A 1,389-Patient Study Challenges Assumptions About Fluorescent Dyes in Glioblastoma Surgery
For surgeons operating on glioblastoma, the boundary between malignant tissue and functioning brain can be almost impossible to see under ordinary white light. A large European study now suggests that one of the dyes used to illuminate that boundary may be associated with more complete removal of the tumor visible on contrast-enhanced magnetic resonance imaging than another widely used compound. In a retrospective analysis of 1,389 patients treated at 12 neurosurgical centers, sodium fluorescein was linked to higher rates of contrast-enhancement-defined gross total resection than 5-aminolevulinic acid, or 5-ALA. The finding could reignite debate over which fluorescent technique best guides glioblastoma surgery, but the study’s statistical analyses also deliver a crucial warning: differences between hospitals, surgeons and patient populations may explain at least part of the apparent advantage.
Glioblastoma is the most aggressive primary malignant brain tumor in adults. Its cells infiltrate surrounding tissue, sending microscopic extensions beyond the solid mass visible on routine imaging. Surgery therefore has two competing goals: remove as much tumor as possible while preserving brain regions responsible for movement, speech, vision, memory and other essential functions. The extent of resection matters because residual malignant tissue can fuel recurrence, yet the tumor’s invasive margins rarely form a clean anatomical border. Fluorescence-guided surgery attempts to make this difficult judgment more visible in real time, allowing surgeons to distinguish tumor-associated tissue from normal brain while working through a microscope.
The two dyes examined in the study illuminate different biological phenomena. Sodium fluorescein is a yellow fluorescent compound that remains largely within the bloodstream and can leak into tissue where the blood–brain barrier has been disrupted. Because high-grade gliomas often contain abnormal, permeable blood vessels, fluorescein can accumulate in regions that also appear bright after gadolinium contrast on magnetic resonance imaging. Under an appropriate microscope filter, the dye can make these areas glow yellow-green. Its signal therefore tends to reflect vascular permeability and contrast enhancement, rather than the direct presence of every infiltrating tumor cell.
5-ALA works through a different pathway. After oral administration, the compound enters the heme biosynthesis pathway, a normal cellular process used to produce hemoglobin. Tumor cells can accumulate the downstream metabolite protoporphyrin IX, which fluoresces a vivid pink-red under blue-violet excitation. In glioblastoma surgery, this signal can reveal malignant tissue that may not be obvious under white light. However, fluorescence intensity varies with tumor biology, cellular density, metabolism, oxygenation and the timing of administration. Areas of infiltrating tumor may produce weak or absent fluorescence, while tissue that is visibly enhanced on MRI does not always correspond perfectly to the strongest 5-ALA signal.
The investigators analyzed prospectively maintained clinical databases from 12 European neurosurgical centers, focusing on consecutive patients with locally documented IDH-wildtype glioblastoma who underwent fluorescence-guided surgery. Of the total cohort, 650 patients received sodium fluorescein alone, 457 received 5-ALA alone and 282 underwent surgery using both fluorescent agents. The primary outcome was contrast-enhancement-defined gross total resection, or CE-GTR, meaning that postoperative imaging showed no remaining tumor within the contrast-enhancing compartment. The researchers also assessed the contrast-enhancing extent of resection, postoperative Karnofsky Performance Status—a scale of functional independence—and overall survival.
The raw results appeared to favor fluorescein decisively. CE-GTR was achieved in 74.9 percent of patients in the sodium fluorescein group, compared with 62.6 percent in the 5-ALA group and 68.4 percent among patients receiving both dyes. The overall difference between the three groups was statistically significant, with a probability value below 0.001. In a conventional multivariable analysis that adjusted for measured patient and tumor factors, patients treated with 5-ALA had lower odds of achieving CE-GTR than those treated with sodium fluorescein. The odds ratio was 0.63, with a 95 percent confidence interval from 0.49 to 0.81.
