metabolic states and heterogeneity across cellular, tissue, and organ systems, this field now demonstrates transformative potential in advancing biological discovery, clinical diagnostics, and pharmacological development.¹⁵¹ Single-cell multi-omics technologies enable the simultaneous measurement of multiple molecular features, such as gene expression, chromatin accessibility, and protein levels, within individual cells.^152,153 While single-cell technologies offer high resolution and the ability to capture cellular heterogeneity,^154,155 they also present challenges, including the need for specialized equipment, high costs, and complex data analysis. Despite these limitations, single-cell technologies have transformed biological research, providing critical insights into complex systems, disease mechanisms, and personalized medicine approaches.

4.5. AI-assisted tissue image analysis: transforming histology with deep learning

As the foundational method for pathological analysis, traditional histological workflows are labor-intensive, requiring significant amounts of chemical reagents and time. Currently, machine learning and deep learning are progressively driving the digital transformation of histological visualization, offering more sustainable, efficient, and cost-effective alternatives to conventional methods.^156,157 Gabriele et al. proposed a deep learning system based on multi-instance learning (MIL) for pan-cancer whole-slide image (WSI) analysis. This framework leverages diagnostic reports as training labels, circumventing the need for costly and time-consuming pixel-level manual annotations. In clinical deployment, the system enables pathologists to exclude 65–75% of slides while maintaining 100% sensitivity. Trained at an unprecedented scale, this model achieves diagnostic accuracy comparable to that of clinical pathologists—marking the first demonstration of AI reaching expert-level performance in pathology. This breakthrough represents a pivotal step in translating AI-driven pathology from research laboratories into real-world clinical practice.¹⁵⁸ These AI-driven innovations have revolutionized data-driven predictive modeling across various fields,¹⁵⁹ including medicine, and histology is no exception. Recent advancements demonstrate that deep learning-based virtual staining techniques can achieve chemical staining-quality results for unstained tissue, enabling basic morphological examination without the need for traditional staining protocols.^160,161 Furthermore, tissue transparency techniques have evolved to allow the removal of hard tissue components using specific solvents,¹⁶² facilitating the study of bones and teeth, and even enabling whole-body transparency.¹³⁰ These developments provide researchers with a diverse array of visualization methods to process and display biological tissue information according to their specific needs.

With advancements in computer vision, the field of microscopy imaging has also benefited significantly. For example, Julia et al. introduced Comprehensive Analysis of Tissues across Scales (CATS), an innovative imaging technology capable of generating high-density architectural maps of brain tissue across multiple spatial scales, ranging from millimeter-level regional organization to nanometer-scale synaptic structures. This advanced platform is applicable to various chemically fixed brain specimens, including both rodent and human tissue samples (Fig. 6A and B).¹⁵⁷ Larger deep learning models with increased parameters often yield superior performance, and the development of large multimodal models holds promise for groundbreaking advancements in biological imaging, potentially surpassing human recognition capabilities.¹⁶³ A notable example is the application of AI in reconstructing brain connectomes. Sven et al. developed Segmentation-guided Contrastive Learning of Representations (SegCLR), a self-supervised machine learning technique that produces cell representations directly from 3D imagery and segmentations. When applied to human and mouse cortex volumes, SegCLR enables an accurate classification of cellular subcompartments and achieves performance equivalent to supervised approaches while requiring 400-fold fewer labeled examples (Fig. 6C–E).¹⁶⁴ Accurate segmentation of neurons and their processes within large-scale electron microscopy datasets is essential for brain reconstruction, whether the data originate from fruit flies, mice, or humans.^165,166 In another breakthrough, Despina et al. developed a rapid, label-free method to analyze biopsy samples by sequentially assessing the physical phenotypes of isolated abnormal suspended cells within 30 min. This approach enables the quantification of colonic inflammation levels and accurately distinguishes between healthy and tumor tissues in both murine and human colon biopsy specimens, offering the potential for intraoperative detection of pathological changes in solid biopsies (Fig. 6F and G).¹⁶⁶ AI is also transforming our understanding of complex biological systems, such as the human immune system.¹⁶⁷ In cancer research, Elham Azizi and his colleagues discussed the role of AI in addressing critical challenges, such as integrating heterogeneous data, quantifying and modeling cells, and identifying causal regulatory networks involved in tumor development, metastasis, and dysregulation.¹⁶⁸

Fig. 6 AI-assisted technologies enhancing precision in clinical diagnosis and treatment. (A) Comprehensive Analysis of Tissues across Scales (CATS) technology that enables dense mapping of brain tissue structures, spanning from millimeter-scale regions to nanoscale synaptic details, in various chemically fixed brain preparations. (B) CATS-generated super-resolved tissue images of synaptic structures in the mouse brain (reproduced from ref. 153 with permission from Nature, copyright 2024). (C) SegCLR that produces cell representations from 3D imagery and segmentations. Visualized embeddings of representative human (D) and mouse (E) cells (reproduced from ref. 159 with permission from Nature, copyright 2023). (F) Analysis of biopsy samples through sequential assessment of physical phenotypes in singularized suspended cells dissociated from tissues. (G) Scatter plots comparing cell size versus deformation in cells isolated from transfer colitis tissue samples (TC) and healthy murine colon tissue (Control), along with the corresponding histograms of cell size and deformation (reproduced from ref. 161 with permission from Nature, copyright 2023).

AI has demonstrated transformative potential in biological visualization, with its integration into conventional techniques representing a frontier in data science. However, biomedical applications face critical challenges rooted in the inherent complexity of biological data. The high-dimensional, spatiotemporal, and multiscale nature of biological systems—spanning from molecular interactions to organ-level dynamics—strains the data representation capacity of AI. Multimodal integration of heterogeneous data types (e.g., genomic sequences, medical imaging) encounters semantic disparities that hinder feature alignment, while inconsistent annotation standards and domain knowledge dependence complicate training data preparation. Models optimized for specific data distributions frequently falter when confronting device variability, population heterogeneity, or novel pathologies, particularly in dynamic processes like embryonic development or rare disease studies. These limitations are compounded by interdisciplinary barriers: life sciences, computer science, and clinical medicine exhibit fundamental divergences in conceptual frameworks, technical infrastructure, and variable integration paradigms. Such systemic challenges not only impede the clinical translation of AI-enhanced visualization tools but also question their scientific validity in decoding complex living systems. Addressing these issues demands coordinated innovation in domain-adaptive methodologies and cross-disciplinary collaboration to achieve paradigm shifts in biological visualization technology.

