Updated:
In Vivo Fluorescence Imaging: Applications in Preclinical Research and Small Animal Studies
SUMMARY
Fluorescence in vivo imaging has become the standard tool for longitudinal preclinical studies. This guide covers the applications side: what it enables in oncology (tumor tracking, metastasis, therapy response), infectious disease (bacterial infection, antimicrobial efficacy), neuroscience (through-skull brain imaging with NIR-II), cardiovascular research (atherosclerosis, ischemia-reperfusion), and immunology (cell tracking, biodistribution). Includes fluorophore recommendations by application and workflow best practices from acquisition to publication-ready data.
Why fluorescence imaging became a standard in preclinical research
The rise of in vivo fluorescence imaging is driven by three shifts in how preclinical research is done today.
Longitudinal studies are now widely used. Regulatory and journal expectations increasingly require repeated measurements on the same animal across a study. Fluorescence imaging is non-invasive and non-ionizing, which makes it compatible with repeated imaging over weeks or months without harming the animal.
Animal numbers are under pressure. The 3R principles (Replacement, Reduction, Refinement) push researchers to extract more data from fewer animals. By acquiring repeated longitudinal measurements, a single mouse can replace multiple cross-sectional cohorts, increasing statistical power while substantially reducing animal use and inter-animal variability
Publication standards demand quantitative data. Journals expect numeric fold changes with confidence intervals, not just representative images. This requires quantitative acquisition with calibrated instruments, standardized ROI methods, and statistically valid replicates. Modern in vivo imaging system for small animals now include NIST-traceable calibration natively, which makes cross-day and cross-instrument comparison possible.
For researchers, in vivo optical imaging techniques share one practical advantage: they use visible or near-infrared light rather than ionizing radiation, so the same animal can be imaged again and again. A light source excites the fluorophore, and the emitted light is collected and quantified. Because in vivo optical imaging depends on how light travels through biological tissues, the optical properties of the tissue (absorption, scattering, autofluorescence) set the practical limits, and researchers pick wavelengths accordingly. In vivo applications now span molecular imaging of receptor expression, cellular activity in immune and tumor cells, and disease progression in models of human disease, giving researchers a real-time visualization that other imaging modalities rarely provide.
Optical imaging among preclinical imaging modalities
Preclinical research draws on several imaging modalities, and optical imaging occupies a specific niche among them. Positron emission tomography (PET) visualizes biological processes using positron-emitting radionuclides, magnetic resonance imaging (MRI) resolves soft-tissue anatomy, and computed tomography (CT) and ultrasound add structural context. Each of these medical imaging techniques carries its own cost and infrastructure. In vivo fluorescence imaging is cost-effective compared to PET or MRI, needs no radiotracers, and gives real-time visualization of molecular processes in living animals. Optical imaging was first used preclinically over 20 years ago, and it has since become a routine choice for longitudinal molecular imaging in small animals. For the comparison between fluorescence, bioluminescence, and NIR-II specifically, our guide to in vivo imaging technologies covers that axis; here we stay on optical imaging and its applications across preclinical studies.
Detector sensitivity for low-light in vivo imaging
In vivo fluorescence and bioluminescence are low light applications, so detector sensitivity sets the practical detection limit. In vivo imaging systems use CCD or InGaAs detectors: typical CCD detectors have peak quantum efficiency in the mid-visible range, while InGaAs detectors cover the wavelength range 900 to 1700 nm required for NIR-II. At very low levels of signal, two noise sources dominate: readout noise from the sensor electronics and background noise from thermal effects in the chip. Deep cooling of cameras can reduce noise to improve imaging, and the Newton platform cools its sensor down to -90 degrees C to ensure high sensitivity for faint emitted light. A wide f/0.70 aperture (the widest lens in the range) collects more light per exposure, which raises the signal to noise ratio without lengthening acquisition. These sensitivity factors (aperture and cooling) determine the ability to detect weak fluorescent or bioluminescent signals above background levels and are critical across the in vivo imaging applications mentionned below. Because in vivo optical imaging reads faint signals from a whole living organism, we benchmark quantum efficiency, readout noise, and light source stability on every camera before it is used for in vivo optical imaging.
