ICG Fluorescence Imaging in Minimally Invasive Surgery

When a surgeon switches from white-light imaging to fluorescence mode, the system interrogates a different optical signal. Excitation reaches the tissue, ICG emits near-infrared radiation, the returned light travels through the scope, and spectral separation helps the detector distinguish emission from unwanted light. Processing then turns detector measurements into a visible image.

The useful engineering question is where information is gained, attenuated or altered along that journey. This article follows the chain from illumination to display, with particular attention to rigid optics. Its evaluation suggestions are engineering discussion points, not measured opentic specifications or clinical acceptance criteria.

What ICG fluorescence adds to white-light imaging

White-light endoscopy principally presents anatomical reflectance in the visible spectrum. Fluorescence introduces a signal associated with a fluorophore and its distribution in a particular biological setting. The two images answer different questions: reflected light supplies structural context, while fluorescence can make an application-specific distribution easier to visualize. Interpreting that distribution still requires the procedure, acquisition conditions and clinical context.

ICG is a near-infrared fluorophore. Typical excitation lies in the high-700 to low-800-nm region, with emission commonly detected at longer wavelengths in the low-800-nm region. Excitation and emission occupy overlapping spectral regions rather than two infinitesimally narrow lines; emission is generally shifted toward longer wavelengths. Published photophysical work also shows that the molecular environment and aggregation affect fluorescence. Solvent, protein binding, concentration and measurement conditions therefore belong beside any quoted spectral peak. A single wavelength pair is an inadequate specification for an entire imaging chain. 5

Green on the monitor is a display choice, not the colour of the emitted fluorescence. The NIR fluorescence signal is invisible to the human eye. The processor can assign green, blue or other pseudocolours, or grayscale, alone or overlaid on visible anatomy. A legend and an understanding of the selected mode are more informative than the hue itself. In a conceptual demonstration, changing the lookup table could make the same recorded signal look quite different without changing the photons originally detected.

For technical discussions, separate three objects: the physical emission, the camera data and the rendered view. Ask which one a screenshot or specification describes. An attractive overlay may be useful for communication, but it does not supply the acquisition settings, calibration or physiological interpretation needed to turn that view into a measurement.

From excitation to display: the complete fluorescence imaging chain

The source supplies excitation energy; its spectral conditioning and delivery optics determine what reaches the object. Tissue containing ICG then becomes the source of fluorescence collected by the imaging optics. The return path must preserve usable emission while rejecting enough unwanted illumination to make detection meaningful. Fluorescence filter design addresses this separation between excitation and emission. 7

SYSTEM-LEVEL IMAGING CHAIN

1. Excitation sourceGenerate illumination
2. Excitation spectral filteringCondition the illumination band
3. Light deliveryCarry excitation to the field
4. Tissue + ICGInteraction with the fluorophore
5. Fluorescence emissionReturned NIR signal
6. RIGID OPTICAL SCOPECollect and relay the optical image
7. Spectral separationSuppress unwanted wavelengths
8. Detector / cameraConvert light into measurements
9. Image processingConstruct the displayed representation
10. DisplayPresent fluorescence and anatomical context

Spectral separation may be implemented in a camera head, coupler, filter assembly or another dedicated optical module. The diagram is a functional sequence, not an assembly drawing: it does not place a filter inside every rigid scope. Because reflected excitation can overwhelm a weak emission signal, a useful design review asks both what the filter passes and what it rejects, at the angles and conditions of use.

Architecture changes the integration problem. A shared optical path and sensor can acquire channels sequentially; separate detectors can acquire visible and NIR information through different spectral branches. Simultaneous and sequential acquisition place different demands on signal balance, motion handling and registration. These are design alternatives, not a claim that every commercial system uses a particular arrangement.

A practical review should follow one configuration through the whole chain. Identify the excitation mode, return-light path, detector channel and processing mode before comparing images. When a component changes, record which downstream settings change with it. This makes it possible to distinguish an optical improvement from a new exposure setting or display map, instead of assigning every visible difference to the scope.

The optical role of the rigid endoscope

The rigid endoscope collects and relays an image through a constrained mechanical envelope. For fluorescence integration, the proposed optical review should cover visible transmission, NIR spectral transmission, resolution, distortion, stray light, ghosting, chromatic correction and field consistency. Mechanical alignment and the connection to the camera complete that review; a central sharp image alone leaves much of the field unexplored.

Visible-light image quality and NIR fluorescence transmission are related but distinct requirements. The material stack—including glass, coatings, bonding materials and interfaces—needs characterization over the intended spectral regions. A white-light image cannot provide a missing NIR transmission curve. Equally, a transmission number without wavelength, geometry and measurement conditions leaves the buyer unable to determine what was actually measured.