5. Applications of visualization technologies in biomedical research

In clinical practice, visualization technologies are increasingly vital for disease diagnosis and treatment. One notable advancement is tissue clearing, which holds significant promise for investigating neuronal pathologies. For instance, the SWITCH (system-wide control of interaction time and kinetics of chemicals) clearing and immunolabelling protocol has been recently applied in Alzheimer's research. This method allows brain-wide labeling and imaging of amyloid-beta (Aβ) peptides, which play a critical role in the onset and progression of Alzheimer's dementia (Fig. 7A).¹⁶⁹ SWITCH allowed microscopic mapping of Aβ plaques and identified several novel subcortical hubs, where Aβ accumulates (Fig. 7B), offering new insights into the pathophysiology of Alzheimer's and potential therapeutic strategies. Tissue clearing also plays a significant role in cancer research, particularly in visualizing micrometastases. Whole animal clearing techniques, such as nanobody (VHH)-boosted 3D imaging of solvent-cleared organs (vDISCO), have enabled the visualization of micrometastases throughout an entire rodent at single-cell resolution (Fig. 7C–I).¹⁷⁰ These methods provide a detailed view of the biodistribution of cancer therapeutics, including the ability to track fluorescently labeled therapeutic antibodies across metastatic sites, enhancing the evaluation of treatment efficacy. Beyond cancer and neurobiology, tissue clearing is invaluable for studying vascular dynamics, such as vessel repair and remodeling during development or after injury. It is particularly useful for investigating blood vessel rupture in the brain (e.g., aneurysm and stroke) and myocardial infarctions.¹³⁴

Fig. 7 Emerging visualization technologies for clinical diagnosis. (A) Spatiotemporal mapping of Alzheimer's disease Aβ deposition using 3D imaging techniques. (B) Representative images of amyloid labeling in the brain from a 12-month-old mouse (reproduced from ref. 164 with permission from Nature, copyright 2019). Scale bars, 1000 μm. (C)–(G) 3D visualization of a transparent mouse body imaged by light-sheet microscopy, demonstrating single-cell metastases identified in the brain through full-body scans. (H) Images acquired with a 1.1× objective, providing an overview of metastatic distribution. (I) High-resolution images acquired with a 12× objective, revealing detailed cellular structures (reproduced from ref. 161 with permission from Cell, copyright 2019).

Traditional 2D pathology is often limited by sampling bias, which can obscure the full range of tissue characteristics and lead to diagnostic inaccuracies. To address this, 3D imaging technologies, such as open-top light-sheet microscopy (OTLS) and micro-CT, provide more detailed tissue information, significantly reducing these biases. For instance, the TriPath platform, developed by Andrew and colleagues, utilized deep learning algorithms to process and analyze 3D tissue volume data. This platform has shown promise in predicting clinical outcomes and offering new insights into patient prognosis,¹⁷¹ highlighting the potential of 3D pathology in clinical applications and guiding the future of pathological diagnosis.

Another significant advancement is the use of digital extended reality technologies, which are widely used in biomedical science due to their accessibility and affordability. In biomedical engineering, virtual reality (VR) and AR have dramatically enhanced the visualization and interaction with microscopic images, molecular data, and anatomical datasets.¹⁷² For example, expansion microscopy, a nanoscale imaging technique, has been combined with VR to enlarge and analyze cell structures that are too small to visualize using conventional light microscopy. Expansion microscopy increased the tissue sample volume by 100-fold, enabling detailed visualization of tissues, molecules, and cellular interactions. The 2D expansion microscopy images were integrated with VR to create interactive 3D models with a 360° view, providing an immersive analytical experience.

In critical care medicine, AI-powered imaging tools demonstrate significant clinical relevance for time-sensitive conditions. The FDA-cleared Viz LVO system utilizes deep learning to analyze computed tomography angiography (CTA) scans, detecting large vessel occlusions (LVOs) in cerebral arteries with 98.7% accuracy. By automating vascular reconstruction and enabling real-time alerts, it reduces door-to-puncture time by 50% (45 vs. 90 minutes) and improves triage efficiency through multi-institutional coordination. Studies report a 32% reduction in missed LVOs at non-specialized centers. Caption AI's convolutional neural network guides ultrasound probe positioning via an AR interface, ensuring ASE-compliant views (e.g., four-chamber). Its quality assessment module achieves diagnostic-grade images in 91% of cases (vs. 47% with conventional methods) by quantifying left ventricular edge sharpness (index > 0.85). Integrated with portable devices, it reduces cardiac screening costs by 64% in primary care, highlighting the role of AI in democratizing specialized diagnostics.

Despite the remarkable achievements of AI-based methods in biology, several challenges remain. Biological data are often noisy, biased, and highly heterogeneous in quality and quantity. In many cases, determining ground truth is difficult, as even manual annotations are not infallible, limiting the accuracy and generalizability of AI models. Additionally, complex interdependencies between biological datasets may lead to data leakage.¹⁷³ Beyond classification and prediction, biologists aim to use AI to extract biological knowledge and guide the design of new experiments and translational strategies. However, the black-box nature of many machine-learning methods poses a significant obstacle, making interpretable machine learning an attractive alternative.¹⁷⁴

Therefore, visualization technologies are transforming biomedical research by providing deeper insights into complex biological systems. From tissue clearing and 3D imaging to AI-driven analysis and AR/VR technologies, these tools hold immense potential to revolutionize clinical applications and advance our understanding of disease mechanisms.

6. Challenges and future directions in biological tissue visualization

6.1. Data complexity and integration

Biological visualization technologies play a crucial role in presenting vast amounts of complex biological data in an intuitive and accessible manner, enabling researchers to uncover patterns and relationships within biological systems. These tools facilitate the identification of tissue structures and functions, as well as the interactions between different tissues, thereby advancing research in biology and medicine. However, as the volume and diversity of data continue to grow, traditional data presentation methods are struggling to meet researchers’ demands for efficiency and clarity. This highlights the urgent need for innovative tools and methods to process and interpret these complex biological datasets. For example, Jaclan et al. proposed the TraNCE framework, which can automate the design of distributed analyses for complex biomedical data types, enabling scalable processing of heterogeneous data.¹⁷⁵ Similarly, Thierry et al. enhanced the open-source software ARX with advanced search algorithms, improving its ability to handle high-dimensional data in terms of processing performance and availability, thereby facilitating data sharing.¹⁷⁶ Mari et al. generated single-nucleus RNA-sequencing data for approximately 400 000 cells to trace cerebellum development from early neurogenesis to adulthood in humans, mice, and marsupial opossums. Their study unveiled shared and lineage-specific gene-expression programs governing cerebellar cellular development,¹⁷⁷ highlighting the power of single-cell technologies in uncovering evolutionary and developmental mechanisms. Similarly, Jiang et al. explored the establishment and maintenance of mouse kidney structures using single-cell resolution spatially resolved transcriptome maps, identifying 40 intercellular communication events related to migration during kidney organogenesis.¹⁷⁸ Advances in genomics sequencing, image processing, and medical data management have further supported the collection and integration of clinical data, while the development of algorithms and program frameworks has provided an effective approach for managing large-scale biological datasets.

6.2. Technological advancements and integration

The integration of deep learning and AI has revolutionized data-driven predictive modeling across various fields. In the future, biological visualization technologies are expected to enable comprehensive, multidimensional information visualization of human tissues, offering deeper insights into the complexity of biological systems. This advancement will provide robust support for biomedical research and promote the development of personalized and precision medicine. For example, Oshma et al. analyzed vaginal metagenomic data and trained deep neural networks (DNNS) to predict full-term birth (TB) and preterm birth (PTB) with 84.10% accuracy. This AI-based strategy holds significant potential for clinicians in assessing preterm birth risk and addressing related health issues.¹⁷⁹ Ehteshami et al. designed a convolutional neural network (CNN)-based model that combines stroma features to distinguish invasive breast cancer from benign lesions.¹⁸⁰ Ding et al. proposed a method for predicting cross-level molecular profiles, including gene mutations, copy number variations, and functional protein expression from whole-slide images, focusing on the spatial analysis of cancer tiles.¹⁸¹ The widespread adoption of AI in clinical settings will depend on the availability of phenotypically rich datasets for model development and the clinical validation of AI-generated insights. The widespread adoption of AI in clinical settings will hinge on two pivotal elements: the establishment of comprehensive phenotypic datasets to facilitate model development, and the rigorous clinical validation of AI-generated insights. By integrating advanced visualization technologies, these approaches will significantly enhance the efficiency and accuracy of data interpretation, thereby unlocking new possibilities for medical research and clinical practice.

