Folate receptor overexpression can be visualized in real time during pituitary adenoma endoscopic transsphenoidal surgery with near-infrared imaging

Full access

OBJECTIVE

Pituitary adenomas account for approximately 10% of intracranial tumors and have an estimated prevalence of 15%–20% in the general US population. Resection is the primary treatment for pituitary adenomas, and the transsphenoidal approach remains the most common. The greatest challenge with pituitary adenomas is that 20% of patients develop tumor recurrence. Current approaches to reduce recurrence, such as intraoperative MRI, are costly, associated with high false-positive rates, and not recommended. Pituitary adenomas are known to overexpress folate receptor alpha (FRα), and it was hypothesized that OTL38, a folate analog conjugated to a near-infrared (NIR) fluorescent dye, could provide real-time intraoperative visual contrast of the tumor versus the surrounding nonneoplastic tissues. The preliminary results of this novel clinical trial are presented.

METHODS

Nineteen adult patients who presented with pituitary adenoma were enrolled. Patients were infused with OTL38 2–4 hours prior to surgery. A 4-mm endoscope with both visible and NIR light capabilities was used to visualize the pituitary adenoma and its margins in real time during surgery. The signal-to-background ratio (SBR) was recorded for each tumor and surrounding tissues at various endoscope-to-sella distances. Immunohistochemical analysis was performed to assess the FRα expression levels in all specimens and classify patients as having either high or low FRα expression.

RESULTS

Data from 15 patients (4 with null cell adenomas, 1 clinically silent gonadotroph, 1 totally silent somatotroph, 5 with a corticotroph, 3 with somatotrophs, and 1 somatocorticotroph) were analyzed in this preliminary analysis. Four patients were excluded for technical considerations. Intraoperative NIR imaging delineated the main tumors in all 15 patients with an average SBR of 1.9 ± 0.70. The FRα expression level of the adenomas and endoscope-to-sella distance had statistically significant impacts on the fluorescent SBRs. Additional considerations included adenoma functional status and time from OTL38 injection. SBRs were 3.0 ± 0.29 for tumors with high FRα expression (n = 3) and 1.6 ± 0.43 for tumors with low FRα expression (n = 12; p < 0.05). In 3 patients with immunohistochemistry-confirmed FRα overexpression (2 patients with null cell adenoma and 1 patient with clinically silent gonadotroph), intraoperative NIR imaging demonstrated perfect classification of the tumor margins with 100% sensitivity and 100% specificity. In addition, for these 3 patients, intraoperative residual fluorescence predicted postoperative MRI results with perfect concordance.

CONCLUSIONS

Pituitary adenomas and their margins can be intraoperatively visualized with the preoperative injection of OTL38, a folate analog conjugated to NIR dye. Tumor-to-background contrast is most pronounced in adenomas that overexpress FRα. Intraoperative SBR at the appropriate endoscope-to-sella distance can predict adenoma FRα expression status in real time. This work suggests that for adenomas with high FRα expression, it may be possible to identify margins and to predict postoperative MRI findings.

ABBREVIATIONS FRα = folate receptor alpha; GTR = gross-total resection; ICG = indocyanine green; iMRI = intraoperative MRI; MOCR = medial opticocarotid recess;NF = nonfunctional; NIR = near infrared; SBR = signal-to-background ratio.

OBJECTIVE

Pituitary adenomas account for approximately 10% of intracranial tumors and have an estimated prevalence of 15%–20% in the general US population. Resection is the primary treatment for pituitary adenomas, and the transsphenoidal approach remains the most common. The greatest challenge with pituitary adenomas is that 20% of patients develop tumor recurrence. Current approaches to reduce recurrence, such as intraoperative MRI, are costly, associated with high false-positive rates, and not recommended. Pituitary adenomas are known to overexpress folate receptor alpha (FRα), and it was hypothesized that OTL38, a folate analog conjugated to a near-infrared (NIR) fluorescent dye, could provide real-time intraoperative visual contrast of the tumor versus the surrounding nonneoplastic tissues. The preliminary results of this novel clinical trial are presented.

METHODS

Nineteen adult patients who presented with pituitary adenoma were enrolled. Patients were infused with OTL38 2–4 hours prior to surgery. A 4-mm endoscope with both visible and NIR light capabilities was used to visualize the pituitary adenoma and its margins in real time during surgery. The signal-to-background ratio (SBR) was recorded for each tumor and surrounding tissues at various endoscope-to-sella distances. Immunohistochemical analysis was performed to assess the FRα expression levels in all specimens and classify patients as having either high or low FRα expression.

RESULTS

Data from 15 patients (4 with null cell adenomas, 1 clinically silent gonadotroph, 1 totally silent somatotroph, 5 with a corticotroph, 3 with somatotrophs, and 1 somatocorticotroph) were analyzed in this preliminary analysis. Four patients were excluded for technical considerations. Intraoperative NIR imaging delineated the main tumors in all 15 patients with an average SBR of 1.9 ± 0.70. The FRα expression level of the adenomas and endoscope-to-sella distance had statistically significant impacts on the fluorescent SBRs. Additional considerations included adenoma functional status and time from OTL38 injection. SBRs were 3.0 ± 0.29 for tumors with high FRα expression (n = 3) and 1.6 ± 0.43 for tumors with low FRα expression (n = 12; p < 0.05). In 3 patients with immunohistochemistry-confirmed FRα overexpression (2 patients with null cell adenoma and 1 patient with clinically silent gonadotroph), intraoperative NIR imaging demonstrated perfect classification of the tumor margins with 100% sensitivity and 100% specificity. In addition, for these 3 patients, intraoperative residual fluorescence predicted postoperative MRI results with perfect concordance.

CONCLUSIONS

Pituitary adenomas and their margins can be intraoperatively visualized with the preoperative injection of OTL38, a folate analog conjugated to NIR dye. Tumor-to-background contrast is most pronounced in adenomas that overexpress FRα. Intraoperative SBR at the appropriate endoscope-to-sella distance can predict adenoma FRα expression status in real time. This work suggests that for adenomas with high FRα expression, it may be possible to identify margins and to predict postoperative MRI findings.

Pituitary adenomas have an estimated prevalence of 15%–20% in the US, and they account for approximately 10% of intracranial tumors.10 Although they are predominantly benign, pituitary adenomas can still cause significant disability from compression of the adjacent neural structures and hypersecretory syndromes.

While prolactinomas and some somatotroph adenomas can be medically managed with cabergoline and octreotide, respectively, surgical management remains the primary treatment for nonfunctional (NF) adenomas and other functional adenomas.13,21 Currently, the endoscopic transsphenoidal approach is the most common surgical approach for resecting pituitary adenomas. The extent of resection is based on the surgeon’s visual and tactile impression of the tissue. Achieving complete resection can be difficult, and the tumor recurrence rate can be as high as 20% after surgery.24,34 Some surgeons have resorted to implementing MRI scanners in the operating room to maximize extent of resection, but this remains expensive and limited in availability.11 In addition, intraoperative MRI has been shown to result in a high false-positive rate and is not currently recommended.16 This study proposes a novel optical contrast technique that takes advantage of the pituitary adenoma’s overexpression of folate receptors.