Consider an engineering comparison between two scopes with similar central white-light sharpness. The proposed test would keep the source, target, coupler and camera fixed, then examine NIR throughput and image behaviour across the field. If the fluorescence views differ, follow-up measurements should investigate transmission, focus, illumination and processing separately. That example is a troubleshooting method, not evidence that either design performs better.

Scope transmission contributes to signal delivery, but system detectability also involves the fluorophore, excitation, tissue, distance, filters, detector and acquisition settings. The scope specification should describe the scope measurement; a system sensitivity claim needs a system experiment. TG 311 treats fluorescence performance through multiple characterization tasks rather than one component label. 3

For procurement, request a traceable distinction between component evidence and assembled-system evidence. A useful scope report identifies the tested specimen, configuration, wavelength range, field positions and limitations. A useful integration report identifies what happened after that scope was connected to the intended platform. Keeping those records separate makes later substitutions and failure investigations substantially easier to interpret.

Visible-to-NIR parfocality and spectral performance

Visible-to-NIR parfocality concerns the axial relationship between best-focus conditions in the two spectral regions. Parfocal optical design aims to minimise visible-to-NIR axial focus shift, bringing best-focus positions sufficiently close to a common image plane for the intended system. It does not mean the physical path lengths must be identical.

Refractive index varies with wavelength. Glass dispersion and the way an optical design corrects chromatic aberration therefore influence focus behaviour across a broad spectrum. The complete arrangement of powered elements matters; saying only that NIR has a longer wavelength does not explain a particular focus shift. General imaging guidance describes wavelength-dependent performance and the importance of designing for the operating spectrum. 6

VISIBLE / NIR FOCUS RELATIONSHIP

A · Separated best-focus planes

VisibleNIR

Axial separation may require a focus adjustment between modes.

B · Parfocal design target

VisibleNIR

Bring best-focus planes closer for the intended imaging configuration.

For an engineering evaluation, define the visible and NIR bands first. Use a documented target, working distance, coupler and focus procedure, then compare the focus settings or axial positions associated with the chosen sharpness criterion. Record field position as well as centre performance. The proposed method should state whether it measures the scope alone or the complete camera-coupler-scope assembly.

A meaningful specification states the acceptance basis. Final parfocal performance also depends on the coupler, camera optics, detector architecture and system configuration; scope correction alone cannot guarantee the system result. How much residual shift is usable depends on imaging requirements and depth of focus. A photograph of two apparently sharp modes does not establish a numerical tolerance. The diagram illustrates a design direction, not zero shift, perfect focus coincidence or a promise that refocusing is never required.

In use, a sufficiently large mismatch can present as loss of sharpness or the need to adjust focus during a mode change. That is an optical and workflow issue. This article makes no inference from improved focus management to clinical outcomes. Parfocality is a focus-management problem across spectral regions, not a marketing synonym for fluorescence compatibility.

Why fluorescence brightness is not a standalone performance metric

A bright fluorescence image is not the same thing as a sensitive fluorescence system. Display intensity sits at the end of a chain containing optical losses, electronic acquisition and image mapping. The following engineering checklist identifies variables to hold constant or record when investigating an apparent brightness difference; it is not a calibrated model of their relative contributions.

What changes the fluorescence signal?

Factor	Where it occurs	Effect on displayed signal
ICG concentration	Fluorophore environment	Response requires characterization; a linear relationship cannot be presumed throughout the range.
Tissue depth	Excitation and return paths	Attenuation and scattering change the signal reaching the scope.
Excitation irradiance	Illumination at the target	Changes excitation conditions; document the delivered illumination.
Working distance	Object and collection geometry	Changes illumination and collection conditions, often together.
Scope spectral transmission	Optical relay	Changes the returned spectral signal delivered downstream.
Emission filter	Detection path	Passband and rejection affect desired signal and unwanted leakage.
Detector sensitivity	Sensor	Spectral response and noise influence measurable signal.
Exposure / gain	Acquisition electronics	Changes recorded level and noise behaviour; saturation can hide differences.
Image processing	Processing and display	Mapping, thresholds and normalization alter the rendered appearance.

