Intraoperative fluorescence-guided surgery — ICG and 5-ALA under near-infrared cameras reveal tumor margins invisible to the naked eye
Fluorescence-guided surgery begins well before the incision: a fluorescent agent must be delivered and given time to accumulate selectively in tumor tissue relative to surrounding normal tissue. The pharmacokinetics of the chosen agent — intravenous indocyanine green, oral 5-aminolevulinic acid, or an emerging tumor-targeted molecular probe — dictate the dose, timing, and route, and directly determine how sharp the eventual tumor-to-background contrast will be in the operating room.
Indocyanine green (ICG) is an FDA-approved tricarbocyanine dye first used in the 1950s for cardiac output and hepatic function testing. Injected intravenously, it binds plasma proteins (mainly albumin) almost immediately, is confined to the vascular compartment, and is cleared exclusively by hepatocytes into bile with no renal excretion or systemic toxicity at clinical doses — a safety profile that has made it the most widely used intraoperative fluorophore worldwide.
ICG accumulates in tumors through the Enhanced Permeability and Retention (EPR) effect: tumor neovasculature is leaky and disorganized, with poor lymphatic drainage, so protein-bound dye extravasates into the tumor interstitium and lingers there for hours while it washes out of normal, tightly-junctioned vasculature. This makes ICG useful for a wide range of applications — sentinel lymph node mapping in breast cancer and melanoma, perfusion assessment in bowel anastomoses, biliary anatomy delineation (cholangiography), and liver tumor margin visualization, where the surrounding cirrhotic parenchyma clears dye faster than the tumor, leaving a fluorescent rim or nodule.
Dosing and timing vary by application: for liver tumor imaging, 0.5 mg/kg is typically given 24–48 hours preoperatively to allow background hepatic clearance while tumor retention persists; for lymphatic mapping, a peritumoral injection of 5 mg/mL ICG is given minutes before incision, tracking drainage in real time. Because ICG lacks a true molecular target, its selectivity is entirely a function of tissue perfusion and clearance kinetics rather than receptor biology.
5-aminolevulinic acid (5-ALA, marketed as Gliolan) works through an entirely different mechanism: it is a metabolic precursor in the heme biosynthesis pathway. Given orally, 5-ALA is taken up and converted intracellularly to protoporphyrin IX (PpIX), a naturally fluorescent molecule that emits red light (~635 nm) under blue-violet excitation (~400–410 nm). In malignant glioma cells, disrupted heme biosynthesis and altered ferrochelatase/ABC-transporter activity cause PpIX to accumulate to far higher concentrations than in surrounding normal brain, producing a striking visual contrast between bright red tumor and dark blue-violet normal tissue under a modified surgical microscope.
Unlike ICG, 5-ALA fluorescence reflects tumor cell metabolism directly rather than vascular leakage, making it particularly effective for infiltrative high-grade gliomas whose margins are otherwise indistinguishable from edematous brain on white light alone. Tumor-targeted molecular probes represent a third category: engineered fluorophore-ligand conjugates such as pafolacianine (folate-FITC, OTL38), approved for ovarian cancer, bind folate receptor-alpha overexpressed on tumor cells, giving true molecular specificity independent of perfusion or metabolism — a design strategy expected to expand rapidly as more receptor-targeted probes reach clinical approval.
Once fluorophore has accumulated in tissue, a specialized camera system must excite it at the correct wavelength, reject the much brighter excitation light, and capture the comparatively faint emitted fluorescence — all while overlaying that signal onto a normal white-light surgical view in real time, at full video frame rates, without disrupting the surgeon's natural view of the field.
Near-infrared light in the 700–900 nm range (NIR-I, or the "first optical window") penetrates several millimeters to centimeters into tissue with minimal absorption by hemoglobin and water and negligible tissue autofluorescence — properties that make it ideal for imaging fluorophores like ICG through blood, fat, and several millimeters of overlying tissue. Commercial platforms exploit this window with dedicated laser or LED excitation sources tuned to ~805 nm paired with high-sensitivity CCD or CMOS sensors filtered to reject the excitation wavelength and pass only the red-shifted emission at ~830 nm.
