Use of Optical Fluorescence Agents During Surgery for Pituitary Adenomas: Current State of the Field

Abstract

Despite improvements in optical visualization technology for pituitary adenoma (PA) surgery, differentiation of normal pituitary from abnormal tumor tissue remains a surgical challenge. During neurosurgical procedures for other tumor types, 5-aminolevulinic acid (5ALA) has become a focus of investigation based on its high specificity in differentiating tumor tissue. However, the role of 5-ALA and other optical fluorescent agents in PA surgery remains less clear. Using PRISMA guidelines, a systematic literature review to identify reports describing 5-ALA and other optical agents for fluorescence-guided surgery for PA was performed. Eleven research studies met inclusion criteria and were reviewed. In two studies, 5-ALA was not shown to be effective in aiding PA resection using standard neurosurgical endoscopic/microscopic approaches. 5-ALA photodynamic therapy was evaluated in two in-vitro models with inconsistent results. Intraoperative use of indocyanine green (ICG) concluded with varying results, but showed a tendency towards improved differentiation of functional PA. OTL38 showed potential for intraoperative identification of nonfunctioning PA, particularly in tumors with high folate receptor expression. One study reported clinically useful fluorescence following sodium fluorescein administration. We conclude that selected optical fluorescent agents, including ICG and folate receptors, are most likely to hold promise for clinical use in differentiating PA from normal tissue.

Introduction:

Pituitary adenomas (PA) are among the most common intracranial tumors, accounting for 10–15% of all intracranial neoplasms [1]. PA can be classified as either sub-centimeter microadenomas, or macroadenomas with a maximal dimension of at least one centimeter [2]. Functional, hormone-secreting PA can cause clinical syndromes related to anterior pituitary hormone hypersecretion and are often discovered as microadenomas [2–6]. Conversely, since patients often seek care after experiencing neurologic symptoms such as visual loss from compression of the optic apparatus, hormonally silent adenomas are typically macroadenomas at diagnosis [2, 7, 8].

Despite their often-small size, surgery is first-line therapy for functional PA causing acromegaly, Cushing’s disease and for a small subset of prolactinomas failing medical management [9–12]. During surgery, neurosurgeons currently perform a series of subjective visual and tactile assessments to differentiate normal versus abnormal pituitary gland when selecting sequential pathologic specimens and performing tumor resection, with a 30–45% rate of residual tumor and less than 5% rate of new hypopituitarism in resection of functional PA [13, 14]. Although intraoperative MRI may identify residual tumor in up to 30% of cases, its sensitivity for microadenomas ranges from 40–80% [15–18]. The consequences of residual functional microadenoma tissue are often significant: the probability of endocrinological cure at a subsequent surgery is much lower, and the morbidity and mortality from continued hormone hypersecretion may be dire [19–26]. Accordingly, intraoperative adjuncts to improve the differentiation of tumor from normal gland and expand the surgeon’s ability to detect residual tumor while protecting the normal pituitary gland have the potential to improve endocrinological, quality of life and survival outcomes in surgery for functional PA.

Even in macroadenomas, surgical visualization and tissue differentiation can be challenging. As many as 35% of PA invade surrounding tissues including dura, bone, paranasal sinuses and the suprasellar space [27]. Although many tumors are benign, invasive tumors may have an aggressive phenotype resulting in tumor recurrence if residual tumor persists after an operation [27]. Furthermore, macroadenomas are often adjacent to vital neurovascular structures and can compress the residual pituitary gland into a thin rim, making the differentiation between tumor and normal structures increasingly important [2].

In other areas of surgical neuro-oncology where tumor visualization is challenging, optical fluorescence agents including indocyanine green (ICG), OTL38, sodium fluorescein and 5-aminolevulinic acid (5-ALA) fluorescence guided surgery have been utilized to improve resection rates. Since ICG binds to plasma proteins and remains in the vasculature, ICG fluorescence is widely used for neurosurgical and ophthalmologic intraoperative angiography as well as other cardiovascular applications [28]. Initial studies have reported the use of ICG to differentiate PA from normal pituitary tissue [29, 30].

