Updated:
In Vivo Imaging for Preclinical Research: A Guide to Fluorescence Technologies (NIR-I, NIR-II, and Beyond)
SUMMARY
Choosing the right in vivo imaging system depends on three questions: what biological process do you want to visualize, how deep does it sit, and what level of quantitative detail do you need. This guide compares NIR-I, NIR-II, and bioluminescence imaging modalities, and walks through the technical criteria that actually matter when selecting a small animal imaging platform for preclinical research.
Choosing the right in vivo imaging system for preclinical research depends on three intertwined questions:
- What biological process do you want to visualize ?
- How deep does it sit in the tissue ?
- What level of quantitative detail do you need ?
The answers determine whether NIR-I fluorescence, NIR-II fluorescence, or bioluminescence is the appropriate imaging modality. This guide walks through each technology, where it excels, where it falls short, and how to make a decision that fits the experiment rather than the equipment available. It also reflects recent developments in detector design and probe chemistry that are providing in vivo insights once impossible to obtain.
We work with researchers running longitudinal oncology studies, infectious disease models, gene therapy biodistribution, and translational drug development. The pattern we see most often is the same modality being applied to every experiment, regardless of whether it is the right one. Modality choice is the single decision that has the largest downstream impact on data quality, statistical power, and reproducibility. Get it right, and the rest of the workflow runs smoothly. Get it wrong, and no amount of acquisition tuning will recover the missing signal.
What in vivo imaging means in preclinical research
In vivo imaging refers to the non-invasive visualization of biological processes inside living animals, typically rodents in preclinical contexts. Unlike ex vivo or in vitro techniques, in vivo optical imaging captures the dynamic behavior of cells, tissues, and molecules in their native environment, with the same animal followed across multiple time points.
The preclinical value of this approach is twofold. First, it dramatically reduces the number of animals required for a study, since each subject serves as its own internal control across the entire time course. Second, it preserves the biological context that is lost the moment tissues are excised. Tumor microenvironment, immune cell trafficking, drug pharmacokinetics, and gene expression dynamics all behave differently in a living organism than in dissected samples. Capturing these molecular processes in real time, within intact biological tissues, is the central advantage that distinguishes in vivo molecular imaging from purely ex vivo or histological methods.
Modern preclinical fluorescence imaging has therefore become a backbone technology in oncology, immunology, neurology, infectious disease, and biodistribution research. The modality you choose, however, dictates what you can actually see. A bioluminescence reporter cannot give you the spatial detail of a fluorescent probe. A NIR-I fluorophore cannot reach the depth that NIR-II achieves. A tomographic modality like computed tomography or MRI captures anatomy but misses the molecular signal optical imaging provides. These other imaging modalities remain complementary, but they cannot replace optical readouts when the goal is to track disease progression or measure drug efficacy at the molecular level.
Three optical modalities dominate preclinical imaging work today: visible and near-infrared fluorescence (VIS/ NIR-I imaging, 700 to 900 nm), short-wave infrared fluorescence (SWIR imaging, 1000 to 1700 nm), and bioluminescence. Each operates on different physical principles and addresses different biological questions. We treat them as complementary, not competing.
The current generation of in vivo imaging platforms also integrates X-Ray and 3D BLI Optical Tomography capabilities, which add structural reference to the optical signal. This kind of multi-modal fusion is increasingly the standard for translational research, where the question "where is the signal coming from" matters as much as "how much signal is there".
The three optical modalities at a glance
The table above summarizes the practical differences. Below we explain why each criterion matters in real experiments.
Penetration depth, scattering, and signal in biological tissues
Penetration depth is governed by how strongly tissue absorbs and scatters photons at a given wavelength. Visible light barely penetrates a few hundred microns. NIR-I extends this to about 1 to 2 millimeters of usable signal. NIR-II in vivo imaging reaches 5 to 10 millimeters in soft tissue, depending on the organ. Bioluminescence emits across multiple wavelengths and is depth-limited primarily by absorption, not scattering.
Tissue scattering drops with wavelength. At 1500 nm, scattering is roughly an order of magnitude lower than at 700 nm. This is the core physical advantage of NIR-II imaging: signals stay coherent through deeper tissue, producing sharper images with higher signal-to-noise ratios.