Small animal imaging: capacity and throughput
Throughput matters as much as sensitivity for labs running large cohorts of small animals. The Newton platform images 1 to 10 mice per session, and its field of view scales from 6 x 6 cm for macro imaging up to 20 x 20 cm, which also accommodates rat imaging and larger fixtures. This capacity makes an in vivo imaging system for small animals a shared resource across projects rather than a single-user tool. For the animal models used in preclinical studies (mice, and rats when the model requires them), a heated bed at 37 degrees C, EQUAFLOW breathers, and a motorized Z-axis keep laboratory animals stable and reproducibly positioned. Imaging living animals at several time points, rather than sacrificing separate cohorts of small animals, is what lets a single living organism carry a full longitudinal dataset, and it is why we treat throughput and animal welfare as part of the same design for small animals.
X-ray co-registration for anatomical validation
For studies that need anatomical context, the Newton platform adds X-ray co-registration, overlaying a fluorescent or bioluminescent signal onto an X-ray image of the same animal. The X-ray channel ties the signal to skeletal and soft-tissue anatomical structures, which supports translational validation when a functional readout has to be matched to a specific organ or lesion. Combining multispectral imaging with X-ray in one session lets us map several biological processes onto a single anatomical reference, providing insights that neither channel gives alone. Because the probe signal originates inside biological tissues, correlating it with X-ray landmarks improves how we interpret deep-tissue results across models of human disease and their biological tissues.
Oncology applications: tumor tracking, metastasis, and therapy response
Oncology remains the largest application area for in vivo fluorescence imaging. Three research questions dominate:
Tumor growth kinetics. In xenograft or syngeneic models, fluorescent tumor cells (transfected with GFP, mCherry, or luciferase for bioluminescence) can be tracked from implantation through growth, plateau, and response to therapy. Imaging sessions on days 3, 7, 14, and 21 post-implantation produce growth curves that would previously have required 4 separate animal cohorts. The Vilber Newton FT platform includes a serial acquisition mode that automatically organizes images acquired over multiple time points, making it easier to monitor tumor progression and extract longitudinal signal kinetics from the same animals.
Metastasis detection. Distant metastases (lung, liver, bone) are difficult to detect with standard fluorescence in the visible range because tissue scattering limits the penetration depth. NIR-I fluorescence (700 to 900 nm) improves depth by reducing scattering, and NIR-II (1 000 to 1 700 nm) improves it further, allowing detection of metastases several millimeters below the surface. For deep organ metastases, NIR-II probes combined with the appropriate detector reveal signals that NIR-I misses entirely.
Therapy response monitoring. Comparing tumor volume or fluorescence intensity across a treated cohort and a control cohort is the classic readout for therapy response. The advantage of fluorescence over caliper-based tumor measurement is that fluorescence tracks live tumor cells, not necrotic tissue, and reveals response earlier and more accurately.
Model selection matters as much as fluorophore choice.
- Subcutaneous xenografts: easy to image, minimal depth issues, standard NIR-I fluorophores work well
- Orthotopic xenografts: closer to human biology, require deeper penetration, favor NIR-II or bioluminescence
- Patient-derived xenografts (PDX): better translational value, similar imaging requirements to orthotopic
- Syngeneic mouse models: essential for immunotherapy studies, require multiplex imaging to track both tumor and immune cells
- Genetically engineered mouse models (GEMM): spontaneous tumor development, often require whole-body imaging with high sensitivity
Fluorophore recommendations for oncology:
- NIR-I (700–900 nm): indocyanine green (ICG) for tumor perfusion, IR-780 for tumor cell labeling, IRDye 800CW conjugated to antibodies for targeted imaging
- NIR-II (1 000–1 700 nm): emerging small-molecule NIR-II dyes and antibody conjugates for deep-tissue tumor imaging
- Bioluminescence: luciferase-expressing tumor cells for maximum sensitivity, useful when depth or scatter compromises fluorescence
Our guide to developing and validating NIR-II probes covers the probe chemistry and validation workflow in detail.
