Indocyanine Green Fluorescence in Endoscopic Resection of Non-Functioning Pituitary Adenomas

It is important to distinguish between the tumour and healthy tissue. This allows for maximum tumour removal with minimal damage to surrounding structures. In neurosurgery, various dyes are already used for specific medical conditions and have been shown to improve treatment outcomes. One example is 5-aminolevulinic acid (5-ALA), which is most commonly used in surgeries for primary brain tumours but also shows potential benefit in surgeries for meningiomas and brain metastases. Sodium fluorescein (SF) offers similar benefits in these pathologies. Both dyes accumulate in tumour tissues due to a compromised blood-brain barrier (BBB). The use of both dyes has been investigated in endoscopic endonasal surgeries. 5-ALA does not yet show significant benefit in these surgeries. SF has shown some potential benefits, but so far the literature is sparse. It is suggested that there is a quantitative difference in the degree of fluorescence of tumour tissue compared to surrounding structures, but these differences cannot be subjectively determined intraoperatively. On the other hand, ICG has proven to be a promising intraoperative marker. After intravenous administration, the dye causes the veins to fluoresce, including the capillaries. Because ICG can highlight capillary flow, it has increasing potential in endoscopic pituitary surgery. As adenomas have significantly lower capillary density compared to normal pituitary tissue, they do not fluoresce after ICG administration, whereas the pituitary gland does.

To date, no randomised, prospective studies have been published, which compare surgical complications in patients operated on with the conventional method and those operated on with ICG. In our exploratory study, the investigators aimed to determine the feasibility of intraoperative use of ICG dye and whether it would result in fewer complications and more frequent complete removal of the adenoma. If feasibility is confirmed, this could demonstrate the potential added value of ICG in endoscopic endonasal surgery of pituitary adenomas, resulting in better patient outcomes.

A prospective, randomised, pilot clinical study, single-blinded for patients and for radiological and ophthalmological outcome assessment, was conducted in the Clinical Department of Neurosurgery at the University Medical Centre Ljubljana, Slovenia. The study was approved by the Medical Ethics Committee of the Republic of Slovenia (No. 0120-56/2019/6). It was not registered with the clinical trial registry prospectively, because it was considered small, pilot, single-centre investigation and was not required by the institution where the study was taking place at the time it was initiated. The investigators included all non-functional pituitary macroadenoma patients surgically treated at the hospital during the period between February 2019 and March 2023, who were not allergic to iodine, were older than 18 years, were not pregnant, and consented to participate. Patients were randomly divided into two groups: one group received ICG during surgery, while the other served as the control group. Randomisation was performed using block randomisation in a 1:1 ratio. Blocks of sizes 4 and 6 were used, with participants randomly assigned within each block to ensure a balanced number of participants in both groups. The randomisation sequence was generated by an independent researcher. Given the exploratory nature of the study and limited patient availability at our centre, the sample size was not based on an a priori power calculation. Instead, it was determined by the number of eligible patients who could be recruited within the available time frame.

Surgical technique In all patients, an endoscopic endonasal transsphenoidal approach was used to remove the pituitary macroadenoma, which the investigators extended with a transplanar, transtubercular (intensive suprasellar tumour extension) or transpterigoid (parasellar extension into the lateral compartment) approach if necessary. In all cases, the investigators operated only through the left nostril, lateralised the middle turbinate, performed a posterior septotomy and a sphenoidotomy. For extended approaches, where there was an intraoperative cerebrospinal fluid (CSF) leak, the investigators made a nasoseptal (Hadad) flap. The posterior sphenoidotomy extended from one anterior loop of the intracavernous carotid artery to the other; in cases of extension into the cavernous sinus, the investigators extended the access on the ipsilateral side with a posterior ethmoidotomy, drilling of the lateral wall of the sphenoid sinus, and a modified transpterigoid approach.