While recent advancements have significantly enhanced biological visualization capabilities, critical challenges persist in technical integration paradigms. A formidable challenge lies in multimodal data fusion processes, where algorithmic bottlenecks emerge from the need to synchronize heterogeneous data types (e.g., molecular dynamics, cellular imaging, and omics datasets) with varying spatiotemporal resolutions. These challenges are exemplified by substantial temporal resolution mismatches between techniques such as fluorescence microscopy, which operates at millisecond timescales to capture rapid molecular interactions, and dynamic mechanical analysis (DMA), which measures material properties over seconds to minutes, creating temporal alignment gaps that obscure dynamic biological processes. Spatial resolution disparities further complicate integration, as millimeter-scale clinical imaging modalities like CT scans cannot be directly correlated with micrometer-scale cellular features visualized through immunofluorescence histology, requiring computationally intensive spatial registration algorithms to bridge these scales. Additionally, structural incompatibilities arise between discrete, text-based patient metadata (e.g., clinical records) and high-dimensional imaging data (e.g., MRI volumetric datasets), demanding specialized frameworks to reconcile tabular, textual, and multidimensional numerical formats. The compounded effects of these temporal, spatial, and structural heterogeneities amplify computational complexity, often necessitating trade-offs between analytical precision and interpretative coherence in multimodal integration pipelines.

Although current methodologies face persistent challenges in reconciling multidimensional discrepancies, emerging solutions demonstrate promising pathways.¹⁸² For example, Justin et al. pioneered an NLP-driven framework that harmonizes genomic profiles with radiological imaging through semantic mapping, achieving a 15% improvement in metastatic cancer outcome prediction.¹⁸³ Similarly, Carrillo-Perez et al. developed a machine learning-based late fusion architecture integrating RNA-seq, WSI, and epigenetic data, which surpassed unimodal performance by 12.8% in NSCLC subtyping accuracy while maintaining computational efficiency through optimized gradient-based weighting.¹⁸⁴ These advances highlight the feasibility of overcoming dimensional complexity through hybrid AI strategies and standardized intermediate representations, though systematic validation across diverse clinical cohorts remains imperative.

6.3. Future prospects

The future directions of biological visualization should prioritize interdisciplinary collaboration frameworks integrating computational biology, materials science, and human–computer interaction design. Particularly promising avenues include the development of adaptive visualization algorithms capable of handling biological complexity gradients and the creation of hybrid systems that effectively bridge data visualization tools with programmable biomaterials. Such innovations could enable synergistic solutions for challenges like volumetric rendering of soft matter interfaces and dynamic visualization of metabolic networks.

The innovative development and integration of advanced biological visualization technologies have significantly advanced our understanding of tissue structure and function. These studies underscore the immense potential of advanced biological visualization technologies in integrating multidimensional data and elucidating dynamic tissue alterations. With the progressive advancement of information visualization techniques, biomedical research will be empowered to more efficiently decode the multi-layered structures and functions of complex biological systems, thereby providing robust support for both disease mechanism investigations and precision medicine initiatives.

7. Conclusions

In summary, the advancement of biological visualization technologies has not only provided new perspectives and tools for scientific research but also empowered researchers to decode the complexities of biological systems – revealing nature's design principles for next-generation biomaterials. Looking ahead, as these technologies become increasingly intelligent and multidimensional, they will drive transformative breakthroughs in biology and medicine by bridging tissue-scale decoding with functional biomaterial innovation, ultimately enabling advanced regenerative therapies.

Author contributions

Haitao Cui and Haijun Cui: conceptualization; Zhiyuan Zhao: writing the original draft; Haitao Cui, Haijun Cui, and Zhiyuan Zhao: review & editing.

Data availability

No primary research results, software, or code have been included, and no new data were generated or analyzed as part of this review.

Conflicts of interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgements

This study was supported by the National Key R&D Program of China (2023YFB4605800), the Fundamental Research Funds for the Central Universities of China (2023CDJKYJH056), the National Natural Science Foundation of China (32471470 and 32201098), the Natural Science Foundation of Chongqing (CSTB2023NSCQ-MSX0498, CSTB2023NSCQ-MSX0163, and cstc2021jcyj-cxttX0002), and the Start-up Funding from Chongqing University.