OTL38, a folate analog conjugated to cyanine dye, targets a folate receptor that is overexpressed in nonfunctional PA. Intraoperative use of OTL38 dye has shown that signal-to background ratio can be correlated with folate receptor expression, which could be used to predict adenomas from normal pituitary tissue[31, 32]. Other surgeons have investigated the use of sodium fluorescein, a yellow-green fluorophore commonly used in ophthalmology and visible to the naked eye. Fluorescein accumulation in brain tumor tissue is thought to be a result of vascular permeability and disruption of the blood brain barrier [33, 34].

The increasing popularity of 5-ALA in the resection of malignant gliomas has led to a resurgence of interest in its use in pituitary surgery [35, 36]. 5-ALA fluorescence enables tumor detection through a process known as Photodynamic Diagnosis. Infusion of exogenous 5-ALA bypasses ALA/heme feedback control and results in the accumulation of Protoporphyrin-IX (PpIX). It is speculated that tumor cells have lower ferrochetalase activity, one of the rate limiting steps in heme biosynthesis, resulting in the accumulation of PpIX [37–39]. Excitation of accumulated PpIX in tumor cells by blue-violet light (440nm), allows for visualization during surgery. In glioblastoma, 5-ALA is associated with gross total resection rates up to 65%, compared with conventional microsurgery rates of 36% without 5-ALA [8]. As a result of this success, there is interest in 5-ALA fluorescence guided surgery to improve resection rates and gland preservation in PA and other common intracranial tumors. In this study, a systematic review of the literature was conducted to evaluate studies assessing the use of 5-ALA and other optical fluorescent agents in the treatment of PA.

Methods:

A systematic literature review using PubMed was conducted according to the PRISMA guidelines ^[40] using the terms “PA” and “5-ALA”, “ALA”, “aminolevulinic acid”, “PpIX”, “Protoporphyrin IX”, “OTL38”, “optical fluorescence”, “indocyanine green”, “ICG”, and “fluorescent agent”.

There was no limitation on publication date. Duplicates, non-English and abstract only articles were excluded. Studies that examined the utilization of 5-ALA and other optical fluorescent agents in PA cell lines and human pituitary tumor operations were included.

Results:

The initial search resulted in 176 articles, after which 145 articles remained following removal of duplicates. Titles and abstracts were screened for relevance by two reviewers (S.C, D.A.D), resulting in the exclusion of 130 articles. Fifteen articles subsequently underwent full text review, of which eleven studies were included and four were excluded. Reasons for exclusion following full text review included one study being a review article, another an editorial, one study utilized a non-fluorescent agent, and another studied cerebrospinal fluid leaks during pituitary surgery using fluorescent agents.

Two studies looked at the intraoperative use of 5-ALA during PA surgery. Two additional studies investigated 5-ALA uptake in PA cell lines and effects of photodynamic therapy (PDT). See Table 1 and Table 2 for brief outlines of the studies. Seven studies were identified that assessed other intraoperative optical fluorescent agents for PA surgery. Please see Table 3 for a brief outline.

Table 1.

5-ALA fluorescence guided surgery.

Study	Year	Study type	No. of patients	Tumor type	Fluorescence	Postoperative outcomes	Comments
Eljamel et al.[41]	2009	prospective, single center	N = 30	14 NFA, 12 secreting PA (2 GH, 3 ACTH, 2 prolactin, 5 gonadotrophins) and 4 pituitary cysts.	Photodiagnostic system: 80.8% sensitivity and 75% specificity. Optical biopsy system: 95.5% sensitivity and 100% specificity	3 macroadenomas recurrences	PpIX-induced fluorescence detection helpful in locating 6 microadenomas when MRI scan were normal and macroadenomas do uptake and retain the photosensitizer
Marbacher et al. [36]	2014	prospective, single center	N = 458	12 PA	8.33% (1/12) positive fluorescence	n/a

Table 2.