The investigators have opened the dura of the sella turcica with a circular incision. These steps allowed maximal exposure of the macroadenoma, which could then be removed. The investigators did not use curettage as there were no blind spots in the field of view; every part of the surgical field could be seen with the camera. After exposing the intrasellar space, the investigators first administered ICG dye intravenously to improve differentiation between the pituitary gland and the tumour. The carotid artery was stained after about 15-20 seconds, the pituitary gland after 20-30 seconds. After complete macroscopic removal of the tumour, the investigators administered ICG a second time to check the surgical field for any remaining tumour residue. The investigators used a technique of stepwise tumour removal with systematic keyhole aspiration along the border between the adenoma and the pituitary gland. First, the investigators removed part of the tumour at the base of the sella turcica, then along the medial walls of the cavernous sinuses, while the suprasellar part gradually folded downwards. After complete removal, closure was performed using an artificial dural substitute (DuraGen Secure, Integra LifeSciences, Princeton, NJ, USA) as an onlay graft, which was additionally reinforced with TachoSil patches (Takeda Pharmaceuticals, Tokyo, Japan). The diaphragma sellae was further supported with Spongostan (Ethicon, Somerville, NJ, USA), and the reconstruction was sealed with DuraSeal fibrin sealant (Integra LifeSciences, Princeton, NJ, USA). In the three cases where a Hadad flap was used, the flap was first positioned and then reinforced with TachoSil, Spongostan, and DuraSeal sealant. At the end of the procedure, Rapid Rhino nasal tampons (Smith \& Nephew, Watford, UK) were inserted into both nostrils and removed 1-2 days postoperatively.

Indocyanine green dye The investigators used the ICG dye (Serb, Paris, France). 25 mg of the ICG dye was dissolved in 10 mL of sterile water. After exposing the intrasellar space, the investigators injected 5 mL of the solution (12.5 mg ICG) intravenously as a bolus during surgery and after macroscopic removal of the macroadenoma.

Optics For endoscopic endonasal transsphenoidal surgery, the investigators used a rigid endoscope Karl Storz 7230AA Hopkins II Telescope 0°, 4 mm/18 cm, Tuttlingen, Germany. For this study, an endoscope with ICG light detection, Karl Storz 8710AGA Hopkins II Rigid ENT Scope 0°, 5.8 mm/20 cm, was used in the ICG group. During the operation, the light source could be switched between white and near-infrared light using a switch on the foot pedal.

Radiological evaluation All patients underwent MR imaging of the head with contrast medium before and 3 to 6 months after surgery according to pituitary gland protocol. Tumour dimensions were measured in the transverse, sagittal and coronal planes. Tumour volume was calculated as half the product of all three dimensions. Tumours were classified according to Knosp based on coronal sequences and Hardy based on sagittal sequences. The parasellar extension was additionally assessed according to Miranda's classification, which describes into which compartment of the cavernous sinus the tumour had invaded.

Assessment of complications Intraoperative complications such as cerebrospinal fluid (CSF) leak, damage to the carotid artery and cavernous nerves were assessed during surgery. For postoperative complications, the investigators paid attention to postoperative CSF leak, tumour residue, visual disturbances, postoperative haematoma, meningitis, double vision and facial dysaesthesia. CSF leak was confirmed by increased beta trace levels in nasal discharge or by direct endoscopic examination of the surgical site by otolaryngologist. Tumour remnants were assessed by MRI examination according to pituitary gland protocol 3-6 months after surgery by radiologist, blinded to the treatment modality. Any radiological finding suggestive of residual tumour, including ambiguous finding, was classified as a residual. All patients underwent a visual field examination before and 3-6 months after surgery by an ophthalmologist, who was also blinded to the treatment modality. Haematoma was confirmed by CT scan of the head. Meningitis was determined by microbiological isolation of bacteria from CSF samples, while double vision and facial dysaesthesia were determined based on the patient's clinical presentation.

Statistical analysis of data The data were analysed using the statistical programme SPSS 25 (IBM Corp., Armonk, USA). P values of less than 0.05 were considered statistically significant.

Demographic and clinical characteristics of patients that were normally distributed were presented as arithmetic mean and standard deviation, while asymmetrically distributed variables were presented as median and first and third quartiles. 95% confidence intervals were derived from sample standard deviation. Categorical variables were presented as relative and absolute frequencies.

To compare the proportions between the groups of patients with and without ICG dye, the investigators used the chi-square test or Fisher's exact test.