References

1. O. Rozenblatt-Rosen, J. W. Shin, J. E. Rood, A. Hupalowska, K. Ardlie, M. Clatworthy, P. Carninci, W. Enard, W. Greenleaf, H. Heyn, E. Lein, J. Z. Levin, S. Linnarsson, E. Lundberg, K. Meyer, N. Navin, G. Nolan, S. Teichmann, T. Voet, X. Zhuang, A. Regev, H. Heyn and Human Cell Atlas Standards and Technology Working Group, Nat. Biotechnol., 2021, 39, 149–153.
2. B. Torrado, B. Pannunzio, L. Malacrida and M. A. Digman, Nat. Rev. Methods Primers, 2024, 4, 80.
3. C. A. J. van Winkel, F. R. Pierik, A. H. Brouwers, D. J. A. de Groot, E. G. E. de Vries and M. N. Lub-de Hooge, Nat. Rev. Clin. Oncol., 2024, 21, 852–866.
4. T. Zhang, S. Guo, F. Li, X. Lan, Y. Jia, J. Zhang, Y. Huang and X.-J. Liang, Adv. Drug Delivery Rev., 2022, 188, 114446.
5. E. H. K. Stelzer, F. Strobl, B.-J. Chang, F. Preusser, S. Preibisch, K. McDole and R. Fiolka, Nat. Rev. Methods Primers, 2021, 1, 73.
6. Y. Chen, S. Chauhan, C. Gong, H. Dayton, C. Xu, E. D. De La Cruz, Y.-Y. W. Tsai, M. S. Datta, G. B. Rosoklija, A. J. Dwork, J. J. Mann, M. Boldrini, K. W. Leong, L. E. P. Dietrich and R. Tomer, Nat. Biomed. Eng., 2024, 8, 1109–1123.
7. A. K. Glaser, K. W. Bishop, L. A. Barner, E. A. Susaki, S. I. Kubota, G. Gao, R. B. Serafin, P. Balaram, E. Turschak, P. R. Nicovich, H. Lai, L. A. G. Lucas, Y. Yi, E. K. Nichols, H. Huang, N. P. Reder, J. J. Wilson, R. Sivakumar, E. Shamskhou, C. R. Stoltzfus, X. Wei, A. K. Hempton, M. Pende, P. Murawala, H.-U. Dodt, T. Imaizumi, J. Shendure, B. J. Beliveau, M. Y. Gerner, L. Xin, H. Zhao, L. D. True, R. C. Reid, J. Chandrashekar, H. R. Ueda, K. Svoboda and J. T. C. Liu, Nat. Methods, 2022, 19, 613–619.
8. Y. Chen, X. Li, D. Zhang, C. Wang, R. Feng, X. Li, Y. Wen, H. Xu, X. S. Zhang and X. J. C. R. Yang, Cell Rep., 2020, 33, 108349.
9. P. J. Withers, C. Bouman, S. Carmignato, V. Cnudde, D. Grimaldi, C. K. Hagen, E. Maire, M. Manley, A. Du Plessis and S. R. Stock, Nat. Rev. Methods Primers, 2021, 1, 18.
10. E. Lev and C. M. Boyce, iScience, 2020, 23, 101534.
11. C. Rabut, S. Yoo, R. C. Hurt, Z. Jin, H. Li, H. Guo, B. Ling and M. G. Shapiro, Neuron, 2020, 108, 93–110.
12. H. A. Alturkistani, F. M. Tashkandi and Z. M. Mohammedsaleh, Global J. Health Sci., 2015, 8, 72–79.
13. J. Zhao, X. Yu, X. Shentu and D. Li, Cell Tissue Res., 2024, 396, 1–18.
14. Y. Mao and S. A. Wickström, Nat. Rev. Mol. Cell Biol., 2024, 25, 654–670.
15. C. F. Guimarães, L. Gasperini, A. P. Marques and R. L. Reis, Nat. Rev. Mater., 2020, 5, 351–370.
16. H. Shroff, I. Testa, F. Jug and S. Manley, Nat. Rev. Mol. Cell Biol., 2024, 25, 443–463.
17. M. W. Mullowney, K. R. Duncan, S. S. Elsayed, N. Garg, J. J. J. van der Hooft, N. I. Martin, D. Meijer, B. R. Terlouw, F. Biermann, K. Blin, J. Durairaj, M. Gorostiola González, E. J. N. Helfrich, F. Huber, S. Leopold-Messer, K. Rajan, T. de Rond, J. A. van Santen, M. Sorokina, M. J. Balunas, M. A. Beniddir, D. A. van Bergeijk, L. M. Carroll, C. M. Clark, D.-A. Clevert, C. A. Dejong, C. Du, S. Ferrinho, F. Grisoni, A. Hofstetter, W. Jespers, O. V. Kalinina, S. A. Kautsar, H. Kim, T. F. Leao, J. Masschelein, E. R. Rees, R. Reher, D. Reker, P. Schwaller, M. Segler, M. A. Skinnider, A. S. Walker, E. L. Willighagen, B. Zdrazil, N. Ziemert, R. J. M. Goss, P. Guyomard, A. Volkamer, W. H. Gerwick, H. U. Kim, R. Müller, G. P. van Wezel, G. J. P. van Westen, A. K. H. Hirsch, R. G. Linington, S. L. Robinson and M. H. Medema, Nat. Rev. Drug Discovery, 2023, 22, 895–916.
18. G. Novakovsky, N. Dexter, M. W. Libbrecht, W. W. Wasserman and S. Mostafavi, Nat. Rev. Genet., 2023, 24, 125–137.
19. A. Bodén, F. Pennacchietti, G. Coceano, M. Damenti, M. Ratz and I. Testa, Nat. Biotechnol., 2021, 39, 609–618.
20. Z. Lu, Y. Liu, M. Jin, X. Luo, H. Yue, Z. Wang, S. Zuo, Y. Zeng, J. Fan, Y. Pang, J. Wu, J. Yang and Q. Dai, Nat. Methods, 2023, 20, 735–746.
21. S. Daly, J. Ferreira Fernandes, E. Bruggeman, A. Handa, R. Peters, S. Benaissa, B. Zhang, J. S. Beckwith, E. W. Sanders, R. R. Sims, D. Klenerman, S. J. Davis, K. O’Holleran and S. F. Lee, Nat. Commun., 2024, 15, 1940.
22. M. Komatsu, A. Sakai, A. Dozen, K. Shozu, S. Yasutomi, H. Machino, K. Asada, S. Kaneko and R. Hamamoto, Biomedicines, 2021, 9, 720.
23. C. H. McCollough and P. S. Rajiah, Radiology, 2023, 309, e230803.
24. H.-J. Smith, Radiology, 2021, 62, 1481–1498.
25. T. Claesson, Stockholm: Universitetsservice US AB, 2001.
26. J. Xu and B. M. Tsui, IEEE Trans. Med., 2013, 33, 61–73.
27. T. M. Buzug, in Springer Handbook of Medical Technology, ed. R. Kramme, K.-P. Hoffmann and R. S. Pozos, Springer Berlin Heidelberg, Berlin, Heidelberg, 2011, pp. 311–342 DOI:10.1007/978-3-540-74658-4_16.
28. A. M. Cormack, J. Appl. Phys., 2004, 35, 2908–2913.
29. X. Pan, J. Siewerdsen, P. J. La Riviere and W. A. Kalender, Med. Phys., 2008, 35, 3728–3739.
30. C. H. McCollough and F. E. Zink, Med. Phys., 1999, 26, 2223–2230.
31. K. Nikolaou, T. Flohr, A. Knez, C. Rist, B. Wintersperger, T. Johnson, M. F. Reiser and C. R. Becker, Int. J. Cardiovasc. Imaging, 2004, 20, 535–540.
32. T. G. Flohr, C. H. McCollough, H. Bruder, M. Petersilka, K. Gruber, C. Süβ, M. Grasruck, K. Stierstorfer, B. Krauss, R. Raupach, A. N. Primak, A. Küttner, S. Achenbach, C. Becker, A. Kopp and B. M. Ohnesorge, Eur. Radiol., 2006, 16, 256–268.