ICG fluorescence can show concentration-dependent aggregation and quenching effects; its environment also affects response. Intensity therefore requires a validated quantitative relationship before it is used to infer concentration. The cited photophysical study is evidence about fluorescence behaviour, not a basis for a clinical administration protocol. 5

SNR describes signal relative to noise; SBR compares a selected signal region with a selected background. A report should specify the measurement method, equation, regions of interest (ROIs), background selection and subtraction, exposure, gain and processing stage. Different definitions or background choices can produce different numbers from the same scene. Neither ratio is a direct statement of clinical accuracy. Performance-method literature provides experimental approaches to several such characterization questions. 4

Depth-related detectability is similarly conditional. Tissue optical properties, fluorophore concentration and distribution, geometry, excitation conditions and system sensitivity all affect the observed result. NIR fluorescence is useful for surface and near-surface information; it is not unrestricted deep-tissue imaging comparable to CT or MRI. A universal penetration depth would conceal the conditions needed to interpret the measurement. 3

For a comparison, retain both the acquisition record and the displayed output. Lock settings where possible; document automatic controls where locking is unavailable. Include a repeat capture after repositioning to reveal setup dependence. If only a processed screenshot is available, describe it as an illustration of that display mode rather than a measurement of transmission, perfusion or detection limit.

Clinical applications: evidence, labeling and context

The clinical question determines what a fluorescence pattern might mean. Vascular distribution, biliary visualization and lymphatic mapping involve different biological contexts. A claim should identify the procedure and the evidence supporting it before attaching a benefit to the technology.

For example, current U.S. IC-GREEN labeling provides a jurisdiction-specific reference; labeling varies by product and jurisdiction. Its January 2026 revision includes defined uses for visualizing vessels, blood flow and tissue perfusion; extrahepatic biliary ducts; and lymph nodes and lymphatic vessels during lymphatic mapping in adults with cervical and uterine cancer. Population restrictions differ across indications, and specified fluorescence-device requirements apply. These are boundaries of that label, not worldwide indications or approval of any opentic device. 1

The EAES consensus organizes evidence by abdominal surgical application, including cholecystectomy, colorectal perfusion assessment and selected lymphatic mapping contexts. Recommendation strength and evidence quality vary by application. A consensus recommendation is distinct from drug labeling and cannot establish authorization for a particular scope or assembled platform. Its publication date also does not make every included study contemporary. 2

Clinical use should be described at the right evidence level

Context	What can be said	What should not be assumed
Regulatory labeling	Describe the named label, jurisdiction, population and device conditions.	A global indication or permission for every system configuration.
Consensus guidance	Identify the procedure and recommendation with its evidence context.	Uniform certainty or regulatory authorization across all applications.
Clinical study	Report the actual population, comparator and endpoint studied.	Transfer of outcomes to another procedure or device without justification.
Engineering validation	Describe measured optical or system behaviour under stated conditions.	Proof of clinical benefit, complication prevention or diagnostic accuracy.
Marketing statement	Match the exact claim to approved supporting evidence.	That a compelling image itself substantiates the claim.

For thoracic discussions, use a separate procedure-specific evidence review. This article does not extrapolate the abdominal EAES consensus to thoracoscopic segment identification, perfusion assessment or lymphatic workflows. Any such proposed use needs its own study context, applicable labeling and platform documentation before a clinical statement is made.

ICG should not be presented as a universal tumor-specific agent. A signal in an oncological procedure may concern lymphatic drainage or another distribution process; it is not automatically a cancer diagnosis. The defined lymphatic use in the U.S. label illustrates why the actual visualization task matters. 1

Administration is a clinical decision governed by applicable labeling, contraindications, warnings and institutional protocols. The U.S. contraindication concerns prior hypersensitivity to indocyanine green; anaphylaxis has been reported. Sodium iodide content and interference with thyroid radioactive iodine uptake studies are separate label considerations, not a blanket “iodine allergy” contraindication. 1 This article discusses imaging technology and evidence context. It provides no instructions for ICG administration or clinical decision-making.

System integration and compatibility

A mechanically compatible scope is not necessarily spectrally compatible. Start an integration review with four separate questions: does the connection fit, does the optical combination form the intended image, are the spectral characteristics suitable, and has the fluorescence configuration been evaluated? A positive answer to the first leaves the other two open.

The proposed configuration record should identify the scope, camera head, coupler, illumination source, cable, relevant filters, processor and software mode. Record operating distance and focus procedure alongside the hardware. This gives an OEM or distributor a reproducible description rather than an open-ended compatibility statement attached to a connector type.

Review illumination and detection together. The source band must be considered alongside the intended fluorophore and the return-path passband. Next examine visible and NIR transmission, focus, field coverage and detector response. Finally examine what processing does to the acquired data. This ordering helps separate spectral mismatch from poor focus or a display setting.

When visible and NIR images are overlaid, assess spatial correspondence across the field. A separate-detector design and a shared-path sequential design need architecture-appropriate checks; they need not use the same registration procedure. For a proposed moving-target test, document acquisition timing as well as spatial alignment so that a temporal discrepancy is not mistaken for a static optical offset.