Leading clinical systems include Novadaq/Stryker's SPY-PHI and SPY Elite (open and laparoscopic vascular/perfusion imaging), Intuitive Surgical's Firefly integrated into the da Vinci robotic platform, and Medtronic's PINPOINT endoscopic fluorescence system — each fusing a standard white-light color channel with a monochrome NIR fluorescence channel, typically pseudo-colored green for visual salience, and composited into a single live video stream displayed on the surgeon's monitor or robotic console. A second-generation NIR-II window (1000–1700 nm) is under active investigation because longer wavelengths scatter even less, promising deeper tissue penetration and sharper margins, though clinically approved NIR-II fluorophores remain limited.
For 5-ALA, detection instead uses a modified operating microscope (e.g., Zeiss BLUE400 module) with a blue-violet excitation light source and a long-pass barrier filter that blocks the excitation wavelength while transmitting the red PpIX emission, switchable by foot pedal so the surgeon alternates between conventional white light and fluorescence mode throughout resection.
Modern systems do not simply replace the white-light image with a fluorescence image — they computationally fuse both channels at full video rate so the surgeon retains normal anatomic orientation while fluorescence appears as a colored overlay only where signal exceeds a threshold. This requires the NIR and color sensors to be co-registered and synchronized, with processing latency kept under roughly 100 ms to avoid perceptible lag during dissection.
Camera sensitivity is critical: clinical-grade systems can detect ICG concentrations down to roughly 0.1 nanomolar under standard operating room lighting, and dynamic range/gain settings are adjustable so faint residual signal at a resection margin remains visible without saturating the bright signal from the main tumor mass. Newer platforms add quantitative fluorescence — reporting a numeric intensity or tumor-to-background ratio rather than a purely qualitative glow — moving the technology from subjective visualization toward reproducible, threshold-based intraoperative decision-making.
This is the operative heart of fluorescence-guided surgery: as the surgeon dissects, the fused white-light/fluorescence display continuously shows which tissue is glowing and which is not, allowing the resection plane to be adjusted moment to moment rather than relying solely on preoperative imaging, palpation, or visual inspection that cannot reliably distinguish tumor from adjacent inflamed or edematous tissue.
In practice, the surgeon works from a monitor or eyepiece showing the surgical field in natural color with a translucent fluorescence overlay — commonly rendered in green for ICG systems or as native red for 5-ALA/PpIX under the microscope. Bright, well-demarcated fluorescence corresponds to tumor tissue with intact vascular leakiness (ICG) or elevated porphyrin metabolism (5-ALA); dim or absent signal corresponds to normal parenchyma. As dissection proceeds along the tumor capsule or infiltrative edge, the surgeon watches the fluorescence boundary in real time, extending resection wherever glow persists and stopping where it fades.
The practical value is greatest at anatomically ambiguous margins — the interface between tumor and a major vessel, bile duct, or eloquent brain tissue — where the fluorescence signal can reveal that disease extends further than the visually apparent tumor edge, or conversely that healthy critical structures are free of tumor and can be preserved. In hepatobiliary surgery, ICG cholangiography during liver resection routinely detects an additional rim of tumor or confirms clear bile duct margins in cases where white-light inspection alone would have been ambiguous.
Quantitative tumor-to-background ratio (TBR) thresholds are increasingly used to standardize the subjective "glow" into an objective decision rule: a TBR above roughly 2.0 is generally taken as a positive fluorescence signal warranting continued resection, while ratios approaching background are treated as a stopping point, subject always to the surgeon's anatomic judgment and frozen-section confirmation where available.