33. S. Leng, Z. Yu, A. Halaweish, S. Kappler, K. Hahn, A. Henning, Z. Li, J. Lane, D. L. Levin and S. Jorgensen, J. Med. Imaging, 2016, 3, 043504.
34. J. C. Benson, K. Rajendran, J. I. Lane, F. E. Diehn, N. M. Weber, J. E. Thorne, N. B. Larson, J. G. Fletcher, C. H. McCollough and S. Leng, Am. J. Neuroradiol., 2022, 43, 579–584.
35. W. Zhou, J. I. Lane, M. L. Carlson, M. R. Bruesewitz, R. J. Witte, K. K. Koeller, L. J. Eckel, R. E. Carter, C. H. McCollough and S. Leng, Am. J. Neuroradiol., 2018, 39, 1733–1738.
36. K. Rajendran, B. A. Voss, W. Zhou, S. Tao, D. R. DeLone, J. I. Lane, J. M. Weaver, M. L. Carlson, J. G. Fletcher and C. H. McCollough, Invest. Radiol., 2020, 55, 91.
37. B. Klintström, L. Henriksson, R. Moreno, A. Malusek, Ö. Smedby, M. Woisetschläger and E. Klintström, Eur. Radiol. Exp., 2022, 6, 31.
38. K. Rajendran, F. Baffour, G. Powell, K. Glazebrook, J. Thorne, N. Larson, S. Leng, C. McCollough and J. Fletcher, Skeletal Radiol., 2023, 52, 23–29.
39. F. I. Baffour, K. Rajendran, K. N. Glazebrook, J. E. Thorne, N. B. Larson, S. Leng, C. H. McCollough and J. G. Fletcher, Eur. Radiol., 2022, 32, 7079–7086.
40. S. Si-Mohamed, S. Boccalini, P.-A. Rodesch, R. Dessouky, E. Lahoud, T. Broussaud, M. Sigovan, D. Gamondes, P. Coulon, Y. Yagil, L. Boussel and P. Douek, Diagn. Interventional Imaging, 2021, 102, 305–312.
41. S. A. Si-Mohamed, S. Boccalini, H. Lacombe, A. Diaw, M. Varasteh, P.-A. Rodesch, R. Dessouky, M. Villien, V. Tatard-Leitman, T. Bochaton, P. Coulon, Y. Yagil, E. Lahoud, K. Erhard, B. Riche, E. Bonnefoy, G. Rioufol, G. Finet, C. Bergerot, L. Boussel, J. Greffier and P. C. Douek, Radiology, 2022, 303, 303–313.
42. R. Symons, A. Pourmorteza, V. Sandfort, M. A. Ahlman, T. Cropper, M. Mallek, S. Kappler, S. Ulzheimer, M. Mahesh, E. C. Jones, A. A. Malayeri, L. R. Folio and D. A. Bluemke, Radiology, 2017, 285, 980–989.
43. S. T. Francis, N. M. Selby and M. W. Taal, Am. J. Kidney Dis., 2023, 82, 491–504.
44. C. Wang, X. Chen, L. Wang, M. Makihata, H.-C. Liu, T. Zhou and X. J. S. Zhao, Science, 2022, 377, 517–523.
45. K. Revett, in Innovations in Intelligent Image Analysis, ed. H. Kwaśnicka and L. C. Jain, Springer Berlin Heidelberg, Berlin, Heidelberg, 2011, pp. 127–161 DOI:10.1007/978-3-642-17934-1_7.
46. E. F. Cox, C. E. Buchanan, C. R. Bradley, B. Prestwich, H. Mahmoud, M. Taal, N. M. Selby and S. T. Francis, Front. Physiol., 2017, 8, 696.
47. X. Wang and J. Pu, Small Methods, 2024, 2301170.
48. Y. Jiang, A. V. Edwards and G. M. Newstead, Radiology, 2021, 298, 38–46.
49. C. G. Radhakrishna and P. Ciuciu, Bioengineering, 2023, 10, 158.
50. J. Cho, B. Gagoski, T. H. Kim, Q. Tian, R. Frost, I. Chatnuntawech and B. J. B. Bilgic, Bioengineering, 2022, 9, 736.
51. M. B. Hossain, K.-C. Kwon, S. M. Imtiaz, O.-S. Nam, S.-H. Jeon and N. J. B. Kim, Bioengineering, 2022, 10, 22.
52. E. Shimron and O. Perlman, Bioengineering, 2023, 10, 492.
53. K. Lei, A. B. Syed, X. Zhu, J. M. Pauly and S. V. Vasanawala, Bioengineering, 2023, 10, 92.
54. P. N. Wells, Phys. Med. Biol., 2006, 51, R83.
55. J. M. Sanches, A. F. Laine and J. S. Suri, Ultrasound imaging, Springer, 2012.
56. C. Kollmann, Ultrasound Med. Biol., 2015, 41, 622.
57. A. P. Sarvazyan, O. V. Rudenko and W. L. Nyborg, Ultrasound Med. Biol., 2010, 36, 1379–1394.
58. J. J. Dahl, Ultrasound Med. Biol., 2015, 41, 3259–3260.
59. C. Kasai and K. Namekawa, Jpn. J. Appl. Phys., 1987, 26, 9.
60. H. Hu, H. Huang, M. Li, X. Gao, L. Yin, R. Qi, R. S. Wu, X. Chen, Y. Ma and K. J. N. Shi, Nature, 2023, 613, 667–675.
61. G. B. J. N. R. C. Lim, Nat. Rev. Cardiol., 2023, 20, 212.
62. L. Zhang, C. Marcus, D. Lin, D. Mejorado, S. J. Schoen, T. T. Pierce, V. Kumar, S. V. Fernandez, D. Hunt, Q. Li, I. I. Shuvo, D. Sadat, W. Du, H. Edenbaum, L. Jin, W. Liu, Y. C. Eldar, F. Li, A. P. Chandrakasan, A. E. Samir and C. Dagdeviren, Nat. Electron., 2024, 7, 77–90.
63. M. Lin, Z. Zhang, X. Gao, Y. Bian, R. S. Wu, G. Park, Z. Lou, Z. Zhang, X. Xu and X. J. N. B. Chen, Nat. Biotechnol., 2024, 42, 448–457.
64. B. Li, K. Xu, D. Feng, H. Mi, H. Wang and J. Zhu, Denoising Convolutional Autoencoder Based B-mode Ultrasound Tongue Image Feature Extraction, ICASSP 2019 – 2019 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), Brighton, UK, 2019, pp. 7130–7134.
65. P. Kokil and S. Sudharson, Comput. Methods Programs Biomed., 2020, 194, 105477.
66. D. Perdios, M. Vonlanthen, A. Besson, F. Martinez, M. Arditi and J.-P. Thiran, Deep Convolutional Neural Network for Ultrasound Image Enhancement, 2018 IEEE International Ultrasonics Symposium (IUS), Kobe, Japan, 2018, pp. 1–4.
67. R. J. Van Sloun, R. Cohen and Y. C. Eldar, Proc. IEEE, 2019, 108, 11–29.
68. M. N. Gurcan, L. E. Boucheron, A. Can, A. Madabhushi, N. M. Rajpoot and B. Yener, IEEE Rev. Biomed. Eng., 2009, 2, 147–171.
69. M. Titford, J. Histotechnol., 2009, 32, 9–19.
70. T. Rando, J. Neuropathol. Exp. Neurol., 2008, 67, 633.
71. M. Titford, Biotech. Histochem., 2005, 80, 73–78.
72. N. C. Foot, Stain Technol., 1933, 8, 101–110.
73. C. C. Pollitt, Equine Vet. J., 1996, 28, 38–46.