Compatibility should be established against a defined system configuration. If a camera, coupler or software mode is substituted, identify which parts of the earlier evidence remain applicable and which tests need repeating. That is a change-control recommendation, not a universal validation protocol. This article supplies no opentic compatibility list or performance guarantee.

How to evaluate an ICG fluorescence system

Begin with the intended imaging task, then select measurements that can answer it. AAPM TG 311 offers professional guidance for performance evaluation, not a mandatory international pass/fail standard. Kanniyappan and colleagues describe test methods addressing properties such as uniformity, spatial resolution, field of view, depth of field, crosstalk, sensitivity and linearity. 3 4

The matrix below is a proposed review framework. Its purpose is to separate component contributions from complete-system behaviour; it specifies neither opentic thresholds nor universal acceptance limits.

ICG fluorescence evaluation framework

Property	Scope contribution	System-level influence	Common interpretation error
Visible transmission	Visible relay throughput	Source, coupling and detector response	Treating white-light brightness as an NIR test.
NIR transmission	Spectral throughput	Filtering and sensor response	Using one spectral number as system sensitivity.
Parfocality	Chromatic focus behaviour	Coupler, sensor position and focus criterion	Assuming fluorescence labeling guarantees identical focus.
Resolution	Optical image transfer	Sampling, processing and target geometry	Equating display pixel count with resolved detail.
Fluorescence brightness	Returned signal delivery	Exposure, gain and display mapping	Reading a screenshot as calibrated concentration.
Sensitivity	Optical losses contribute	Excitation, noise and detection criterion	Attributing a complete-system limit to the scope alone.
Uniformity	Field-dependent behaviour	Illumination, vignetting and correction	Inspecting only the central image.
Spectral crosstalk	Spectral relay characteristics	Filter rejection and channel separation	Treating leaked illumination as fluorescence.
SNR / SBR	Signal and stray-light contributions	Noise, background definition and processing	Calling a ratio clinical accuracy.
Registration	Geometry and distortion	Channel architecture, calibration and timing	Assuming every platform needs the same correction.

Sensitivity needs a defined measurement method and detection criterion under stated test conditions. Before accepting a demonstration, ask what fluorophore and spectral range it targets, how excitation is delivered and where separation occurs. Request the visible/NIR characterization and focus method. Establish how exposure and gain were controlled, how sensitivity was defined and whether the displayed information is qualitative or quantitatively validated.

Then review reproducibility. A useful engineering record includes configuration identifiers, test conditions, analysis method, repetitions and limitations. Keep raw or minimally processed data where the system permits, alongside representative display captures. Document any automatic normalization that could conceal variation between captures. For a product change, use this record to decide which comparison remains valid rather than relying on visual memory.

The decision to accept an integration belongs to a defined requirement and review process. Optical characterization supports that decision; clinical evidence and regulatory assessment answer different questions. Keeping those evidence streams connected without merging them produces a more useful technical discussion than asking whether a fluorescence image simply looks impressive.

Key takeaways

• Follow the complete excitation-to-display chain when interpreting fluorescence performance.
• Evaluate visible/NIR transmission and focus separately from connector fit and display brightness.
• Define acquisition conditions and analysis methods before comparing systems.
• Match clinical statements to a specific label, procedure and evidence source.
• Treat the diagrams and evaluation framework here as educational engineering material, not product test results.

References

Sources checked 6 October 2026. Label information is a U.S. example; applicability elsewhere requires separate review.

1. DailyMed. IC-GREEN — indocyanine green prescribing information. Revised January 2026. Indications, population/device conditions and safety sections.

2. Cassinotti E, et al. European Association for Endoscopic Surgery (EAES) consensus on Indocyanine Green (ICG) fluorescence-guided surgery. Surgical Endoscopy. 2023;37(3):1629–1648. DOI: 10.1007/s00464-023-09928-5. Author-hosted full text.

3. Pogue BW, et al. AAPM Task Group Report 311: Guidance for performance evaluation of fluorescence-guided surgery systems. Medical Physics. 2024;51:740–771. DOI: 10.1002/mp.16849. University-hosted full text.

4. Kanniyappan U, et al. Performance test methods for near-infrared fluorescence imaging. Medical Physics. 2020;47(8):3389–3401. DOI: 10.1002/mp.14189.

5. Kraft JC, Ho RJY. Interactions of indocyanine green and lipid in enhancing near-infrared fluorescence properties: the basis for near-infrared imaging in vivo. Biochemistry. 2014;53(8):1275–1283. DOI: 10.1021/bi500021j. Photophysical study; not an administration guide.

6. Edmund Optics. Wavelength Effects on Performance. Technical application note on wavelength-dependent imaging behaviour.

7. Edmund Optics. Introduction to Fluorescence Filters. Technical application note on spectral separation.