Folate receptor alpha (FRα) overexpression has been reported in pituitary adenomas, especially NF adenomas, which overexpress FRα more than 20 times above the level of the normal pituitary gland and surrounding intracranial structures.7,8,18 This makes FRα an attractive target for specific labeling of pituitary adenoma with a fluorescent dye that can be visualized intraoperatively.32 OTL38 (On Target Laboratories) is a folic acid analog conjugated to an analog of indocyanine green (ICG), a near-infrared (NIR) fluorescent dye. This conjugated dye binds folate receptors. Fluorescence in the NIR range has advantages over visible light by virtue of its superior tissue penetration and lack of autofluorescence in normal tissues.25 We hypothesized that OTL38 would selectively bind to the folate receptors on pituitary adenomas that overexpress FRα and allow intraoperative visualization of neoplastic tissue.

Methods

Study Design

This prospective cohort study was approved by the University of Pennsylvania Institutional Review Board, and all patients gave informed consent. Adult patients over the age of 18 years who presented with a pituitary tumor were eligible for this study. Pregnancy and history of allergy to Benadryl (Johnson & Johnson) or OTL38 were the main exclusion criteria. All patients underwent preoperative MRI of the brain and sella with and without gadolinium contrast (Table 1). Patients were informed that enrollment in the study would not substantially change the scope of surgery, as resection would proceed with conventional techniques. Margin biopsies, however, would be obtained to study the value of intraoperative NIR imaging.

TABLE 1.

Clinical and pathological characteristics of the study patients

Case No.Age (yrs), SexIHC FindingPreop Biochemical Status*Clinical Features at PresentationPreop/Postop Visual Field DeficitMax Tumor Dimension (mm)Postop ComplicationsHormonal Changes
Nonsecretory adenoma
 567, FNegativeMild hyperprolactinemia, hypogonadotropic hypogonadismIncidental MRI findingNo/no20NoneEuhormonal
 780, MNegativeEuhormonalBitemporal hemianopiaYes/no17NoneEuhormonal
 1245, MNegativeHypogonadotropic hypogonadismRt temporal hemianopia, headachesYes/yes28NoneResolving hypogonadism
 1652, FNegativeEuhormonalIncidental MRI findingNo/no15Self-limited, mild SIADHEuhormonal
 139, FLH, FSH, rare TSHElevated alpha subunit, mild hyperprolactinemiaSeizure, headacheNo/no55Transient DIEuhormonal
 267, FGHHypogonadotropic hypogonadismBitemporal hemianopiaYes/no37Transient SIADHEuhormonal
ACTH-secreting adenoma
 936, FACTHMild hyperprolactinemia, hypercortisolemiaInfertilityNo/no13Transient DIEuhormonal
 1434, FACTHHypercortisolemiaCushing’s diseaseNo/no10NoneReduced hypercortisolemia
 1554, MACTHMild hypercortisolemia, mildly elevated IGF-1, hypogonadotropic hypogonadismHeadacheNo/no11NoneResolved hypogonadism & hypercortisolemia w/ persistent ACTH elevation
 1836, FACTHMild hyperprolactinemia, mild hypercortisolemiaGalactorrhea, weight gainNo/no15NoneEuhormonal
 1951, FACTHElevated IGF-1Incidental MRI finding, increased ring sizeNo/no6NoneEuhormonal
GH-secreting adenomas
 645, MGH, rare ACTHElevated IGF-1, hypogonadotropic hypogonadismAcromegaly, hypogonadismYes/yes38NoneHypothyroid, hypogonadal, persistent IGF-1 elevation
 874, MGHHypogonadotropic hypogonadism, elevated IGF-1Polyuria/polydipsia, new-onset DMYes/no30NoneNormalized IGF-1 & persistent hypogonadism
 1174, FGHElevated IGF-1Weight gain, new-onset DM, increased ring size, bitemporal hemianopiaYes/no10Transient SIADHHypothyroid, adrenal insufficiency
Mixed adenoma
 366, MGH, ACTHMild hyperprolactinemia, hypogonadotropic hypogonadism, elevated ACTHHeadache, ptosis, diplopiaNo/no28NoneEuhormonal

ACTH = adrenocorticotropic hormone; DI = diabetes insipidus; DM = diabetes mellitus; FSH = follicle-stimulating hormone; GH = growth hormone; IGF-1 = insulin-like growth factor–1; IHC = immunohistochemical; LH = luteinizing hormone; SIADH = syndrome of inappropriate antidiuretic hormone secretion; TSH = thyroid-stimulating hormone.

Mild prolactin elevation defined as < 60 ng/ml.

Mild cortisol defined as urine free cortisol < 2 times the upper limit of normal.

IGF-1 elevation defined as < 2 times the upper limit of normal.

OTL38 Administration

Patients who were enrolled in this study were instructed to stop taking any folate supplements, including multivitamins, 48 hours prior to surgery to reduce interactions with OTL38. Two to 4 hours prior to surgery, patients were injected with 25 mg Benadryl to minimize allergic reactions, and 0.025 mg/kg OTL38 was infused over the course of 1 hour. Patients were monitored for adverse reactions during and 30 minutes after infusion, and then they were checked in for surgery.

Surgical Approach

A uninostril or binostril endoscopic endonasal approach was performed in all patients. The endonasal approach was performed by otorhinolaryngologists, primarily using an endoscopic camera and endoscope system (Storz). The choice of the uninostril or binostril approach was made based on patient anatomy.6,29 A standard transnasal approach was employed by opening both sphenoid sinuses, removing the sphenoid rostrum, and resecting the sphenoid intersinus septum. The mucosa of the sella was removed in an atraumatic fashion to keep the surgical field dry. After exposing the sella face and sella floor, the sella was opened using standard techniques, and the dura was kept intact as much as possible to visualize NIR signal through the dura. Once the dura was opened, the NIR camera system was used to visualize fluorescence within the tumor. Tumor resection was performed using standard Storz endoscopic equipment, and, on completion, the NIR camera system was brought back in to inspect the resection margins. On satisfactory completion of the surgery, closure was performed jointly by otorhinolaryngology and neurosurgery personnel depending on the need for CSF leak repair. Abdominal fat was harvested only in select cases; in other instances, either a free mucosal graft or nasoseptal flap was used.

NIR System

In all cases, the NIR signal was visualized using the Visionsense Iridium camera system (Visionsense). This system is FDA approved for perfusion imaging in plastic and reconstructive surgery (for example, to assess tissue flap vascularity). The Visionsense Iridium camera system was coupled to a dedicated 4-mm outer-diameter endoscope (Fig. 1). The excitation source is a laser tuned to the NIR range (785 nm), and the sensor was filtered for emissions in the 800- to 835-nm range. The endoscope consists of a 4-mm outer-diameter scope, which features a dual optical path design, allowing the separate and independent use of white light and NIR light. Having separate paths for visible light and NIR images allows very faint fluorescence images to be acquired in the presence of strong white light. Image processing was performed in real time and displayed at 1080-pixel video resolution, with recording at 720 pixels.