5-ALA based photodynamic therapy

Study	Year	Cell lines	Flow cytometry	Spectrofluorometric analysis	5-ALA concentration (ug/mL)	Viability assay	PDT	Comments
Nemes et al.[42]	2016	GH3	51.4%	147046	25–200	No significant change	635nm, 25J/cm2 for 250 seconds, 9.3cm distance	There is uptake and metabolism of 5-ALA in the 2 PA cell lines. PDT had no significant effect on the survival of cells.
		AtT-20	62.6%	112295	25–200	No significant change
Neumann et al. [43]	2016	GH3	n/a	n/a	7.5–16.5	Significant dose dependent decrease	635nm, 18.75J/cm2 for 625 seconds, 9.3cm distance	GH3 and human PA cells line showed dose dependent response to PDT therapy. Different histological subtypes have different sensitivities to 5-ALA PDT
		Human PA: 5 null cell, 2 corticotroph, 4 gonadotrophic, 2 somatotrophic, 2 chromophobic	n/a	n/a	12.5–100	Significant dose dependent decrease

Table 3.

Other optical fluorescence agents

Study	Year	Study type	No. of patients	Tumor type	Fluorescence agent	Comments
Litvack et al. [44]	2012	prospective, single center	N = 9	4 null cell adenomas, 1 adrenocorticotrophic adenoma, 2 growth hormone secreting adenoma, 1 prolactinoma, 1 mammosomatotrophic adenoma	10 or 25mg ICG bolus	hypervascularity in prolactinoma and Acromegaly, NFMA had varying degree of contrast in fluorescence
Hide et al. [29]	2015	prospective, single center	N = 26	n/a	12.5mg ICG bolus	- average time of peak color value of normal pituitary (29.5 seconds) and PA (31 seconds) was not significant - median peak intensity of normal pituitary (123) and PA (136.5)
Sandow et al.[45]	2015	prospective, single center	N = 22	13 Acromegaly, 6 Cushing’s, 3 presented with other symptoms. (6 macroadenomas, 15 microadenoma, 1 MRI negative)	25mg ICG bolus	11 cases directly visualized, 11 cases indirectly visualized. Lesions detected in all microadenoma and negative MRI patients.
Verstegen et al. [30]	2016	prospective, single center	N = 10	4 NFMA, 4 Cushing’s, 1 Acromegaly, 1 Prolactinoma	5mg ICG bolus	1 PA positive fluorescence, 8 had weak or no fluorescent signal and 1 could not be discriminated from normal pituitary. Mean FCR 1.5 ± 0.2
da Silva et al. [46]	2016	prospective, single center	N = 1	n/a	1g of sodium fluorescein 20%	Strongly positive sodium fluorescein enhancement
Lee et al. [32]	2017	prospective, single center	N = 15	6 NF adenomas, 3 somatotrophic adenoma, 5 corticotropic adenomas, 1 corticotroph and somatotroph cells	25mg Benadryl and 0.025mg/kg OTL38 infused over course of 1 hour given 2 to 4 hours prior to surgery	- high folate overexpression (H > 200) had SBR of 3 ± 0.27 - low folate expression had a SBR of 1.6 ± 0.43 - NIR fluorescence: 86% sensitive and 89% specific for all tumors - Visible light: 80% sensitive, 89% specific for all tumors
Cho et al.[31]	2018	prospective, single center	N = 14	9 null-cell adenomas, 2 silent gonadotrophic adenomas, and 3 silent somatotrophic adenomas	25mg Benadryl and 0.025mg/kg OTL38 infused over course of 1 hour given 2 to 4 hours prior to surgery	- NIR fluorescence: 100% sensitive, 100% specific - without fluorescence: 83% sensitive.100% specific

5-ALA for PA Surgery

Eljamel et al. evaluated 5-ALA fluorescence guided surgery in 30 patients with PA, consisting of 20 macroadenomas (12 of which were invasive), 6 microadenomas, and 4 pituitary cysts, using two different visualization systems. First, an optical biopsy system using a laserbased probe was utilized to serially assess each quadrant of the gland. A Gallium-nitride laser unit and a compact spectrometer pass through a beam splitter into the probe, 20cm long and 3mm in diameter, to help visualize fluorescence below the gland surface and detect deeper adenomas. The optical biopsy had a 95.5% sensitivity (detection of PA tissue) and 100% specificity (detection of normal tissue). This system was able to identify 6 pituitary microadenomas that were not visualized on MRI. After the optical biopsy system was used, it was removed and a photodiagnostic filter was attached to the endoscope and the PA was exposed. The photodiagnostic endoscope filter had 80.8% sensitivity and 75% specificity [41].