74. Y. K. Tao, D. Shen, Y. Sheikine, O. O. Ahsen, H. H. Wang, D. B. Schmolze, N. B. Johnson, J. S. Brooker, A. E. Cable, J. L. Connolly and J. G. Fujimoto, Proc. Natl. Acad. Sci. U. S. A., 2014, 111, 15304–15309.
75. D. A. Orringer, B. Pandian, Y. S. Niknafs, T. C. Hollon, J. Boyle, S. Lewis, M. Garrard, S. L. Hervey-Jumper, H. J. L. Garton, C. O. Maher, J. A. Heth, O. Sagher, D. A. Wilkinson, M. Snuderl, S. Venneti, S. H. Ramkissoon, K. A. McFadden, A. Fisher-Hubbard, A. P. Lieberman, T. D. Johnson, X. S. Xie, J. K. Trautman, C. W. Freudiger and S. Camelo-Piragua, Nat. Biomed. Eng., 2017, 1, 0027.
76. H. Tu, Y. Liu, D. Turchinovich, M. Marjanovic, J. K. Lyngsø, J. Lægsgaard, E. J. Chaney, Y. Zhao, S. You, W. L. Wilson, B. Xu, M. Dantus and S. A. Boppart, Nat. Photonics, 2016, 10, 534–540.
77. S. González and Y. Gilaberte-Calzada, Int. J. Cosmet. Sci., 2008, 30, 1–17.
78. S. Al-Janabi, A. Huisman and P. J. Van Diest, Histopathology, 2012, 61, 1–9.
79. A. Madabhushi and G. Lee, Med. Image Anal., 2016, 33, 170–175.
80. H. Wang, Y. Rivenson, Y. Jin, Z. Wei, R. Gao, H. Günaydın, L. A. Bentolila, C. Kural and A. Ozcan, Nat. Methods, 2019, 16, 103–110.
81. Y. Wu, Y. Luo, G. Chaudhari, Y. Rivenson, A. Calis, K. de Haan and A. Ozcan, Light: Sci. Appl., 2019, 8, 25.
82. Y. Jo, H. Cho, W. S. Park, G. Kim, D. Ryu, Y. S. Kim, M. Lee, S. Park, M. J. Lee, H. Joo, H. Jo, S. Lee, S. Lee, H.-S. Min, W. D. Heo and Y. Park, Nat. Cell Biol., 2021, 23, 1329–1337.
83. S. Cheng, S. Fu, Y. M. Kim, W. Song, Y. Li, Y. Xue, J. Yi and L. Tian, Sci. Adv., 2021, 7, eabe0431.
84. C. Ounkomol, S. Seshamani, M. M. Maleckar, F. Collman and G. R. Johnson, Nat. Methods, 2018, 15, 917–920.
85. Y. Zhang, K. de Haan, Y. Rivenson, J. Li, A. Delis and A. Ozcan, Light: Sci. Appl., 2020, 9, 78.
86. K. de Haan, Y. Zhang, J. E. Zuckerman, T. Liu, A. E. Sisk, M. F. P. Diaz, K.-Y. Jen, A. Nobori, S. Liou, S. Zhang, R. Riahi, Y. Rivenson, W. D. Wallace and A. Ozcan, Nat. Commun., 2021, 12, 4884.
87. X. Xing, J. Shi, Y. Jia, Y. Dou, Z. Li, B. Dong, T. Guo, X. Cheng, X. Li, H. Du, Y. Hu, S. Jia, J. Zhang, Z. Li and J. Ji, Cancer Commun., 2022, 10, e003984.
88. H. M. Lai, Y. Tang, Z. Y. H. Lau, R. A. A. Campbell, J. C. N. Yau, C. C. Y. Chan, D. C. W. Chan, T. Y. Wong, H. K. T. Wong, L. Y. C. Yan, W. K. K. Wu, S. H. Wong, K.-W. Kwok, Y.-K. Wing, H. H. N. Lam, H.-K. Ng, T. D. Mrsic-Flogel, V. C. T. Mok, J. Y. K. Chan and H. Ko, Nat. Methods, 2022, 19, 1137–1146.
89. I. Laevsky, in Atlas of Human Pluripotent Stem Cells: Derivation and Culturing, ed. M. Amit and J. Itskovitz-Eldor, Humana Press, Totowa, NJ, 2012, pp. 105–113 DOI:10.1007/978-1-61779-548-0_7.
90. V. Schacht and J. S. Kern, J. Invest. Dermatol., 2015, 135, 1–4.
91. R. Piña, A. I. Santos-Díaz, E. Orta-Salazar, A. R. Aguilar-Vazquez, C. A. Mantellero, I. Acosta-Galeana, A. Estrada-Mondragon, M. Prior-Gonzalez, J. I. Martinez-Cruz and A. Rosas-Arellano, Int. J. Mol. Sci., 2022, 23, 1426.
92. K. Im, S. Mareninov, M. F. P. Diaz and W. H. Yong, Biobanking: Methods and Protocols, 2019, pp. 299–311.
93. J. M. Taube, K. Roman, E. L. Engle, C. Wang, C. Ballesteros-Merino, S. M. Jensen, J. McGuire, M. Jiang, C. Coltharp and B. Remeniuk, J. Immunother. Cancer, 2021, 9, e002197.
94. E. C. Stack, C. Wang, K. A. Roman and C. C. Hoyt, Methods, 2014, 70, 46–58.
95. K. Merrett, R. M. Cornelius, W. G. McClung, L. D. Unsworth and H. Sheardown, J. Biomater. Sci., Polym. Ed., 2002, 13, 593–621.
96. D. Raković and D. B. Uskoković, Društvo za istraživanje materijala Srbije, 2010, 221–240.
97. G. van Samkar, P. P. S. Balraadjsing, H. Hermanns, I. V. Hoogendijk, M. W. Hollmann, S. A. J. Zaat and M. F. Stevens, Reg. Anesth. Pain Med., 2020, 45, 381.
98. M. Essani, B. Charbonnier, N. Stephant, H. Moussi, P. Weiss, J. Le Bideau and P. Abellan, Mater. Adv., 2023, 4, 2474–2486.
99. K. S. Dae, J. H. Chang, K. Koo, J. Park, J. S. Kim and J. M. Yuk, ACS Omega, 2020, 5, 14619–14624.
100. J. Xin, S. Han, M. Zheng, C. Xu, Z. Huang, B. Wang, C. Yu, F. An and Y. Ren, Chin. Chem. Lett., 2024, 35, 109165.
101. D. G. M. Tiernan and A. T. Devitt, Ir. J. Med. Sci., 2019, 188, 987–992.
102. Ž. Mitić, A. Stolić, S. Stojanović, S. Najman, N. Ignjatović, G. Nikolić and M. Trajanović, Mater. Sci. Eng., C, 2017, 79, 930–949.
103. A. Diaspro, F. Federici and M. Robello, Appl. Opt., 2002, 41, 685–690.
104. R. Carriles, L. E. Zavala-García and L. Sánchez-Segura, in Microscopic Techniques for the Non-Expert, ed. S.-K. Kamaraj, A. Thirumurugan, S. S. Dhanabalan and S. A. Hevia, Springer International Publishing, Cham, 2022, pp. 53–81 DOI:10.1007/978-3-030-99542-3_3.
105. T. Takeda, S. Iwatsuki, S. Nozaki, A. Okada, K. Mizuno, Y. Umemoto and T. Yasui, Andrology, 2023, 11, 1147–1156.
106. C. Deng, M. Zheng, J. Xin and F. An, J. Colloid Interface Sci., 2023, 651, 384–393.
107. K. Yeh, I. Sharma, K. Falahkheirkhah, M. P. Confer, A. C. Orr, Y.-T. Liu, Y. Phal, R.-J. Ho, M. Mehta, A. Bhargava, W. Mei, G. Cheng, J. C. Cheville and R. Bhargava, Nat. Commun., 2023, 14, 5215.
108. K. F. Mallett and E. M. Arruda, Acta Biomater., 2017, 56, 44–57.
109. J. E. Ackerman, I. Bah, J. H. Jonason, M. R. Buckley and A. E. Loiselle, Microsyst. Nanoeng., 2017, 35, 2716–2724.
110. E. Song, Y. Huang, N. Huang, Y. Mei, X. Yu and J. A. Rogers, Microsyst. Nanoeng., 2022, 8, 77.