Fig. 1.
Fig. 1.

Visionsense equipment setup for NIR imaging. Upper: The excitation laser for NIR imaging is attached to a 4-mm endoscope. Copyright Visionsense. Published with permission. Lower: Images are displayed on the leftmost screen. Figure is available in color online only.

DSouza et al. recently compared the Visionsense Iridium system with 5 other commercially available fluorescent visualization systems, including those manufactured by PerkinElmer (Solaris), Quest, Novadaq, and others.5 They concluded that the Visionsense Iridium system demonstrates the highest sensitivity for NIR fluorescent dye: its sensitivity is in the picomolar range for IRDye 800 (LI-COR Biosciences), which has excitation and emission characteristics that are similar to OTL38 and ICG. In addition to high sensitivity, the Visionsense system features a “smart image processing algorithm to produce a wide dynamic range,” which allows for better comparisons of high and low signal-to-background ratios (SBRs).5

Study Procedure

On encountering the tumor, the Visionsense Iridium 4-mm endoscope was used to simultaneously view white light and NIR images (Fig. 2). These 2 images were superimposed on the video output in real time to allow intraoperative analysis of the NIR signal. The NIR signal through the dura was recorded, as was the NIR signal through the pituitary adenoma after the dura was opened. After pituitary exposure, specimens were obtained of the mass and coded by the attending surgeon (J.Y.K.L. or M.S.G.) as consistent with the tumor based on the surgeon’s impression under white light (yes vs no) and NIR fluorescence (yes vs no) and then sent to the pathology laboratory for histopathological diagnosis. Surgery then proceeded in the standard-of-care manner without the use of NIR fluorescence.

Fig. 2.
Fig. 2.

Case 7. This patient had an NF adenoma, high FRα expression on final pathological analysis, and intraoperative NIR fluorescence. A and B: Preoperative sagittal (A) and coronal (B) MR images without contrast showing a 17-mm mass in the sella. C and D: Intraoperative endoscopic imaging with the dura open shown with white light only (C) and the NIR signal superimposed on white light (D). The MOCRs can be seen clearly and are used as a measure of distance from the endoscope to the target. E and F: Intraoperative endoscopic image of the sella after resection of the gross tumor. The residual tumor was identified on the right, both with white light only (E) and the NIR signal superimposed on the white light (F). This fluorescent specimen was confirmed on the final pathological analysis to be consistent with pituitary adenoma. G and H: Intraoperative endoscopic image of sella contents after resection of the residual neoplasm with white light only (G) and NIR superimposed on white light (H). No obvious fluorescence is identified. I: Example of SBR and MOCR calculation. Five points are chosen in the gross tumor itself, with the Visionsense camera calculating NIR signal for the points (75 + 74 + 66 + 61 + 57)/5 = 67 ± 7.9. Five points are chosen in the surrounding bone close to the tumor but outside of the obvious specimen (22 + 29 + 34 + 19 + 17)/5 = 24 ± 7.1. The SBR is thus 67/24 = 2.8. The black triangles denote the locations of the MOCRs, spanning roughly 50% of the field, that were used to approximate the endoscope distance.

After the tumor had been resected and the attending neurosurgeon was satisfied that complete resection had been achieved based on white light visualization, NIR imaging was used to identify areas of residual disease (Fig. 2E and F). Areas were biopsied at the discretion of the senior surgeon (J.Y.K.L. or M.S.G.), and specimens were coded and sent to the pathology laboratory.

Patients were admitted to the intensive care unit following surgery, and no adverse outcomes were reported. Postoperative MRI was performed on postoperative Day 1, and the patients were seen approximately 2 and 4 weeks after resection.

Immunohistochemistry for FRα

The specimens obtained from each patient were chemically fixed, embedded in paraffin blocks, and cut to 5-μm-thick sections. These permanent sections were stained with H & E. Furthermore, immunohistochemical analysis against FRα was performed using murine monoclonal antibodies against the folate receptor (1:20 dilution of NCL-L-FRα, Leica Biosystems). To quantify FRα expression in each sample, H-scores were calculated by a single, blinded neuropathologist (M.M.L.) using a positive control as the standard, by summing the percentages of adenoma cells that stained strongly (3+) multiplied by 3, the percentage of adenoma cells that stained moderately (2+) multiplied by 2, and the percentage of adenoma cells that stained weakly (1+) multiplied by 1 (Fig. 3 and Table 2).4,28 The H-score ranges from 0 to 300, with 300 implying strong staining in all cells.

Fig. 3.
Fig. 3.

Immunohistochemical examination demonstrating the level of FRα expression in pituitary adenomas. A: Kidney is used as the positive control with an H-score of 300. Scale bar = 300 µm. B: Normal adenohypophysis with an H-score of 130. Scale bar = 600 µm. C and D: Case 12. Low-power (C) and high-power (D) views of the NF adenoma with a high H-score (270). Scale bar = 1 mm and 200 µm, respectively. E and F: Case 2. Low-power (E) and high-power (F) views of the somatotroph adenoma with an intermediate H-score of 100. Scale bar = 500 µm and 200 µm, respectively. G and H: Case 8. Low-power (G) and high-power (H) views of the somatotroph adenoma with a low H-score of 5. Scale bar = 600 µm and 200 µm, respectively. Figure is available in color online only.

TABLE 2.

Results of immunohistochemical analysis of FRα

Case No.PathologyFRα Staining Intensity% Cells StainedH-ScoreSBR Close (MOCR >0.7)SBR Middle (MOCR 0.4–0.7)SBR Far (MOCR <0.4)
Nonsecretory adenoma
 5Null cell0001.5NANA
 7Null cell1+, 2+, 3+20, 60, 202002.92.82.8
 12Null cell2+, 3+30, 702703.9NA3.3
 16Null cell1+5521.61.5
 1Gonadotroph1+, 2+, 3+25, 25, 502253.73.12.8
 2Somatotroph1+, 2+80, 101002.42.2NA
ACTH-secreting adenoma
 9Corticotroph0002.2NANA
 14Corticotroph1+55NA2.41.8
 15Corticotroph1+10101NANA
 18Corticotroph1+, 2+5, 5152.31.81.6
 19Corticotroph2+10201.81.7NA
GH-secreting adenoma
 6Somatotroph1+552.2NA1.9
 8Somatotroph1+552.1NA1.1
 11*Somatotroph000NANA2.5
Mixed adenoma
 3Corticosomatotroph1+40401.61.21

NA = not applicable (images were not taken at this distance).

Patient presented with an adenoma and tuberculum sellae meningioma.

NIR Image Analysis of SBR and Laser Distance

The data from 15 patients were analyzed using Visionsense software (VSPlayer v1.8.05.01; Fig. 2I). Between 3 and 5 region of interest points were placed on the adenoma, and then an additional 3–5 points of interest were placed on the surrounding normal tissue (such as the bone of the tuberculum sella or carotid prominences) to determine the signal-to-background ratio (SBR). Because the intravenous administration of OTL38 causes normal mucosa and skin to fluoresce when the laser source is sufficiently close, fluorescence in the mucosa that was more proximal to the laser than the pituitary mass was disregarded. In general, the surgeon removed the mucosa from the sella and surrounding sella as part of the normal surgical procedure.