Autofluorescence describes the background signal produced by tissue components such as collagen, NADH, and porphyrins. It peaks in the visible spectrum, decreases in NIR-I, and becomes negligible in NIR-II. For experiments requiring quantification of weak signals, this is decisive. Bioluminescence has zero autofluorescence by definition, since no excitation light is involved.
Molecular sensitivity refers to the minimum number of cells or molecules detectable above background. Bioluminescence is the most sensitive modality available, capable of detecting fewer than 100 cells in some contexts. Fluorescence requires more abundant signal but offers spatial resolution and multiplexing capabilities that bioluminescence cannot match.
NIR-I vs NIR-II: when the spectral window changes everything
The shift from NIR-I vs NIR-II is not incremental. Moving from 700 to 900 nm into the 1000 to 1700 nm window changes what is biologically observable.
In NIR-I, photons interact with hemoglobin and water in ways that produce significant attenuation past a few millimeters. NIR-II sits in a region where the optical properties of biological tissue change favorably: water absorption stays moderate, but scattering drops sharply. The result is that signals from deep organs, vasculature in fatty tissue, and tumors located beyond the surface become detectable with the kind of resolution you would normally lose at this depth. Practically, this means a NIR-II image of a mouse abdomen reveals capillary structures and organ boundaries that NIR-I produces only as diffuse, blurred regions.
The physics behind the shift comes down to two effects. First, Rayleigh scattering scales with the inverse fourth power of wavelength, so doubling the wavelength reduces scattering by a factor of 16. Second, autofluorescence from biological molecules drops dramatically beyond 1000 nm, since most endogenous fluorophores have emission peaks in the visible or NIR-I window. The combination produces what NIR-II practitioners describe as a "quiet" background, where weak signals stand out cleanly against a near-zero baseline. The optical scattering properties of biological tissues in the NIR-II imaging window allow micrometer-scale features to remain visible at depth, and tissue autofluorescence drops to a level where high sensitivity detection of fluorescent dyes becomes practical even for faint reporters.
There are three preclinical contexts where NIR-II adoption has changed experimental design:
Vascular and lymphatic imaging. Real-time visualization of blood vessel architecture and lymphatic drainage in deep organs (liver, kidney, mesenteric vasculature) is achievable in NIR-II in ways that NIR-I cannot match. The reduced scattering preserves the structural detail needed to map fine capillary networks. Studies on tumor angiogenesis, ischemia-reperfusion injury, and lymphedema pathology have benefited directly from this capability.
Drug biodistribution, pharmacokinetics, and toxicity. Tracking the real-time spread of fluorescently tagged compounds through deep tissue benefits directly from NIR-II's penetration. Studies that previously required serial sacrifices and ex vivo imaging can now be run longitudinally on the same animal. The ability to follow a drug from injection to clearance in a single subject reduces inter-animal variability and improves the statistical power of pharmacokinetic models.
Translational research and clinical applications. Recent developments in NIR-II hardware, fluorescent dyes, and contrast agents now allow whole-mouse studies that once required invasive surgery. Recent imaging of the mouse brain through intact skin and skull, for example, has been reported in the NIR-II imaging window, opening new ways to evaluate drug efficacy and disease progression in neurology models. NIR-II findings also translate more cleanly into clinical applications, where deeper tissue penetration is a hard requirement. Several intraoperative imaging systems now operate in NIR-II for fluorescence-guided surgery, and preclinical NIR-II data is increasingly used to qualify candidate probes before clinical trials. Our deep dive on how NIR-II is emerging as a key tool in translational research covers how preclinical NIR-II imaging data feeds into drug development pipelines and clinical trial design.
Short wave infrared is not always necessary. For surface-level fluorescence (skin tumors, superficial transplants, ex vivo organ imaging), fluorescence imaging in vivo in the NIR-I window remains efficient and cost-effective. The decision should follow the depth of the biological question, not the perceived sophistication of the technology. NIR-II small animal imaging also requires probes that emit in the SWIR window, and this probe ecosystem, while growing rapidly, is still narrower than the well-established NIR-I catalog. Probe availability should be checked against the experimental design before committing to a SWIR workflow.
Forteams considering custom probes, our guide to developing and validating NIR-II probes covers the validation steps needed before in vivodeployment.
Bioluminescence vs Fluorescence Imaging: How to Choose the Right Reporter for Your Biological Target
The bioluminescence vs fluorescence in vivo debate often gets framed as a sensitivity question. That framing is incomplete. The two modalities answer different biological questions.