Infectious disease and antimicrobial research
Infectious disease research relies heavily on live imaging to track pathogens over time, evaluate host response, and test antimicrobial candidates. Two approaches dominate.
Bioluminescent pathogen strains carry a luciferase gene (typically lux or luc) that emits light when the bacteria are metabolically active. Bioluminescence imaging in infection models produces exceptional signal-to-noise because there is essentially no background in mammalian tissue. Standard strains exist for Staphylococcus aureus, Escherichia coli, Pseudomonas aeruginosa, Mycobacterium tuberculosis, and many others. The signal correlates with bacterial load, which makes bioluminescence a natural fit for tracking infection progression and treatment response.
In vivo bioluminescence suits infection models because there is no excitation light and therefore almost no tissue background. In vivo bioluminescence imaging quantifies metabolically active bacterial load over time, and pairing bioluminescence and fluorescence imaging in one acquisition tracks the pathogen and the host response together. This makes bioluminescence and fluorescence imaging a natural combination for antimicrobial studies, where drug effect on bacterial load and host inflammation are read out side by side. For high performance in these low-signal conditions, in vivo bioluminescence relies on a deeply cooled sensor and a wide aperture to recover faint photon counts.
Fluorescent probes targeting bacterial components offer an alternative when engineered bioluminescent strains are not available or not appropriate. Vancomycin-conjugated fluorophores label Gram-positive bacterial cell walls. Zinc-binding probes detect bacterial biofilms. Antimicrobial peptide conjugates target specific pathogens. These probes work with wild-type bacteria and clinical isolates, which is essential for translational studies.
Common research questions:
- Infection kinetics: how fast does the pathogen colonize the target organ?
- Antimicrobial efficacy: does the candidate drug reduce bacterial load in vivo, and how fast?
- Host response: how does the immune system respond to infection, and where do immune cells accumulate?
- Biofilm formation: does the pathogen form biofilms that resist antibiotic treatment?
Model considerations for infectious disease imaging:
- Sepsis and systemic infection: require whole-body imaging with sensitivity across the animal
- Skin and soft tissue infections: benefit from high-resolution imaging of a defined field of view
- Pulmonary infection: require deep-tissue penetration, favor NIR-II or bioluminescence
- Wound infection: often combined with high-resolution photography and fluorescence overlay
For sepsis studies specifically, bioluminescence with a whole-body imager provides the most robust quantification, since bacterial load in disseminated infection is spatially heterogeneous and requires imaging the entire animal.
Neuroscience applications: brain imaging through intact skull
The skull absorbs and scatters visible and near-infrared light, which historically limited in vivo brain imaging to invasive approaches (cranial windows, skull-thinning) or post-mortem histology. The emergence of NIR-II fluorescence changed this equation.
NIR-II imaging (1 000–1 700 nm) penetrates the intact mouse skull with usable signal. This opens non-invasive neuroscience workflows that were previously impossible:
- Cerebrovascular imaging: NIR-II probes reveal cortical and deeper vascular networks through the skull, useful for stroke models, hemorrhage studies, and cerebral perfusion
- Neurodegeneration models: Alzheimer, Parkinson, ALS models can be imaged longitudinally through the skull, tracking disease progression with limited invasive surgery
- Blood-brain barrier studies: probes designed to cross or fail to cross the BBB can be quantified in the brain versus periphery
- Neurovascular coupling: pairing NIR-II vascular imaging with behavioral or physiological measurements links brain activity and blood flow
The Vilber Newton FT-900 was specifically designed for NIR-II applications, with a detector optimized for the 1 000 to 1 700 nm range. For the technical criteria that make NIR-II possible (detector choice, spectral filtering, sensor sensitivity), our article on NIR-II detectors covers the hardware requirements.