Because the Visionsense endoscope does not have the capacity to measure the distance from the laser source to the fluorescent object, a substitute measurement for estimating this distance was developed. The medial opticocarotid recess (MOCR) provides a useful landmark in the endoscopic approach to the sella.15,17,38 The distance between the 2 MOCRs as a fraction of the overall field size gives a rough estimate of the distance of the laser source to the fluorescent tissue (Fig. 2I).

Data Analysis

Stata (version 10, StataCorp) was used to compute the summary statistics and perform the t-tests and linear regression analysis.

Results

Clinical Data

A total of 19 patients were enrolled between October 2015 and May 2016. Two patients were excluded because the laser serving as the excitation light source malfunctioned during surgery. Two patients were excluded because their final pathological review did not demonstrate pituitary adenoma (craniopharyngioma and Rathke cleft cyst). One patient, who presented with a somatotroph adenoma and tuberculum sellae meningioma, is included in the data analysis, but SBR data were censored due to interference with the NIR signal measurement from the meningioma.

Data from 15 patients were available for analysis. Seven patients were male, and the mean patient age was 53.2 years (range 35–80 years). Six patients had NF adenomas (4 patients had a null cell adenoma, 1 patient had a clinically silent gonadotroph adenoma, and 1 patient had a totally silent somatotroph adenoma), 3 patients had somatotroph adenomas, 5 patients had corticotroph adenomas, and 1 patient had both corticotroph and somatotroph cells, as determined by immunohistochemical analysis (Table 1).

All patients tolerated injections of 0.025 mg/kg OTL38 administered 2 to 4 hours prior to surgery without adverse events. All patients did well after surgery with no unique complications identified. Seven of 9 patients with functional adenomas had postoperative control of hormonal hypersecretion. A surgical cure was not attained in the remaining 2 patients and further medical treatment was required (Table 1). The rate of gross-total resection (GTR) based on postoperative MRI was 73% (Table 3). Residual tumor was identified on MRI only in patients with cavernous sinus invasion or a significant extrasellar tumor, such as a tumor in the sylvian fissure.

TABLE 3.

Summary of preoperative MRI status and postresection MRI and fluorescence status

Case No.PathologyFRα OverexpressionPreop Tumor LocationPostop Tumor Detected
SellaSuprasellarCavernous SinusSylvian FissurePostop Day 1 MRIPostresection NIR Fluorescence
Nonsecretory adenoma
 5Null cellNoYesYesNoNoNoneNone
 7Null cellYesYesYesNoNoNoneNone
 12Null cellYesYesYesYesNoYesYes
 16Null cellNoYesNoNoNoNoneNone
 1GonadotrophYesYesYesYesYesYesYes
 2SomatotrophNoYesYesYesNoNoneNone
ACTH-secreting adenoma
 9CorticotrophNoYesYesNoNoNoneNone
 14CorticotrophNoYesNoNoNoNoneNone
 15CorticotrophNoYesNoYesNoNoneNone
 18CorticotrophNoYesNoNoNoNoneNone
 19CorticotrophNoYesNoNoNoNoneNone
GH-secreting adenoma
 6SomatotrophNoYesYesYesNoYesNA
 8SomatotrophNoYesYesYesNoNoneNone
 11*SomatotrophNoYesNoNoNoNoneNone
Mixed adenoma
 3SomatocorticotrophNoYesNoYesNoYesNone

The patient had a tuberculum sellae meningioma.

The laser excitation source failed after initial visualization of the tumor.

Intraoperative NIR Fluorescence Using the Visionsense System

All pituitary adenomas demonstrated some degree of a NIR fluorescent signal compared with background. Taken as a whole, the mean ± SD SBR was 1.9 ± 0.70 for all samples in all 15 patients. There was significant variability in the range of SBR (1–3.9; interquartile range 1.6–2.7). To explain this wide range in fluorescence, we performed a subgroup analyses using several variables: functional status (known preoperatively), distance from the endoscope to the pituitary gland (determined using the distance between the MOCRs as a proxy and known at time of surgery), time from injection (known at time of surgery), and FRα expression level (not known until after surgery was completed).

FRα Overexpression

Previous studies have demonstrated that NF pituitary adenomas overexpress FRα compared with functioning pituitary adenomas.8,12 Immunohistochemical analysis was performed on paraffin-embedded tissue sections using murine monoclonal antibodies to FRα, sections were examined by a single neuropathologist (M.M.L.), and H-scores (0–300) were determined (Fig. 3 and Table 2).

Three of the 6 NF adenomas demonstrated significant FRα overexpression with H-scores of 200, 225, and 270. These tumors were categorized as “high FRα tumors.” Among the 3 other patients with NF adenomas, 1 patient had a totally silent but growth hormone–staining adenoma. This tumor demonstrated an H-score of 100. In comparison, normal adenohypophysis has an H-score of 90 based on specimens that contain some small amount of normal pituitary gland tissue in the tissue block. Because 1 patient with an H-score of 100 was not significantly above the background H-score of the normal pituitary gland, we chose to categorize this patient as having a “low FRα tumor.” The other 2 NF adenomas were null cell adenomas and demonstrated virtually no FRα overexpression.

Among the 5 patients with corticotroph adenomas, all demonstrated low FRα expression, with H-scores ranging from 0 to 10. The 3 patients with somatotroph adenomas had H-scores ranging from 0 to 5, and 1 patient with somatocorticotroph adenoma had an H-score of 40.

Overall, only 3 patients were considered to have high FRα overexpression, and 12 patients were considered to have low FRα expression.

The NIR SBR of the patients with low H-scores was 1.6 ± 0.43, and the NIR SBR of the patients with high H-scores was 3.0 ± 0.27 (p < 0.01, t-test). Thus, the 3 patients with NF adenoma and high FRα overexpression (defined as > 200) did show significantly higher NIR SBR compared with the other 12 patients.

Functional Status

Six of the 15 patients had NF adenomas. One patient’s tumor positively stained for gonadotroph hormone on immunohistochemistry and 1 patient’s tumor positively stained for somatotroph hormone, but the preoperative laboratory values were not suggestive of hormonal abnormality. Because clinical hormonal status is the primary information that would be known at the time of resection, rather than pathological hormone staining, we plotted SBR versus functional status. The NIR SBR of NF adenomas (2.4 ± 0.81) was higher than the NIR SBR of functional adenomas (1.6 ± 0.48) (p < 0.05). However, because mean values are sensitive to extreme values (e.g., tumors with an FRα H-score > 200), we interpreted the data with a box plot to better evaluate the data (Fig. 4). As can be seen in Fig. 4, there is a significant overlap in the SBRs of functional and NF adenomas: the functional status of the adenoma alone did not explain the variability in SBR.

Fig. 4.
Fig. 4.