Bioluminescence imaging and the role of the enzymatic reaction
Bioluminescence uses a genetically encoded luciferase enzyme that emits light upon reaction with its substrate (D-luciferin or coelenterazine). No external excitation source is required, tissue autofluorescence drops effectively to zero, and the resulting signal is directly proportional to the enzymatic reaction, which itself reflects gene expression or cell viability. This is what makes in vivo bioluminescence such a robust readout for low-abundance targets across longitudinal studies.
This makes bioluminescence the modality of choice for:
- Cell viability assays in cancer models, where signal intensity correlates with viable cell count
- Gene expression reporting using promoter-driven luciferase constructs
- Detection of very small cell populations, including circulating tumor cells and minimal residual disease
- Longitudinal monitoring of metastatic progression at the whole-animal level
Fluorescence imaging and the use of fluorescent dyes
Fluorescence imaging in vivo uses external light to excite a fluorophore (genetically encoded protein, antibody-conjugated dye, or nanoparticle), which then re-emits at a longer wavelength. The advantages are different: spatial resolution is higher, multiplexing across multiple wavelengths is possible, and probes can be targeted to specific molecular markers without genetic modification of the host.
Fluorescence is the modality of choice when you need:
- Anatomical detail and high-resolution imaging
- Multiplexed detection of several markers simultaneously
- Targeted molecular imaging using antibody-dye conjugates or activatable probes
- Translational compatibility, since most clinical optical contrast agents are fluorescent
The most informative preclinical studies often combine both.Bioluminescence provides the sensitivity and quantification needed to detectrare cells or low-level expression, while fluorescence provides the spatialresolution and multiplexing required to characterize what is happening at thecellular level. Modern small animal imaging system platforms support both modalities in a single acquisition session, andmany in vivo imaging systems also combine optical channels with X-ray orcomputed tomography to put the molecular signal back into a precise anatomicalcontext.
A 5-step modality decision framework
Before selecting equipment or designing acquisition protocols, we recommend the following decision sequence. It works for new labs setting up an imaging core and for experienced groups expanding into a new disease area, and it scales naturally as programs move from early target validation to lead optimization for new drugs.
From biological endpoint to imaging modalities
Step 1. Define the biological endpoint. What process are you trying to measure? Cell viability, gene expression, drug distribution, vascular integrity, immune trafficking, tumor growth, or something else? The endpoint dictates what kind of signal you need to capture.
Step 2. Estimate the signal depth. Where in the animal does the process occur? Surface lesions, subcutaneous tumors, orthotopic implants in the liver or brain, or systemic distribution? Use anatomical reference points and existing literature to set the depth target.
Step 3. Match modality to depth and sensitivity needs. For surface to ~2 mm depth, NIR-I fluorescence is usually sufficient. For 2 to 10 mm, NIR-II in vivo imaging offers a real advantage. For molecular sensitivity at any depth (with the caveat that very deep signals attenuate), bioluminescence is unmatched, especially when the readout is gene function rather than anatomy. We covered the specific thresholds in detail in our article on when to use NIR-II and when not to, which includes a decision matrix based on depth, model type, and biological endpoint.
Step 4. Plan multiplexing and complementarity. Will a single modality answer the question, or do you need bioluminescence + fluorescence in the same animal? Plan acquisition order and probe compatibility upfront. Multi-modal in vivo imaging platforms reduce the friction of running combined modalities by integrating acquisition pipelines in a single workflow.
Step 5. Validate with a pilot. Run a small-scale pilot before committing to the full study. Modality assumptions that look correct on paper sometimes fail in practice due to pigmentation, hair, food autofluorescence, or unexpected tissue absorption.
Pigmentation, hair, and other variables in living animals
Pigmentation is one of the most underestimated variables. Black or brown mice (C57BL/6, BALB/c hybrids) have melanin in skin and hair that absorbs strongly in the visible and NIR-I regions. The same fluorescent probe can produce drastically different signal intensities in different mouse strains, even at identical doses. A pilot identifies this before it contaminates a large dataset.
Hair removal is the second variable that catches new users off guard. Even on light-skinned mice, hair follicles produce localized signal absorption and scattering. Most NIR-I protocols require depilation or hairless strain selection. SWIR imaging is more forgiving here, since the longer wavelengths penetrate hair with less attenuation, but the effect is still measurable.