When to use NIR-II vs alternatives in neuroscience:
- Superficial cortex imaging: NIR-I sometimes suffices, but NIR-II gives better resolution
- Deep brain structures (hippocampus, subcortical): NIR-II is often the only non-invasive option
- Longitudinal disease progression: NIR-II supports repeat imaging over weeks or months without cranial windows
- Functional imaging with high temporal resolution: bioluminescence or fluorescence with genetically encoded sensors, depending on the temporal requirements
Fluorophore recommendations for neuroscience:
- NIR-II cerebrovascular imaging: PbS quantum dots, small-molecule NIR-II dyes (CH1055 and derivatives)
- BBB and neurodegeneration: peptide-conjugated NIR-II probes designed to target amyloid or specific brain markers
- Multiplex neuroscience: combine bioluminescence (activity readout) with NIR-II fluorescence (vascular or structural readout) for functional plus anatomical information
Cardiovascular and inflammation research
Vascular imaging is one of the natural strengths of in vivo fluorescence, since blood-borne fluorophores are naturally distributed through the vasculature. Three research areas dominate.
Atherosclerosis and plaque imaging. Fluorescent probes targeting oxidized LDL, macrophages, or matrix metalloproteinases (MMPs) accumulate in atherosclerotic plaques. Longitudinal imaging tracks plaque progression, response to statins or novel therapies, and plaque stability markers. NIR-I probes work for surface arteries, NIR-II for deeper vascular structures.
Ischemia-reperfusion injury. In myocardial or cerebral ischemia models, fluorescent probes reveal the extent of ischemic damage and the effectiveness of reperfusion therapy. Real-time imaging during the reperfusion phase captures the dynamics that endpoint measurements miss. Probes for reactive oxygen species (ROS), apoptosis (annexin V conjugates), or vascular integrity provide complementary readouts.
Inflammation and vascular permeability. Activatable fluorescent probes, which are non-fluorescent in their inactive state and become fluorescent when cleaved by specific enzymes (MMPs, cathepsins, caspases), reveal inflammation with high signal-to-noise. Because the probe generates limited fluorescence before activation, signal accumulation reflects local enzymatic activity and provides improved contrast at sites of inflammation.
Model considerations for cardiovascular imaging:
- Cardiac imaging requires either NIR-II penetration or bioluminescence with luciferase-expressing cardiomyocytes, since the heart is deep and constantly moving
- Vascular imaging in limbs and periphery works well with NIR-I probes
- Whole-body inflammation benefits from activatable probes with low background
The multispectral imaging capability of modern in vivo systems is particularly valuable here, since separating multiple fluorescent probes in the same animal allows tracking of different biological processes simultaneously (perfusion, inflammation, apoptosis).
Immunology and biodistribution studies
Cell tracking and drug biodistribution are two of the most common applications of in vivo animal imaging.
Immune cell tracking. Adoptive cell therapies (CAR-T, CAR-NK, TIL) rely on injected immune cells finding their target tissue. Labeling these cells with a fluorescent dye (DiR, DiI) before injection allows tracking their migration, homing to tumors or lymphoid organs, and persistence over time. This is particularly valuable for evaluating novel cell therapy candidates in preclinical models before clinical translation.
Drug pharmacokinetics and biodistribution. Fluorescently labeled drugs, antibodies, or nanoparticles reveal where the compound accumulates, how long it persists, and how quickly it is cleared. Longitudinal imaging over hours, days, or weeks builds a full PK profile from a single mouse rather than a cohort of animals sacrificed at each time point.
Nanoparticle tracking. In vivo imaging is often the primary tool for evaluating nanoparticle-based drug delivery systems. Fluorescent nanoparticles reveal enhanced permeability and retention (EPR) effects in tumors, off-target accumulation in liver and spleen, and stability of the nanoparticle in circulation.
Gene therapy monitoring. Reporter genes (luciferase, fluorescent proteins) expressed after viral vector delivery allow monitoring of transgene expression in the target tissue. Bioluminescence often outperforms fluorescence here because of the very low background, and because the reporter is only active in cells that received the vector.
Fluorophore recommendations for immunology and biodistribution:
- Cell tracking: DiR (near-infrared lipophilic dye) for long-term labeling, DiI for shorter studies, Vybrant CFSE for proliferation tracking
- Antibody biodistribution: IRDye 800CW conjugates for NIR-I, NIR-II conjugates for deeper tissues
- Nanoparticle tracking: fluorophore choice depends on the particle chemistry, with encapsulated dyes protected from degradation and providing longer imaging windows
- Reporter gene expression: firefly luciferase (bioluminescence) for maximum sensitivity, fluorescent proteins (mCherry, GFP) when live imaging with cellular resolution is required
Choosing fluorophores by application
The fluorophore choice depends on three parameters: the target biology, the depth of the tissue, and the imaging system available. The table below summarizes the recommendations across the applications covered above.