Box plots showing the medians and interquartile ranges of SBR based on whether the tumor was a known functioning or NF adenoma based on preoperative clinical data. Six NF adenomas showed a mean ± SD SBR of 2.4 ± 0.81 with a median of 2.5, and 9 hormone-secreting tumors demonstrated an SBR of 1.6 ± 0.48 with a median of 1.7. The difference between the 2 populations is close to statistical significance (p = 0.07), but there is significant overlap in the interquartile range. Figure is available in color online only.

Endoscope Distance

The signal intensity of light varies by the square of the distance, and as such we hypothesized that the endoscope distance to the sella may contribute to variability in SBR, especially within the tight confines of the nose and sella. In the absence of a reference standard, we measured the MOCR distance on the screen and divided it by the horizontal width of the video itself, which is captured as 720 line pairs in the horizontal dimension: MOCR ratio = (medial opticocarotid recess distance)/(horizontal video capture distance).

The MOCR fraction ranged from 0.33 (which is one-third of the screen, implying that the endoscope is relatively far from the sella) to 1.0 (which means the MOCR is filling the entire sella). Indeed, the MOCR fraction can actually be > 1.0 if the endoscope is introduced into the sella itself; however, we limited calculations of SBR to the more typical distances at which the sella is visualized.

MOCR was not a statistically significant univariate predictor of SBR in all patients (linear regression, p = 0.59). However, multivariate analysis, which takes FRα expression into consideration in the linear regression model, did identify statistical significance. In both adenomas overexpressing FRα and adenomas that do not, the endoscope distance from the adenoma influenced SBR (Figs. 57).

Fig. 5.
Fig. 5.

Case 12. The FRα H-score of 270 demonstrates strong fluorescence, regardless of endoscope distance. A and B: Sagittal (A) and coronal (B) preoperative MR images demonstrating a sellar mass. C and D: With MOCRs spanning approximately 30% of the field with white light (C), the tumor fluoresces with an SBR of 3.3 (D). E and F: With MOCRs spanning approximately 75% of the field with white light (E), the tumor fluoresces with an SBR of 3.9 (F). G and H: Visible light does not show convincing residual fluorescence (G), but NIR shows strong residual fluorescence in the left side of the sella (H). I and J: One-month postoperative sagittal (I) and coronal (J) MR images demonstrate a residual 1-cm hypoenhancing nodule in the left side of the sella.

Fig. 6.
Fig. 6.

Case 8. The FRα H-score of 5 demonstrates weak fluorescence that can be accentuated by moving the endoscope closer. A and B: Coronal (A) and sagittal (B) MR images demonstrating a sellar mass.C and D: With the MOCR spanning approximately 40% of the field with white light (C), the tumor fluoresces with an SBR of 1.1 (D). E and F: With the MOCR spanning approximately 70% of the field with white light (E), the tumor fluoresces with an SBR of 2.1 (F).

Fig. 7.
Fig. 7.

SBR can be increased by simply moving the endoscope closer to the adenoma. This effect can lead to a low FRα–expressing adenoma looking similar to a high FRα adenoma if careful attention is not paid. In both groups of adenomas, SBR increases as the distance between the endoscope and adenoma decreases (i.e., MOCR spans a greater percentage of the field). Figure is available in color online only.

Time From Injection to Visualization

Although the time required for OTL38 to saturate a tumor is proposed to be less than 30 minutes, its half-life in circulation is only approximately 25 minutes, suggesting that clearance of OTL38 from receptor-negative tissues should require 3–5 hours (i.e., 5 half-lives of OTL38 clearance should reduce the level of unbound imaging agent in circulation to < 5% of its initial level and the corresponding concentration of OTL38 in the tissues should soon follow). Because the fraction of OTL38 that is captured by a folate receptor will likely be internalized and retained by receptor-expressing cells, NF adenomas that express high levels of FRα should appear brightly fluorescent within approximately 30 minutes of injection and maintain their bright fluorescence for hours or even days.

In this study, the time from intravenous injection to initial surgical visualization of the adenoma ranged from 166 minutes to as long as 390 minutes, allowing sufficient time for most of the unbound OTL38 to clear from receptor-negative tissues. For the 3 patients with FRα overexpression (regression p value = 0.28) and 5 patients with NF adenomas (regression p value = 0.5), we did not find a significant time dependence for SBR, which is consistent with the rapid uptake of OTL38 by FRα-overexpressing cells.

However, in patients with low FRα expression, who should theoretically not demonstrate any significant NIR signal, we identified a linear trend toward higher SBR as time from injection increased (p = 0.0274; Fig. 8). We discuss this issue in the limitations section in Discussion.

Fig. 8.
Fig. 8.

The time from OTL38 injection to exposure and first imaging ranged from 166 to 372 minutes. The longer time period is correlated with increasing SBR in adenomas with low FRα expression (p = 0.0274). This suggests that there is an enhanced permeation and retention effect with the accumulation of dye in adenomas that do not overexpress FRα. Figure is available in color online only.

NIR as a Predictor of Postoperative MRI Results

All patients underwent postoperative MRI within 48 hours of surgery. Four of 15 patients showed residual neoplasm (Table 3). Excluding the patient in whom the laser malfunctioned in the middle of the procedure, 14 of 15 patients had intraoperative NIR images that were analyzed for fluorescence after excision of the main tumor. One patient demonstrated definite residual fluorescence and neoplasm on MRI, which was expected intraoperatively given the intracranial lateral extent of the neoplasm (Table 3 and Fig. 9). Her second-stage craniotomy performed 96 hours later demonstrated strong fluorescence even 4 days after injection, which is consistent with the anticipated internalization of the dye by the FRα-overexpressing adenoma cells. Two of 3 patients with cavernous sinus involvement seen on preoperative MRI exhibited residual neoplasm in the cavernous sinus or adjacent to the cavernous sinus on postoperative Day 1. In one of these patients (Case 12), the residual neoplasm was not seen using visible light alone, but a residual NIR signal was identified, thus predicting the findings on the follow-up MRI scan obtained 1 month later (Fig. 5).

Fig. 9.
Fig. 9.

Case 1. Comparison of MR images and intraoperative fluorescence of a pituitary adenoma before and after resection. A: Preoperative MR image showing a 25-mm mass in the sella and a 28-mm mass in the left middle cranial fossa. B: Postoperative MR image obtained after endoscopic surgery but before craniotomy, showing the incomplete resection of the sellar mass with cavernous sinus enhancement indicative of residual neoplasm. C and D: Preresection, intraoperative endoscopic images with NIR fluorescence off (C) and on (D), highlighting the sellar mass and surrounding bone and mucosa. E and F: Postresection, intraoperative endoscopic images with NIR fluorescence off (E) and on (F), showing residual fluorescence, especially to the patient’s left side. G and H: Preresection craniotomy and intraoperative NIR imaging with NIR fluorescence off (G) and on (H) during the second surgery obtained 96 hours after OTL38 injection to remove the mass in the cranial fossa.