Marbacher et al. studied 5-ALA fluorescence guided surgery in various primary brain tumors and metastases. 5-ALA was used during every brain tumor case and fluorescence was graded by the surgeon. Findings were dichotomized to positive or negative fluorescence. Out of 458 patients in the study, 12 presented with PA. Though the study did not specify macroadenoma versus microadenoma, only one out of the 12 (8.3%) had positive fluorescence after 5-ALA administration [36].

Nemes et al. demonstrated general uptake of 5-ALA in two different animal PA cell lines, GH3 and AtT-20, supporting the possibility of using 5-ALA during surgery and photodynamic therapy (PDT). Intracellular accumulation of PpIX was measured using flow cytometry and spectrofluorimetrical measurements after cells were incubated in 100g/mL of 5-ALA for 6 or 24 hours. Flow cytometry showed a positive shift of 51.4% and 62.6% for GH3 and AtT-20 cells, respectively, at 6 hours. Spectrofluorimetrical measurements indicated AtT-20 cell size was 1.5 times larger than GH3. Having demonstrated 5-ALA uptake in cells, PDT was attempted and showed no significant change in either cell lines at concentrations of 25–200ug/mL 5-ALA for 6 hours prior to 250 seconds of irradiation with 635nm light [42].

Neumann et al. also looked at 5-ALA PDT in GH3 cell lines and in PA cells from operated patients. Concentration dependent effect on cell death was reported in both cell lines, which were treated with varying concentrations of 5-ALA (7.5–16.5ug/mL GH3 and 12.5100ug/mL human cells) for 4 hours before 625 seconds of irradiation. LD50 was reported at 8.5ug/mL in GH3 cells and between 25–50ug/mL in human PA cells. Corticotroph adenomas were noted to be especially sensitive to PDT compared to other PA subtypes as the majority of cells were killed at concentrations below 70ug/mL [43].

ICG for PA Surgery

Litvack et. al reported their use of ICG to differentiate tumor from normal tissue using a custom photodynamic diagnostic system with telescope filter following standard endoscopic endonasal approach. Removal of the tumor was performed in standard fashion unless the tumor was exposed, in which case a second ICG injection was given. Initial fluorescent examination of the 9 cases showed hypervascularity in the 4 acromegaly and prolactinoma patients (44.4%). In non-functioning PA, a varying degree of contrast between tumor and normal pituitary was noted. No distinct interface was observed in the microadenoma case [44].

Hide et al. studied the use of ICG in endonasal transsphenoidal surgery in 26 patients presenting with PA, though size was not reported. Recorded images from surgery were used for analysis of ICG intensity and time of detection. The average time of peak color value between normal pituitary and tumor tissue was 29.5 seconds and 31 seconds, respectively, which was not statistically significantly different. The median peak color intensity was 123 for normal pituitary and 136.4 for pituitary tumor tissue, though significance could not be assessed due to a small sample size [29].

Sandow et al. studied ICG use during microscopic transsphenoidal surgery for PA. Of 22 lesions, 11 cases (50%) were visualized indirectly where the PA had lower ICG fluorescence than surrounding tissue, while the remaining 11 cases (50%) were visualized directly via ICG uptake. It was noted that all 6 cases of Cushing’s disease were directly visualized. The authors also report visualization of lesions in patients with microadenomas and those with MRI- negative PA. Post-operative MRI showed no evidence of residual tumor in 21 patients and a follow up recurrence rate of 1 out of 20 (5%) patients [45].

Verstegen et al. evaluated NIR fluorescence in pituitary surgery using a NIR-compatible endoscope. Fluorescence intensities were measured 45 seconds after ICG was administered, and in 9 of the 10 cases, normal pituitary exuded at a higher intensity. The PA fluorescence was rated as no signal or weak signal in 8 cases and positive in 1 case. Only in one patient was there no discrimination between normal pituitary and PA, due to suboptimal timing of the image capture. The average fluorescence contrast ratio was reported to be 1.5 ± 0.2 in the study [30].