Practical considerations when selecting a fluorophore:
- Emission wavelength drives tissue penetration. Longer wavelengths (NIR-II) reach deeper structures with less scatter. For surface tissues, NIR-I is often sufficient and less expensive.
- Brightness matters for low-abundance targets. High extinction coefficient and quantum yield combine to determine the practical detection limit.
- Stability under illumination determines imaging duration. Photobleaching limits how many time points can be acquired without signal loss.
- Conjugation chemistry constrains the choice for targeted imaging. Antibody-fluorophore conjugates require reactive dyes (NHS ester, maleimide) compatible with the antibody chemistry.
- Autofluorescence is a background source that varies by wavelength. Blue and green fluorophores compete with tissue autofluorescence from collagen, elastin, and flavins. NIR-I and especially NIR-II have very low tissue autofluorescence, which improves signal-to-noise significantly.
Fluorescent dyes are not the only reporters available. Fluorescent contrast agents (untargeted dyes that highlight perfusion or vascular leak) and targeted probes (antibody or peptide conjugates) answer different questions, and activatable probes report an enzymatic reaction at the target site. When the goal is to screen therapeutic agents or rank new drugs by drug efficacy, a reporter with high sensitivity shortens acquisition and lowers the detection threshold. Matching the labeling technique to the biology, rather than defaulting to one dye, is what keeps quantification reliable, and the same logic applies to the labeling techniques used for cell tracking and biodistribution.
From acquisition to publication: workflow best practices
Reproducible in vivo fluorescence data comes from consistency at every step of the workflow, not from one heroic acquisition. The steps below describe the workflow adopted by labs that publish quantitative in vivo imaging routinely.
Animal preparation is often the most under-controlled step. Before imaging:
- Anesthesia: isoflurane is the standard for mouse imaging, delivered via a precision vaporizer with a heated animal bed to prevent hypothermia. The Newton platform includes EQUAFLOW breathers that deliver equal gas to each nose cone, preventing unwanted awakening during sessions with multiple animals.
- Fur removal: shaving or depilatory cream on the imaging area reduces light absorption and scattering. Black fur strains (C57BL/6) benefit most, since melanin absorbs strongly in the visible range.
- Positioning: the animal must be in a reproducible position across time points. A heated bed with fixed markers and standardized limb positioning is essential for longitudinal work.
- Autofluorescence control: chlorophyll from standard rodent chow produces strong autofluorescence in the gut, particularly at red and NIR-I wavelengths. A chlorophyll-free diet for at least one week before imaging significantly reduces this background.
Imaging parameters are set for reproducibility, not for maximum signal:
- Fixed exposure time across all animals in a comparison ensures that signal intensity is comparable
- Fixed aperture and binning apply the same photon collection settings across the study
- Fixed animal position and field of view keeps the geometry constant
- Multispectral imaging for multiplex studies requires spectral unmixing, ideally applied automatically by the acquisition software
ROI analysis determines the quantitative output:
- Automated ROI placement reduces operator bias, particularly for longitudinal studies where the same anatomical region must be measured across sessions
- Mouse atlas overlay aligns the acquired image with a reference mouse anatomy, allowing organ-specific ROI without manual placement
- Background subtraction applies a consistent method across all images in the study
- Absolute quantification in physical units (radiance, photons/s) supports cross-day and cross-instrument comparison, which is essential for longitudinal studies. NIST-traceable calibration is the technical requirement.
Statistical reporting:
- Biological replicates, not technical replicates, are the unit of statistical analysis
- Report individual data points, mean and standard error, and the exact test used
- Longitudinal data analysis uses repeated-measures ANOVA or mixed models, which account for the within-subject correlation
- Document the analysis pipeline in the methods, or better, in a supplementary file that reviewers can inspect
In Vivo Imaging Specialist & Global Sales Director