For the 3 patients with high FRα expression, the postoperative MRI results correlated perfectly with the intraoperative imaging findings, and the residual fluorescence seen in 2 patients corresponded to the residual neoplasm seen on postoperative MRI (Table 3).

Using NIR to Detect Margins

On completion of standard adenoma resection, the Visionsense Iridium endoscope was introduced to inspect the margins of any potential residual adenoma. Excluding 4 patients from whom margin specimens were not taken, a total of 16 margin samples were collected from 11 patients. For all specimens, the true-positive rate was 86% and the true-negative rate was 89% using NIR, while the true-positive and true-negative rates were 80% and 89%, respectively, using visible light. If we exclude functioning adenomas and only examine the 6 patients with NF adenomas, a total of 5 margin samples were collected; for both NIR and visible light, the true-positive rate was 100% and the true-negative rate was 100%. Focusing on the 3 patients with FRα overexpression, a total of 3 margin samples were obtained. Two samples were positive for tumor, and these same 2 samples were also positive for fluorescence. One margin specimen was negative for fluorescence, and the final pathological analysis yielded only “fragments of crushed normal adenohypophysis.” Hence, the sensitivity, specificity, positive predictive value, and negative predictive value were 100% for these 3 patients.

Discussion

Despite advancements in the surgical approach and visualization using endoscopic and microscopic transsphenoidal techniques, GTR of pituitary adenomas remains a major surgical challenge. A recent published series documents a GTR rate of 50%–70%.16 A frequently studied intraoperative tool is MRI. Intraoperative MRI (iMRI) using a standard field (1.5-T MRI) has been shown to improve GTR rates. Berkmann et al. reported their experience with 109 patients, and demonstrated that after initial resection, GTR was seen in only 17%.2 Using this iMRI information, the authors proceeded to resect more pituitary adenomas, thus raising the GTR rate from 17% to 49% in their cohort of patients. Intraoperative MRI thus promises the user the ability to improve GTR, but the same authors also noted that although iMRI showed residual tumor, residual pituitary adenomas decreased in volume by an average of 90% in half of these patients on follow-up imaging.1 Hence, iMRI may result in false positives, and a recent guideline statement from the Congress of Neurological Surgeons concludes that iMRI is not currently recommended.16 Consequently, adjunct imaging is still needed during pituitary adenoma resection to help achieve GTR.

Pioneering work performed by Evans et al. demonstrated that FRα—a membrane receptor that transports reduced folate—is overexpressed in NF pituitary adenoma cells.7–9 Folate is an essential vitamin necessary for nucleic acid biosynthesis and is transported across the cell membrane in 3 ways: ubiquitously expressed reduced folate carrier, intestinal brush border–specific proton-coupled folate transporter, and folate receptor.27,39,40 Folate receptors exist in 4 forms—FRα, FRβ, FRγ, and FRδ—and FRα has been reported to be overexpressed in ovarian, lung, and breast carcinomas.3,20,33,35 In vivo SPECT imaging of folate receptor binding in healthy volunteers and patients has been performed using 99mTc-labeled etarfolatide, a folate-targeted technetium chelating agent, to demonstrate that folate receptors can be targeted for visualization.32,37

The ability to conjugate a folate molecule to a fluorescent contrast agent has been previously proposed to identify tumor nodules. Specifically, van Dam et al. published the first trial in humans that demonstrated intraoperative optical contrast during surgery for ovarian cancer using a folate-fluorescein (EC17) conjugate to detect intraoperative metastasis.36 Subsequently, Hoogstins et al. studied the use of OTL38 in ovarian carcinomas, which overexpress FRα.14 Based on these preliminary findings, we hypothesized that FRα overexpression in pituitary adenomas may provide a window of opportunity for neurosurgeons to visualize pituitary adenomas during surgery. Ultimately, we propose that this technology may be able to improve the GTR of pituitary adenomas.

NIR Visualization

In this study, we investigated the potential of using intraoperative, fluorescent, folate receptor targeting to improve the GTR of pituitary adenoma. For several reasons, we chose to use OTL38 (the NIR fluorescent, folate-conjugated dye) over EC17 (the visible light, fluorescent, folate-conjugated dye as used in the ovarian cancer trial published by van Dam et al.36). First, there is minimal autofluorescence in the NIR spectrum in CNS tissues. In addition, NIR has a longer wavelength, which can potentially provide a benefit because NIR light can penetrate the normal dura. In our prior work using Second Window ICG, where we deliver an extremely high dose of ICG intravenously and then wait 24 hours before surgical visualization, we have been able to demonstrate the NIR signal through normal dura and normal brain parenchyma.23 Another benefit of NIR is that our group has extensive experience using NIR fluorescence to study gliomas and meningiomas.22,23 One of the main challenges of working in the NIR spectrum is that humans cannot see light in the NIR spectrum. Indeed, most neurosurgeons have experience using NIR during vascular neurosurgery (for example, aneurysm clipping), and the image is only seen on the screen in black and white. In contrast, we chose to employ the Visionsense Iridium system, which simultaneously provides the real-time overlay of visible light and NIR as a “pseudocolor” map of the NIR signal superimposed on the visible light view (Figs. 1 and 2). This allowed the surgeon to work with the endoscope in visible light and then turn on NIR when necessary. The Visionsense Iridium system has been shown to have high sensitivity and a wide dynamic range, making it very suitable for our purposes.5

FRα as a Target for Optical Contrast

In this study, 3 of 6 patients (50%) with NF adenomas demonstrated FRα overexpression based on the immunohistochemical analysis of the paraffin-embedded specimens. The NIR SBR of the 12 adenomas with low FRα was 1.6 ± 0.43, and the SBR of the 3 adenomas with high FRα was 3.0 ± 0.27 (p < 0.01, t-test). Thus, by keeping the endoscope farther from the sella (MOCR < 50% of video screen) and using an SBR cutoff value of 2.5, we could perfectly predict FRα expression in all patients with an area under the receiver operating characteristic curve of 1.0 (Fig. 10). This could be used to predict adenomas that intraoperatively overexpress FRα.

Fig. 10.
Fig. 10.

SBR, as measured intraoperatively after exposing the adenoma through dura, is a strong predictor of FRα expression levels on pathological analysis. The adenomas that overexpress FRα (black triangles) and adenomas that do not (black rectangles) are easily distinguishable on intraoperative SBR, while 1 adenoma with SBR in the intermediate range (black circle) turned out to have intermediate FRα expression. Here, SBRs were measured with the MOCRs spanning 30%–50% of the screen.

An important finding of this pilot study is that the endoscope distance and injection-to-visualization time are critical. SBR can be artificially increased by simply bringing the endoscope closer to the tumor. Figure 7 demonstrates that SBR increased in all tumors as the endoscope moved closer to the tumor. We have found that maintaining an endoscope distance that is farther from the sella (where the MOCR spans less than half of the video screen) leads to the best ability to differentiate high and low SBR, which correspond to high and low FRα expression, respectively. There is obvious patient variability with respect to the distance between the right and left MOCRs, as reported by Nunes et al. (the average MOCR distance was 11 ± 1 mm).29 Nevertheless, in the absence of a laser range finder to provide the absolute distance, we used this internal landmark as a proxy for distance.