OTL38 for PA Surgery

Lee et al. studied OTL38 based NIR imaging to target folate receptor alpha and calculate signal to background ratio (SBR) and folate receptor expression (H-scores). Visionsense Iridium system was utilized which allowed for the simultaneous overlay of visible light and NIR. It was noted that high folate overexpression (H > 200) had SBR of 3 ± 0.27, while low folate expression had SBR of 1.6 ± 0.43. The three PA with high folate overexpression were 3 of the 6 non-functional PA in the study. OTL38 was 86% sensitive and 89% specific for all tumors, but was 100% sensitive and specific when looking only at PA with high folate overexpression. Visible light, on the other hand, had 80% sensitivity and 89% specificity [32].

Cho et al. further focused OTL38 based NIR imaging OTL38 on non-functioning pituitary macroadenomas. Seven null-cell PA and two gonadotrophic PA had high folate receptor overexpression (H-score > 150) and intraoperative SBR of 3.2 ± 0.52, while the remaining two null-cell PA and three somatotrophic PA had low expression (H-score < 100) and SBR of 1.5 ± 0.18. Postoperative analysis of tumor margins collected during surgery showed that the surgeon’s impression of tumor margins without fluorescence was 83% sensitive and 100% specific, while NIR fluorescence was 100% sensitive and specific. NIR fluorescence was also reported to be 100% sensitive in predicting postoperative MRI results in cases with less than Knosp grade 4 cavernous sinus invasion [31].

Sodium Fluorescein for PA Surgery

da Silva et al. performed an introductory experiment studying the use of sodium fluorescein in skull base tumors, with 1 of the 6 being a PA of unstated size. Digital photos were taken pre- and 10 minutes post intravenous injection of 1g of sodium fluorescein 20%. These images were analyzed to calculate the total area presenting the sodium fluorescein wavelength. Comparison of pre- and post- photos results in a statistically significant difference between the two images with a Wilcoxon T test (P = 0.028) [46].

Discussion:

Modern pituitary surgery is currently predicated on direct visual and tactile identification of tumor and differentiation between tumor and normal structures. However, small and/or invasive tumors create challenges for optical visualization, with implications for long-term endocrinological response rates, quality of life and overall survival. Fluorescent dyes have improved lesion visualization in other brain tumor subtypes; 5-ALA is of particular interest based on the experience of glioma surgery [35, 36]. We aimed to review the literature to determine the current status of 5-ALA and other optical fluorescent agents for fluorescence imaging in PA surgery.

We identified two clinical studies in humans where 5-ALA was used to visualize PA during surgery. In both studies, optical assessment of 5-ALA fluorescence was not consistently clinically useful. While one study found an increased though non-uniform rate of 5-ALA in a sample size consisting largely of macroadenomas, the other showed negative fluorescence in the majority of PA cases. 5-ALA optical fluorescence should be limited to research protocols at present.

Since 5-ALA fluorescence can only detect surface lesions, further advances in intraoperative imaging might be required to leverage the use of 5-ALA in PA surgery. Eljamel developed and deployed a novel optical biopsy system utilizing a probe which reported a quantitative (spectrographic) assessment of 5-ALA fluorescence rather than an optical image. This “optical biopsy system” had extremely high sensitivity and specificity, raising interest in its application especially for challenging functional microadenomas such as refractory Cushing’s disease which might not be visualized on pre-operative MRI. However interesting this technique may be, it has not been reported on since Eljamel’s initial manuscript and has been neither validated in another institution nor mass-produced.