111. C. G. Fontanella, C. Salmaso, I. Toniolo, N. de Cesare, A. Rubini, G. M. De Benedictis and E. L. Carniel, Ann. Biomed. Eng., 2019, 47, 1237–1249.
112. S. Budday, T. C. Ovaert, G. A. Holzapfel, P. Steinmann and E. Kuhl, Arch. Comput. Methods Eng., 2020, 27, 1187–1230.
113. M. Griffin, Y. Premakumar, A. Seifalian, P. E. Butler and M. Szarko, J. Visualized Exp., 2016, e54872.
114. D. A. Morrow, T. L. H. Donahue, G. M. Odegard and K. R. Kaufman, J. Mech. Behav. Biomed. Mater., 2010, 3, 124–129.
115. C. N. Maganaris, J. Biomech., 2002, 35, 1019–1027.
116. E. Newell and M. Driscoll, IEEE Trans. Biomed. Eng., 2024, 71, 2678–2683.
117. S. Zhao, M. Arnold, S. Ma, R. Abel, J. Cobb, U. Hansen and O. Boughton, Bone Jt. Res., 2018, 7, 524–538.
118. E. Donnelly, Clin. Orthop. Relat. Res., 2011, 469, 2128–2138.
119. B. Johnson, S. Campbell and N. Campbell-Kyureghyan, Med. Eng. Phys., 2020, 80, 44–51.
120. J. Barsimantov, J. Payne, M. de Lucio, M. Hakim, H. Gomez, L. Solorio and A. B. Tepole, Ann. Biomed. Eng., 2024, 52, 1638–1652.
121. J. Betten, Creep mechanics, Springer Science & Business Media, 2008.
122. E. Kula, Residual stress and stress relaxation, Springer Science & Business Media, 2013.
123. A. Anssari-Benam, H. R. Screen and A. Bucchi, J. Mech. Behav. Biomed. Mater., 2019, 93, 230–245.
124. W. Lear, C. M. Blattner, T. A. Mustoe and J. J. Kruzic, J. Mech. Behav. Biomed. Mater., 2019, 97, 85–89.
125. D. S. Jones, Int. J. Pharm., 1999, 179, 167–178.
126. K. P. Menard and N. Menard, Dynamic mechanical analysis, CRC press, 2020.
127. J. H. Edwards, E. Ingham and A. Herbert, J. Mech. Behav. Biomed. Mater., 2019, 91, 18–23.
128. L. Islamagič, M. J. Tjørnild, L. Carlson Hanse, J. V. Nygaard and J. M. Hasenkam, Proc. Inst. Mech. Eng., Part H, 2023, 237, 435–442.
129. D. S. Richardson, W. Guan, K. Matsumoto, C. Pan, K. Chung, A. Ertürk, H. R. Ueda and J. W. J. N. R. M. P. Lichtman, Nat. Rev. Methods Primers, 2021, 1, 84.
130. S. Zhao, M. I. Todorov, R. Cai, R. A. I. Maskari, H. Steinke, E. Kemter, H. Mai, Z. Rong, M. Warmer, K. Stanic, O. Schoppe, J. C. Paetzold, B. Gesierich, M. N. Wong, T. B. Huber, M. Duering, O. T. Bruns, B. Menze, J. Lipfert, V. G. Puelles, E. Wolf, I. Bechmann and A. Ertürk, Cell, 2020, 180, 796–812.e719.
131. Y. Zhao, O. Bucur, H. Irshad, F. Chen, A. Weins, A. L. Stancu, E.-Y. Oh, M. DiStasio, V. Torous and B. Glass, Nat. Biotechnol., 2017, 35, 757–764.
132. W. Menegas, J. F. Bergan, S. K. Ogawa, Y. Isogai, K. Umadevi Venkataraju, P. Osten, N. Uchida and M. Watabe-Uchida, eLife, 2015, 4, e10032.
133. S. I. Kubota, K. Takahashi, J. Nishida, Y. Morishita, S. Ehata, K. Tainaka, K. Miyazono and H. R. Ueda, Cell Rep., 2017, 20, 236–250.
134. I. Nehrhoff, J. Ripoll, R. Samaniego, M. Desco and M. V. Gómez-Gaviro, Biomed. Opt. Express, 2017, 8, 3110–3118.
135. L. Yildirimer, Q. Zhang, S. Kuang, C.-W. J. Cheung, K. A. Chu, Y. He, M. Yang and X. J. B. Zhao, Biofabrication, 2019, 11, 032003.
136. A. Ertürk, Nat. Methods, 2024, 21, 1153–1165.
137. Z. Jin, Y. Li, K. Yu, L. Liu, J. Fu, X. Yao, A. Zhang and Y. He, Adv. Sci., 2021, 8, 2101394.
138. N. Farahani, A. Braun, D. Jutt, T. Huffman, N. Reder, Z. Liu, Y. Yagi and L. Pantanowitz, J. Pathol. Inf., 2017, 8, 36.
139. J. T. C. Liu, A. K. Glaser, K. Bera, L. D. True, N. P. Reder, K. W. Eliceiri and A. Madabhushi, Nat. Biomed. Eng., 2021, 5, 203–218.
140. L. A. Earl, V. Falconieri, J. L. Milne and S. Subramaniam, Curr. Opin. Struct. Biol., 2017, 46, 71–78.
141. M. C. DiIorio and A. W. J. J. Kulczyk, J. Visualized Exp., 2022, e63387.
142. D. Tegunov and P. Cramer, Nat. Methods, 2019, 16, 1146–1152.
143. E. Y. Chua, J. H. Mendez, M. Rapp, S. L. Ilca, Y. Z. Tan, K. Maruthi, H. Kuang, C. M. Zimanyi, A. Cheng and E. T. Eng, Annu. Rev. Biochem., 2022, 91, 1–32.
144. R. Danev, H. Yanagisawa and M. Kikkawa, Trends Biochem. Sci., 2019, 44, 837–848.
145. P. Zhang, D. Ma, X. Cheng, A. Tsai, Y. Tang, H.-C. Gao, L. Fang, C. Bi, G. Landreth, A. Chubykin and F. Huang, Biophys. J., 2023, 122, 430a.
146. K. Galler, K. Bräutigam, C. Große, J. Popp and U. J. A. Neugebauer, Analyst, 2014, 139, 1237–1273.
147. X. Chen, S. A. Teichmann and K. B. Meyer, Annu. Rev. Biomed. Data Sci., 2018, 1, 29–51.
148. S. Baek and I. Lee, Comput. Struct. Biotechnol. J., 2020, 18, 1429–1439.
149. A. Regev, S. A. Teichmann, E. S. Lander, I. Amit, C. Benoist, E. Birney, B. Bodenmiller, P. Campbell, P. Carninci and M. Clatworthy, elife, 2017, 6, e27041.
150. R. Ahmed, T. Zaman, F. Chowdhury, F. Mraiche, M. Tariq, I. S. Ahmad and A. Hasan, Int. J. Mol. Sci., 2022, 23, 3042.
151. T. Alexandrov, Nat. Metab., 2023, 5, 1443–1445.
152. A. Baysoy, Z. Bai, R. Satija and R. J. N. R. M. C. B. Fan, Nat. Rev. Mol. Cell Biol., 2023, 24, 695–713.
153. L. Chappell, A. J. Russell and T. Voet, Annu. Rev. Genomics Hum. Genet., 2018, 19, 15–41.
154. M. Tellez-Gabriel, B. Ory, F. Lamoureux, M.-F. Heymann and D. Heymann, Int. J. Mol. Sci., 2016, 17, 2142.
155. M. Ke, B. Elshenawy, H. Sheldon, A. Arora and F. M. J. B. Buffa, BioEssays, 2022, 44, 2200084.
156. Y. LeCun, Y. Bengio and G. Hinton, Nature, 2015, 521, 436–444.