Clinical Benefit

GTR of an NF pituitary adenoma is often limited by intrinsic factors such as the circumferential invasion of the cavernous sinus (Knosp Grade 4 invasion).31 Berkmann et al. demonstrated that cytoreduction could be improved in the iMRI suite by taking the time to obtain an MRI scan and then performing further resection.2 Similarly, real-time intraoperative molecular imaging with an optical contrast agent such as folate-NIR dye may be able to replicate the iMRI findings in the subset of pituitary adenomas that overexpress FRα. We believe that intraoperative, folate receptor–targeted, fluorescent tumor imaging may provide important intraoperative information that influences surgical decision making and allows the surgeon to improve resection. In this limited series of 3 patients with high FRα overexpression, the intraoperative fluorescence perfectly predicted the postoperative MRI findings. In 1 patient (Case 7), no residual fluorescence was seen intraoperatively, and no residual tumor was seen on postoperative MRI (Fig. 2). In 2 patients (Cases 1 and 12), intraoperative fluorescence demonstrated a residual adenoma to the left of the sella, and in both cases MRI demonstrated residual adenoma on the left side (Figs. 5 and 10). This suggests a strong correlation between intraoperative fluorescent imaging and postoperative MRI.

With respect to margin detection, we believe that this technique has excellent sensitivity and specificity for patients with high FRα expression. In a margin analysis of 3 patients with high FRα expression, OTL38 showed major clinical value. Three biopsy specimens were obtained from 2 of the patients. There was a perfect correlation between intraoperative NIR fluorescence and the pathological analysis, demonstrating 100% sensitivity, specificity, positive predictive value, and negative predictive value. We recognize, of course, that these results must be extended to a larger population of patients with NF pituitary adenoma.

Limitations

A shortcoming of our study was the small number of patients who had FRα-overexpressing adenomas. Evans et al. concluded that all 11 NF adenomas demonstrated positive FRα overexpression on immunohistochemical analysis, and none of 7 functioning adenomas demonstrated this overexpression.7 In our study, none of 9 functioning adenomas demonstrated FRα overexpression, and only 3 of 6 (50%) of the NF adenomas demonstrated FRα overexpression. Because of this low FRα overexpression, we primarily focused this pilot study on predicting FRα overexpression using pathological immunohistochemical analysis as the gold standard. Without the proper identification of those tumors with high FRα overexpression, this technique could result in significant false-negative results. We will continue to expand this work in nonfunctioning adenomas, but there may also be a role for predicting which patients will benefit from this dye before surgery. Preoperative radionuclide imaging with 99mTc EC20 could be a viable option.19,32,37

Another limitation of this study is that we were only able to analyze 5 margin specimens of adenomas with an FRα H-score > 100 in order to compare visual and tactile discrimination of the tumor to OTL38 fluorescence, which showed that both approaches correctly discriminate neoplasms from normal tissue with 100% sensitivity and specificity. Future studies with larger sample sizes will further examine the validity of using OTL38 for margin discrimination.

A third limitation of the study is the possibility of false positives when using the NIR signal. Many pituitary adenomas in this study did not overexpress FRα and thus should not have demonstrated any significant NIR signal. Nevertheless, we believe that there may be a small accumulation of OTL38, even in the pituitary adenomas as shown in Figs. 6 and 8. This observation likely derives from the fact that the tumor vasculature of a pituitary adenoma is generally leakier than healthy tissue vasculature, leading to an enhanced permeation and retention effect in the tumor that allows some contrast agents to passively accumulate in the tumor tissue more effectively than in normal tissues.26 Because the magnitude of this effect is small, especially with low-molecular-weight contrast agents such as OTL38, significant contrast cannot be seen until the agent is cleared from normal tissues. Moreover, because the contrast agent is not receptor bound, it will wash out of the tumor over time. We suspect that adenomas with little or no FRα expression show increased SBR as time passes because the normal tissues clear OTL38 faster than the tumor. Thus, if OTL38’s injection-to-visualization time exceeds 5–6 hours, adenomas with low FRα expression may demonstrate high SBR, leading to false-positive findings.

Another consideration is that there are other folate receptor variants, such as FRβ, that are expressed on activated, but not quiescent, macrophages. We did not stain for FRβ in this study because we did not see macrophage accumulation in the slide blocks. Nevertheless, this is a potential source of OTL38 uptake that would result in fluorescence.

A final issue that needs to be considered is that there is a learning curve to intraoperatively performing video analysis of the NIR signal using the Visionsense Iridium 4-mm endoscope. Although the endoscope has a flat field of illumination in the visible light spectrum, the NIR excitation light source is brighter in the center than in the periphery. Thus, background points must be taken closer to the tumor rather than at the peripheries of the field. In addition, in the absence of a laser range finder to measure distance, approximating MOCR and screen size is not always easy if the surgeon has not opened the sphenoid sinus wide enough to visualize these anatomical structures.

Conclusions

In this paper, we present the first clinical study utilizing a folate receptor–targeted fluorescent dye for resection of pituitary adenomas. Administering the intravenous folate dye, OTL38, safely provided the optical contrast that allowed the discrimination of high and low FRα expression. We confirmed the findings of previous studies that reported only NF pituitary adenomas overexpress FRα. By just focusing on the 3 patients with confirmed FRα overexpression, OTL38 was able to predict postoperative MRI findings and provide 100% sensitivity and specificity of the margins. Although these results are encouraging, we consider this paper to represent an early-stage innovation. In our future work, we will focus only on NF adenomas in which this modality is the most likely to be clinically relevant. In addition, we will continue to refine the imaging technique and analysis as distance appears to be a critical factor.

Acknowledgments

This study was supported in part by the National Institutes of Health (grant no. R01 CA193556 awarded to S.S.) and the Institute for Translational Medicine and Therapeutics of the Perelman School of Medicine at the University of Pennsylvania (awarded to J.Y.K.L.). In addition, the research reported in this publication was supported by the National Center for Advancing Translational Sciences of the National Institutes of Health (award no. UL1TR000003 awarded to J.Y.K.L.). The content is solely the responsibility of the authors and does not necessarily represent the official views of the NIH.

Disclosures

The authors report the following. Dr. Lee owns stock options in Visionsense. Dr. Low has direct stock ownership in On Target Laboratories. Dr. Singhal holds patent rights over the technologies presented in this article.

Author Contributions

Conception and design: Lee, Grady. Acquisition of data: Lee, Cho, Zeh, Pierce, Martinez-Lage, Grady. Analysis and interpretation of data: Cho. Drafting the article: Cho. Critically revising the article: Lee, Adappa, Palmer, Newman, Learned, White, Kharlip, Snyder, Low, Singhal, Grady. Statistical analysis: Lee, Cho. Administrative/technical/material support: Lee, Low. Study supervision: Lee.

References

If the inline PDF is not rendering correctly, you can download the PDF file here.