This framework is intentionally simple. The most common mistake we see in preclinical imaging is starting with the equipment available rather than the biological endpoint. Modality decisions made in reverse rarely produce optimal data.
Choosing an in vivo imaging system: technical criteria that matter
Once the modality is set, the next question is: what does an in vivo small animal imaging system actually need to deliver? The market is full of platforms with overlapping specifications. The criteria below cut through the noise and focus on what actually impacts data quality across in vivo imaging techniques, from in vivo fluorescence imaging at the surface to deep NIR-II acquisition.
Beyond raw specifications, three operational factors influence day-to-day performance:
Sensors, image formation, and software for vivo detection
Sensor sensitivity. This is the single most underestimated criterion. The detector technology used (the CCD, CMOS, or InGaAs camera for NIR-II) determines the lowest signal you can quantify reliably. Quantum efficiency at your target wavelengths matters more than headline pixel count, because it dictates how cleanly photons are converted into the digital signal that drives image formation. For demanding biomedical research workflows, this is the criterion that ultimately defines whether in vivo detection of low-abundance targets is feasible at all. For NIR-II small animal imaging in particular, the detector choice has outsized impact on data quality, as we cover in detail in why the choice of detectors defines your NIR-II performance.
Acquisition speed. Real-time imaging sessions for vascular dynamics, drug uptake, or fast biological events require frame rates that not all systems support. If your protocols include time-resolved measurements, verify the actual frame rate at the wavelengths you intend to use.
Anesthesia, physiology, and animal handling. A good imaging system is integrated with a complete physiology and physiological support workflow: heated stages, isoflurane delivery, ECG monitoring if needed, and standardized positioning aids. Inconsistent animal handling is a leading cause of irreproducible data.
Software and quantification. The analysis software determines whether your data is publishable. Look for native NIST-traceable calibration, ROI quantification, multi-modal overlay, and export pipelines that integrate with downstream statistical tools. Software that locks data in proprietary formats or requires manual rescaling between modalities introduces friction at the worst possible moment, when reviewers ask for clarification.
The Vilber Newton series was designed around these criteria. Modular configurations support NIR-I, NIR-II, bioluminescence, and X-ray in a single platform, with absolute NIST-traceable calibration that allows quantitative comparison across animals, days, and even instruments. With up to 12 excitation channels (440-980nm) paired with high-efficiency emission filters covering 500 to 1500 nm, the Newton 7.0 provide the spectral flexibility needed for multiplexed studies without modality switching between sessions.
Common pitfalls and how to avoid them
Three errors account for the majority of failed in vivo imaging studies we encounter:
Tissue autofluorescence, positioning, and calibration
Underestimating autofluorescence. Standard rodent chow is rich in chlorophyll and produces strong tissue autofluorescence in NIR-I, particularly in the 600 to 700 nm range. The signal originates in the gut and persists for several hours after feeding. Switching to alfalfa-free diet two weeks before imaging eliminates this background entirely. For NIR-II studies, the issue largely disappears, which is one more reason the modality is gaining ground in fluorescence-heavy disciplines. Beyond food, skin pigments and certain components of bedding material can also contribute. Therefore, if you are quantifying weak signals in NIR-I, control for autofluorescence with a non-injected animal in every cohort.
Inconsistent positioning. Quantification depends on reproducible animal positioning. Small variations in posture, distance to the lens, or head orientation introduce variability that is often misattributed to biological variance. The fix is mechanical: standardized Pads, fixed reference points on the imaging stage, and a written positioning protocol that all operators follow.
Skipping calibration. Without absolute calibration, signal intensities cannot be compared across days or instruments. NIST-traceable calibration is the only reliable way to produce longitudinal data that holds up to peer review. Internal calibration standards that drift over time create artifactual trends that look like biological effects. This pitfall is particularly damaging in long-term oncology studies, where comparing tumor signals across weeks of imaging is the central output. The investment in proper calibration upfront prevents the much larger investment of repeating an entire study after a reviewer flags the data.
A fourth issue worth flagging is single-modality bias. Once a lab adopts a modality, there is a strong tendency to use it for every experiment, even when a different modality would answer the biological question more directly. Resist this. The modality decision should be re-evaluated for each new biological endpoint, not inherited from the last successful study.
In Vivo Imaging Specialist & Global Sales Director