PDT also represents a promising therapy in other skin, bladder and gastrointestinal tumors. 5-ALA based PDT exploits PpIX’s property to act as a photosensitizer, which in combination with oxygen and red light (635nm) resulting in the generation of reactive oxygen species leading to cell death. Two studies examined cell viability post-PDT in GH3 cell lines with conflicting results. However, the studies differed in many aspects such as different cell culture media, different time points when performing cell viability assays, different 5-ALA concentrations and incubation time, as well as different irradiation times. This makes it difficult to compare the results as the two studies had no standardized protocol to follow. On the other hand, both studies did note differential PpIX accumulation in the different cell lines tested which represented different PA subtypes. Thus, PA subtype may play a role in the resulting sensitivity to 5-ALA PDT. However, no clinical reports of PDT for PA exist and this technique remains investigational at this time,

ICG fluorescence based surgery has been attempted in endoscopic transnasal transsphenoidal surgery by several authors. Real-time visualization of blood flow and differences in fluorescence intensity can be utilized to help surgeons identify neighboring neurovascular structures. Some studies have shown a contrast ration in fluorescence intensity allowing for the discrimination of normal pituitary gland from PA regardless of the size of the adenoma. PA subtype may affect fluorescence intensity since hormone-producing PA have been observed to have higher ICG uptake than normal tissue, likely due to the hypervascularization of these tumors. Several key limitations in this technique include the inability to detect ICG fluorescence using standard endoscopes and the variable rate of uptake of ICG between tumor and normal pituitary tissue. As new endoscopes may include lenses or filters for near-infrared wavelengths, the utility of ICG might improve. At this time, we consider adjunctive ICG to be a promising area that should be further explored in PA surgery.

OTL38 has been shown in two small studies to be useful, particularly in the detection of non-functioning PA with high folate overexpression, with 100% sensitivity and specificity. While size did not factor into OTL38 uptake, folate expression did as NIR was also able to detect and differentiate nonfunctional adenomas with high and low folate expression. When analyzing OTL38 detection in other subtypes of PA, the results were similar to that of visible light. While OTL38 may not be useful for some types of PA, those with folate receptor overexpression can likely benefit greatly from this technique to better detect and resect the tumor.

Sodium fluorescein provides an interesting option for tumor identification. In the one study currently available, marked tumor enhancement was seen in PA. Fluorescein is particularly useful because it fluoresces in the visual light spectrum and does not require any changes to the intraoperative optics. Additionally, since fluorescein is commonly used to diagnose cerebrospinal fluid leaks, it might be useful to investigate tumor fluorescence during cases where fluorescein has already been instilled preoperatively. Challenges with fluorescein use include the need to infuse the drug into the subarachnoid space via placement of a lumbar drain and a small but well documented risk of seizures, particularly if the drug is infused rapidly. Though more studies are needed to support sodium fluorescein intraoperative use, it has proven to be a promising adjunct that can be utilized universally due to its low cost and safety.

We identified clear limitations of optical fluorescent agents during PA surgery. Currently, fluorescence techniques can only detect of PA near the surface of the operative field due to their requiring a direct application of polarized light. Novel technological devices could be developed to address deeper lesions, but no market solutions currently exist. Current in vitro studies are limited by a lack of standardized protocols, small sample sizes, and lack of primary human cell cultures. These limitations should be considered in designing future studies and prior to the clinical use of optical fluorescent dyes during PA surgery.

Conclusion:

Identification of small and residual tumor tissue during PA surgery is a major surgical challenge with profound effects on patient outcomes. We conducted a systematic review of the literature and identified four optical fluorescence agents with varying degrees of promise as diagnostic adjuncts during PA surgery. Currently available evidence does not describe a clear clinical benefit from intraoperative fluorescence imaging using 5-ALA, ICG, OTL-38 or fluorescein in PA surgery in terms of extent of resection or neuroendocrine outcomes. In the future, fluorescence techniques could be particularly useful in MRI negative patients with functional PA syndromes. In a handful of small studies, microadenomas were positively identified with the use of 5-ALA and ICG adjuncts, but further validation of these reports is needed. Pituitary subtype and degree of folate expression affects the uptake of fluorescent dyes: further investigation of this relationship is warranted prior to clinical use in functional PA. Folate-based dyes such as OTL-38 may be particularly useful for intraoperative visualization of PA. Directions for future research include investigations of tumor proliferation markers and PA subtype on fluorescence, efficacy of 5-ALA PDT in vivo and ongoing exploration of these optical fluorescent contrast agents impacts on gross total resection rates. At this time, intraoperative fluorescence remains a promising, albeit investigational, adjunctive technique during PA surgery.