Article Information

Correspondence John Y. K. Lee, University of Pennsylvania, 235 S Eighth St., Philadelphia, PA 19106. email: leejohn@uphs.upenn.edu.

INCLUDE WHEN CITING Published online August 25, 2017; DOI: 10.3171/2017.2.JNS163191.

Disclosures The authors report the following. Dr. Lee owns stock options in Visionsense. Dr. Low has direct stock ownership in On Target Laboratories. Dr. Singhal holds patent rights over the technologies presented in this article.

© AANS, except where prohibited by US copyright law.

Headings

Figures

  • View in gallery

    Visionsense equipment setup for NIR imaging. Upper: The excitation laser for NIR imaging is attached to a 4-mm endoscope. Copyright Visionsense. Published with permission. Lower: Images are displayed on the leftmost screen. Figure is available in color online only.

  • View in gallery

    Case 7. This patient had an NF adenoma, high FRα expression on final pathological analysis, and intraoperative NIR fluorescence. A and B: Preoperative sagittal (A) and coronal (B) MR images without contrast showing a 17-mm mass in the sella. C and D: Intraoperative endoscopic imaging with the dura open shown with white light only (C) and the NIR signal superimposed on white light (D). The MOCRs can be seen clearly and are used as a measure of distance from the endoscope to the target. E and F: Intraoperative endoscopic image of the sella after resection of the gross tumor. The residual tumor was identified on the right, both with white light only (E) and the NIR signal superimposed on the white light (F). This fluorescent specimen was confirmed on the final pathological analysis to be consistent with pituitary adenoma. G and H: Intraoperative endoscopic image of sella contents after resection of the residual neoplasm with white light only (G) and NIR superimposed on white light (H). No obvious fluorescence is identified. I: Example of SBR and MOCR calculation. Five points are chosen in the gross tumor itself, with the Visionsense camera calculating NIR signal for the points (75 + 74 + 66 + 61 + 57)/5 = 67 ± 7.9. Five points are chosen in the surrounding bone close to the tumor but outside of the obvious specimen (22 + 29 + 34 + 19 + 17)/5 = 24 ± 7.1. The SBR is thus 67/24 = 2.8. The black triangles denote the locations of the MOCRs, spanning roughly 50% of the field, that were used to approximate the endoscope distance.

  • View in gallery

    Immunohistochemical examination demonstrating the level of FRα expression in pituitary adenomas. A: Kidney is used as the positive control with an H-score of 300. Scale bar = 300 µm. B: Normal adenohypophysis with an H-score of 130. Scale bar = 600 µm. C and D: Case 12. Low-power (C) and high-power (D) views of the NF adenoma with a high H-score (270). Scale bar = 1 mm and 200 µm, respectively. E and F: Case 2. Low-power (E) and high-power (F) views of the somatotroph adenoma with an intermediate H-score of 100. Scale bar = 500 µm and 200 µm, respectively. G and H: Case 8. Low-power (G) and high-power (H) views of the somatotroph adenoma with a low H-score of 5. Scale bar = 600 µm and 200 µm, respectively. Figure is available in color online only.

  • View in gallery

    Box plots showing the medians and interquartile ranges of SBR based on whether the tumor was a known functioning or NF adenoma based on preoperative clinical data. Six NF adenomas showed a mean ± SD SBR of 2.4 ± 0.81 with a median of 2.5, and 9 hormone-secreting tumors demonstrated an SBR of 1.6 ± 0.48 with a median of 1.7. The difference between the 2 populations is close to statistical significance (p = 0.07), but there is significant overlap in the interquartile range. Figure is available in color online only.

  • View in gallery

    Case 12. The FRα H-score of 270 demonstrates strong fluorescence, regardless of endoscope distance. A and B: Sagittal (A) and coronal (B) preoperative MR images demonstrating a sellar mass. C and D: With MOCRs spanning approximately 30% of the field with white light (C), the tumor fluoresces with an SBR of 3.3 (D). E and F: With MOCRs spanning approximately 75% of the field with white light (E), the tumor fluoresces with an SBR of 3.9 (F). G and H: Visible light does not show convincing residual fluorescence (G), but NIR shows strong residual fluorescence in the left side of the sella (H). I and J: One-month postoperative sagittal (I) and coronal (J) MR images demonstrate a residual 1-cm hypoenhancing nodule in the left side of the sella.

  • View in gallery

    Case 8. The FRα H-score of 5 demonstrates weak fluorescence that can be accentuated by moving the endoscope closer. A and B: Coronal (A) and sagittal (B) MR images demonstrating a sellar mass.C and D: With the MOCR spanning approximately 40% of the field with white light (C), the tumor fluoresces with an SBR of 1.1 (D). E and F: With the MOCR spanning approximately 70% of the field with white light (E), the tumor fluoresces with an SBR of 2.1 (F).

  • View in gallery

    SBR can be increased by simply moving the endoscope closer to the adenoma. This effect can lead to a low FRα–expressing adenoma looking similar to a high FRα adenoma if careful attention is not paid. In both groups of adenomas, SBR increases as the distance between the endoscope and adenoma decreases (i.e., MOCR spans a greater percentage of the field). Figure is available in color online only.

  • View in gallery

    The time from OTL38 injection to exposure and first imaging ranged from 166 to 372 minutes. The longer time period is correlated with increasing SBR in adenomas with low FRα expression (p = 0.0274). This suggests that there is an enhanced permeation and retention effect with the accumulation of dye in adenomas that do not overexpress FRα. Figure is available in color online only.

  • View in gallery

    Case 1. Comparison of MR images and intraoperative fluorescence of a pituitary adenoma before and after resection. A: Preoperative MR image showing a 25-mm mass in the sella and a 28-mm mass in the left middle cranial fossa. B: Postoperative MR image obtained after endoscopic surgery but before craniotomy, showing the incomplete resection of the sellar mass with cavernous sinus enhancement indicative of residual neoplasm. C and D: Preresection, intraoperative endoscopic images with NIR fluorescence off (C) and on (D), highlighting the sellar mass and surrounding bone and mucosa. E and F: Postresection, intraoperative endoscopic images with NIR fluorescence off (E) and on (F), showing residual fluorescence, especially to the patient’s left side. G and H: Preresection craniotomy and intraoperative NIR imaging with NIR fluorescence off (G) and on (H) during the second surgery obtained 96 hours after OTL38 injection to remove the mass in the cranial fossa.

  • View in gallery

    SBR, as measured intraoperatively after exposing the adenoma through dura, is a strong predictor of FRα expression levels on pathological analysis. The adenomas that overexpress FRα (black triangles) and adenomas that do not (black rectangles) are easily distinguishable on intraoperative SBR, while 1 adenoma with SBR in the intermediate range (black circle) turned out to have intermediate FRα expression. Here, SBRs were measured with the MOCRs spanning 30%–50% of the screen.

References

TrendMD

Cited By

Metrics

Metrics

All Time Past Year Past 30 Days
Abstract Views 165 165 0
Full Text Views 451 451 27
PDF Downloads 242 242 19
EPUB Downloads 0 0 0

PubMed

Google Scholar