157. J. M. Michalska, J. Lyudchik, P. Velicky, H. Štefaničková, J. F. Watson, A. Cenameri, C. Sommer, N. Amberg, A. Venturino, K. Roessler, T. Czech, R. Höftberger, S. Siegert, G. Novarino, P. Jonas and J. G. Danzl, Nat. Biotechnol., 2024, 42, 1051–1064.
158. G. Campanella, M. G. Hanna, L. Geneslaw, A. Miraflor, V. Werneck Krauss Silva, K. J. Busam, E. Brogi, V. E. Reuter, D. S. Klimstra and T. J. Fuchs, Nat. Med., 2019, 25, 1301–1309.
159. E. J. J. N. m Topol, Nat. Med., 2019, 25, 44–56.
160. S. Koivukoski, U. Khan, P. Ruusuvuori and L. J. L. I. Latonen, Lab. Invest., 2023, 103, 100070.
161. U. Khan, S. Koivukoski, M. Valkonen, L. Latonen and P. J. P. Ruusuvuori, Patterns, 2023, 4, 100725.
162. M. Gonzalez Lopez, B. Huteckova, J. Lavicky, N. Zezula, V. Rakultsev, V. Fridrichova, H. Tuaima, C. Nottmeier, J. Petersen and M. J. S. A. Kavkova, Sci. Adv., 2023, 9, eadi0482.
163. C. Ma, W. Tan, R. He and B. J. N. M. Yan, Nat. Methods, 2024, 1–10.
164. S. Dorkenwald, P. H. Li, M. Januszewski, D. R. Berger, J. Maitin-Shepard, A. L. Bodor, F. Collman, C. M. Schneider-Mizell, N. M. da Costa, J. W. Lichtman and V. Jain, Nat. Methods, 2023, 20, 2011–2020.
165. M. Januszewski and V. J. N. m Jain, Nat. Methods, 2024, 21, 1398–1399.
166. D. Soteriou, M. Kubánková, C. Schweitzer, R. López-Posadas, R. Pradhan, O.-M. Thoma, A.-H. Györfi, A.-E. Matei, M. Waldner, J. H. W. Distler, S. Scheuermann, J. Langejürgen, M. Eckstein, R. Schneider-Stock, R. Atreya, M. F. Neurath, A. Hartmann and J. Guck, Nat. Biomed. Eng., 2023, 7, 1392–1403.
167. E. Berson, P. Chung, C. Espinosa, T. J. Montine and N. Aghaeepour, Nat. Methods, 2024, 21, 1400–1402.
168. J. L. Fan, A. Nazaret and E. Azizi, Nat. Methods, 2024, 21, 1403–1406.
169. R. Gail Canter, W.-C. Huang, H. Choi, J. Wang, L. Ashley Watson, C. G. Yao, F. Abdurrob, S. M. Bousleiman, J. Z. Young and D. A. Bennett, Commun. Biol., 2019, 2, 360.
170. C. Pan, O. Schoppe, A. Parra-Damas, R. Cai, M. I. Todorov, G. Gondi, B. von Neubeck, N. Böğürcü-Seidel, S. Seidel and K. J. C. Sleiman, Cell, 2019, 179, 1661–1676. e1619.
171. A. H. Song, M. Williams, D. F. Williamson, S. S. Chow, G. Jaume, G. Gao, A. Zhang, B. Chen, A. S. Baras and R. J. C. Serafin, Cell, 2024, 187, 2502–2520. e2517.
172. M. Venkatesan, H. Mohan, J. R. Ryan, C. M. Schürch, G. P. Nolan, D. H. Frakes and A. F. Coskun, Cell Rep. Med., 2021, 2, 100348.
173. J. Bernett, D. B. Blumenthal, D. G. Grimm, F. Haselbeck, R. Joeres, O. V. Kalinina and M. J. N. M. List, Nat. Methods, 2024, 21, 1444–1453.
174. O. Rotem and A. Zaritsky, Nat. Methods, 2024, 21, 1394–1397.
175. J. Smith, Y. Shi, M. Benedikt and M. Nikolic, GigaScience, 2021, 10, giab058.
176. T. Meurers, R. Bild, K.-M. Do and F. Prasser, GigaScience, 2021, 10, giab068.
177. M. Sepp, K. Leiss, F. Murat, K. Okonechnikov, P. Joshi, E. Leushkin, L. Spänig, N. Mbengue, C. Schneider, J. Schmidt, N. Trost, M. Schauer, P. Khaitovich, S. Lisgo, M. Palkovits, P. Giere, L. M. Kutscher, S. Anders, M. Cardoso-Moreira, I. Sarropoulos, S. M. Pfister and H. Kaessmann, Nature, 2024, 625, 788–796.
178. S. Jiang, H. Yu, T. Zhao, Y. Lu, X. Li, L. Gao, J. Mi, W. Xia, L. Geng, P. Li, M. Chen, M. Kang, J. Liu and Y. Ke, Innovation, 2025, 6, 100767.
179. O. Chakoory, V. Barra, E. Rochette, L. Blanchon, V. Sapin, E. Merlin, M. Pons, D. Gallot, S. Comtet-Marre and P. Peyret, Biomarker Res., 2024, 12, 25.
180. B. Ehteshami Bejnordi, M. Mullooly, R. M. Pfeiffer, S. Fan, P. M. Vacek, D. L. Weaver, S. Herschorn, L. A. Brinton, B. van Ginneken, N. Karssemeijer, A. H. Beck, G. L. Gierach, J. A. W. M. van der Laak and M. E. Sherman, Mod. Pathol., 2018, 31, 1502–1512.
181. K. Ding, M. Zhou, H. Wang, S. Zhang and D. N. Metaxas, Lancet Digital Health, 2022, 4, e787–e795.
182. S. Steyaert, M. Pizurica, D. Nagaraj, P. Khandelwal, T. Hernandez-Boussard, A. J. Gentles and O. Gevaert, Nat. Mach. Intell., 2023, 5, 351–362.
183. J. Jee, C. Fong, K. Pichotta, T. N. Tran, A. Luthra, M. Waters, C. Fu, M. Altoe, S.-Y. Liu, S. B. Maron, M. Ahmed, S. Kim, M. Pirun, W. K. Chatila, I. de Bruijn, A. Pasha, R. Kundra, B. Gross, B. Mastrogiacomo, T. J. Aprati, D. Liu, J. Gao, M. Capelletti, K. Pekala, L. Loudon, M. Perry, C. Bandlamudi, M. Donoghue, B. A. Satravada, A. Martin, R. Shen, Y. Chen, A. R. Brannon, J. Chang, L. Braunstein, A. Li, A. Safonov, A. Stonestrom, P. Sanchez-Vela, C. Wilhelm, M. Robson, H. Scher, M. Ladanyi, J. S. Reis-Filho, D. B. Solit, D. R. Jones, D. Gomez, H. Yu, D. Chakravarty, R. Yaeger, W. Abida, W. Park, E. M. O’Reilly, J. Garcia-Aguilar, N. Socci, F. Sanchez-Vega, J. Carrot-Zhang, P. D. Stetson, R. Levine, C. M. Rudin, M. F. Berger, S. P. Shah, D. Schrag, P. Razavi, K. L. Kehl, B. T. Li, G. J. Riely, N. Schultz, A. Lisman, G. Zhao, I. de Bruijn, W. K. Chatila, X. Li, A. Kohli, D. Moore, R. Lim, T. Pollard, R. Sheridan, A. Wang, C. Chennault, M. Wilson, H. Zhang, R. Pimienta, S. Rangavajhala, G. Subramanian, J. Garcia, N. Rachuri, K. Boehm, M. Parker, H. Walch, S. Nandakumar, J. Eichholz, A. Kris, P. Manca, X. Bai, T. Agbamu, U. Justin, X. Bi and MSK Cancer Data Science Initiative Group, Nature, 2024, 636, 728–736.
184. F. Carrillo-Perez, J. C. Morales, D. Castillo-Secilla, O. Gevaert, I. Rojas and L. J. Herrera, J. Pers. Med., 2022, 12, 601.

---