Once the visible tumor mass has been removed, the surgeon sweeps the walls of the resection cavity under fluorescence — a step that would be impossible to perform reliably with the naked eye. Small foci of residual tumor, often just a few millimeters across, retain fluorescent signal at the cavity edge and prompt additional targeted resection before the case is closed, directly reducing the rate of positive margins found later on final pathology.
After bulk resection, the cavity wall is systematically inspected under fluorescence mode. Any lingering bright focus is biopsied or directly re-resected, and the process is repeated until the cavity is fluorescence-negative or until further resection would risk unacceptable damage to adjacent functional tissue. In 5-ALA glioma surgery, this step is credited with much of the technique's benefit: infiltrative tumor cells beyond the visibly abnormal, contrast-enhancing region on preoperative MRI are frequently PpIX-positive and would otherwise be left behind.
Fluorescence at the margin is not perfectly specific, however. Reactive astrogliosis, acute inflammation, blood products, and areas of increased vascular permeability from surgical trauma itself can all produce a false-positive glow that mimics residual tumor, particularly with 5-ALA in a category surgeons term "vague" or low-intensity fluorescence as opposed to solid, bright signal. This is why fluorescence-guided margin assessment is used as a real-time adjunct that prompts additional biopsy or resection — not as a standalone diagnostic — and why intraoperative frozen section or neuronavigation-correlated sampling remains the confirmatory step whenever residual signal is ambiguous. Reported specificity for true residual tumor at the cavity margin ranges roughly 75–85% across series, improving further when combined with quantitative TBR thresholds rather than qualitative visual grading alone.
The clinical case for fluorescence-guided surgery ultimately rests on outcome data: does glowing tumor at the margin translate into measurably better surgery? Across randomized trials and large cohort studies spanning neurosurgery, hepatobiliary surgery, and gynecologic oncology, fluorescence guidance consistently raises the rate of complete resection, lowers the rate of positive margins, and — in the best-studied setting, malignant glioma — improves progression-free survival compared with conventional white-light resection alone.
The pivotal evidence for fluorescence-guided resection is the multicenter randomized controlled trial by Stummer and colleagues (Lancet Oncology, 2006), which randomized 322 patients with suspected malignant glioma to resection under 5-ALA fluorescence guidance versus conventional white-light microsurgery. The fluorescence-guided arm achieved complete resection of contrast-enhancing tumor in 65% of patients versus 36% in the white-light arm — a near doubling of the complete resection rate — and 6-month progression-free survival was 41% in the 5-ALA group versus 21% in the control group. This trial led directly to regulatory approval of 5-ALA (Gliolan) for high-grade glioma surgery in Europe and later the United States, and it remains the reference standard demonstrating that a fluorescence-guided technique can change a hard oncologic endpoint, not merely intraoperative visualization.
Subsequent studies across other tumor types have reinforced the pattern, if with smaller effect sizes: ICG-guided hepatectomy for liver tumors and lung nodules improves detection of subcentimeter or subsurface lesions missed by palpation and imaging, sentinel lymph node mapping with ICG reduces the need for more invasive axillary or inguinal dissection while maintaining nodal staging accuracy, and pafolacianine-guided ovarian cancer surgery increased detection of additional malignant tissue not visualized under white light or palpation in a pivotal Phase III trial, supporting its 2021 FDA approval. Reoperation rates for positive margins found on final pathology fall meaningfully across these series when fluorescence guidance is used, translating into fewer repeat surgeries, shorter cumulative recovery time, and — in glioma specifically — a demonstrated survival benefit that has made fluorescence guidance part of standard neuro-oncological practice at most high-volume centers.
In the Stummer et al. 2006 randomized trial (Lancet Oncology), 5-ALA fluorescence-guided resection achieved complete removal of contrast-enhancing glioma in 65% of patients versus just 36% with conventional white-light surgery — nearly doubling the complete resection rate and translating into significantly longer 6-month progression-free survival (41% vs 21%). This single trial remains the strongest causal evidence that intraoperative fluorescence guidance changes hard oncologic outcomes, not just visualization.