The association remained when the researchers used a generalized estimating equation model that accounted for the possibility that patients treated at the same institution might resemble one another. In that analysis, the odds ratio for CE-GTR with 5-ALA compared with fluorescein was 0.57, with a 95 percent confidence interval of 0.36 to 0.90. Sodium fluorescein was also associated with a greater mean contrast-enhancing extent of resection: after adjustment for clustering by treatment center and correction for multiple comparisons, the difference between fluorescein and 5-ALA was 6.7 percentage points, with a confidence interval of 3.3 to 10.0 percentage points.
But the picture changed when the analysis directly incorporated center-specific effects. A fixed-effects logistic regression model, designed to compare patients while controlling more stringently for systematic differences between hospitals, reduced the strength of the association. The odds ratio became 0.65, but the confidence interval widened to 0.37–1.12 and the result was no longer statistically significant. This attenuation suggests that institutional practice and case selection may have influenced the initial comparison. Hospitals using one dye may differ from those using another in surgical philosophy, access to intraoperative imaging, navigation systems, mapping expertise, patient referral patterns or the anatomical difficulty of the tumors they treat. A retrospective study can adjust for known variables, but it cannot fully reconstruct the treatment decisions that shaped each operation.
The survival findings were similarly cautious. Median overall survival was 18.2 months for patients receiving sodium fluorescein, 17.9 months for those receiving 5-ALA and 18.5 months for the dual-fluorescence group. In a Cox proportional-hazards model using center-clustered robust standard errors, 5-ALA was associated with a lower risk of death than fluorescein, with a hazard ratio of 0.84 and a 95 percent confidence interval of 0.74–0.97. Yet that apparent difference disappeared in a center-stratified sensitivity analysis, where the hazard ratio was 0.88 and the confidence interval extended from 0.64 to 1.22. The researchers found no survival association for dual fluorescence. These results underscore why a more complete-looking operation on postoperative MRI cannot automatically be interpreted as a longer life for patients.
A major limitation is embedded in the study’s central measurement. Contrast enhancement is an imperfect surrogate for the full extent of glioblastoma. Gadolinium enhancement generally indicates disruption of the blood–brain barrier, but infiltrating tumor cells can extend into regions that do not enhance. Conversely, some enhancing tissue may contain treatment-related changes, edema or abnormal vasculature rather than a dense population of viable malignant cells. The analysis could not evaluate non-enhancing tumor, the microscopic infiltrative margin or so-called supramaximal resection beyond the conventional enhancing boundary. It therefore cannot establish that sodium fluorescein removes more total tumor, nor that it is intrinsically superior as a biological marker.
The findings also do not overturn the established role of 5-ALA. Earlier randomized and observational studies have shown that 5-ALA fluorescence can increase the likelihood of complete removal of contrast-enhancing malignant glioma compared with white-light surgery, while subsequent research has demonstrated that fluorescent signal may extend beyond the MRI-defined margin. Sodium fluorescein, meanwhile, is relatively inexpensive and can be administered intravenously, but its signal is closely tied to vascular leakage and can be affected by dose, circulation, timing and microscope filters. Dual fluorescence is attractive because it could theoretically combine complementary information: fluorescein may highlight areas of blood–brain barrier disruption, while 5-ALA may reveal metabolically active tumor cells. In practice, interpreting two signals requires technical expertise and does not eliminate the fundamental need to protect eloquent brain.
The new multicenter analysis is consequently less a declaration of a winning dye than a reminder that fluorescent surgery is an imaging strategy, not a substitute for biological certainty. Its large sample and use of several statistical models strengthen the evidence that sodium fluorescein can be associated with greater removal of contrast-enhancing tissue in real-world practice. At the same time, the loss of statistical significance after center fixed effects highlights the danger of treating retrospective comparisons as randomized tests. Future studies will need standardized imaging protocols, detailed molecular and pathological sampling, objective fluorescence measurements and long-term functional outcomes, ideally in prospective randomized designs. For patients facing glioblastoma surgery, the practical message is more nuanced than a simple choice between yellow-green and pink-red: the value of either dye depends on how its signal is interpreted within a broader system of imaging, functional mapping, surgical judgment and postoperative care.
