CLINICAL ARTICLE

IntraoperatIve frozen section analysis is often used during pituitary surgery to establish a diagnosis and to assess margins. Frozen section analysis requires tissue

processing and has inherent limitations, such as prepara-tion time, freezing and sectioning artifact, and degradation or loss of tissue for permanent specimen studies.16,17 Each of these limitations is particularly relevant in cases of pitu-itary microadenomas, because only a small amount of tis-sue is available for analysis and surgical decision making may depend on the intraoperative diagnosis.

Confocal reflectance microscopy (CRM) is an imag-ing modality that generates cellular resolution images of biopsy specimens without the need for tissue processing, sectioning, or staining. Tissue samples can be transported directly from the operating room to the confocal micro-scope, and varying depths of the specimen can be ana-lyzed by changing the focal plane of the microscope. CRM therefore allows real-time analysis of cellular morphology and tissue architecture within a biopsy specimen, and it can provide rapid analysis of the tissue while preserving

ABBREVIATIONS CRM = confocal reflectance microscopy; NA = numerical aperture.SUBMITTED June 23, 2016. ACCEPTED November 21, 2016.INCLUDE WHEN CITING Published online May 26, 2017; DOI: 10.3171/2016.11.JNS161651.

Immediate ex-vivo diagnosis of pituitary adenomas using confocal reflectance microscopy: a proof-of-principle studyMichael A. Mooney, MD,1 Joseph Georges, DO, PhD,1 Mohammedhassan Izady Yazdanabadi, BS,1 Katherine Y. Goehring, PhD,2 William L. White, MD,1 Andrew S. Little, MD,1 Mark C. Preul, MD,1 Stephen W. Coons, MD,2 Peter Nakaji, MD,1 and Jennifer M. Eschbacher, MD2

Departments of 1Neurosurgery and 2Neuropathology, Barrow Neurological Institute, St. Joseph’s Hospital and Medical Center, Phoenix, Arizona

OBJECTIVE The objective of this study was to evaluate the feasibility of using confocal reflectance microscopy (CRM) ex vivo to differentiate adenoma from normal pituitary gland in surgical biopsy specimens. CRM allows for rapid, label-free evaluation of biopsy specimens with cellular resolution while avoiding some limitations of frozen section analysis.METHODS Biopsy specimens from 11 patients with suspected pituitary adenomas were transported directly to the pathology department. Samples were immediately positioned and visualized with CRM using a confocal microscope located in the same area of the pathology department where frozen sections are prepared. An H & E–stained slide was subsequently prepared from imaged tissue. A neuropathologist compared the histopathological characteristics of the H & E–stained slide and the matched CRM images. A second neuropathologist reviewed images in a blinded fashion and assigned diagnoses of adenoma or normal gland.RESULTS For all specimens, CRM contrasted cellularity, tissue architecture, nuclear pleomorphism, vascularity, and stroma. Pituitary adenomas demonstrated sheets and large lobules of cells, similar to the matched H & E–stained slides. CRM images of normal tissue showed scattered small lobules of pituitary epithelial cells, consistent with matched H & E–stained images of normal gland. Blinded review by a neuropathologist confirmed the diagnosis in 15 (94%) of 16 images of adenoma versus normal gland.CONCLUSIONS CRM is a simple, reliable approach for rapidly evaluating pituitary adenoma specimens ex vivo. This technique can be used to accurately differentiate between pituitary adenoma and normal gland while preserving biopsy tissue for future permanent analysis, immunohistochemical studies, and molecular studies.https://thejns.org/doi/abs/10.3171/2016.11.JNS161651KEY WORDS adenohypophysis; confocal reflectance microscopy; frozen section analysis; pituitary adenoma; rapid diagnosis; tissue architecture; pituitary surgery

©AANS, 2017 J Neurosurg May 26, 2017 1

M. A. Mooney et al.

J Neurosurg May 26, 20172

the entire specimen for future permanent-section staining, molecular analysis, and biobanking. This technology is currently used in several fields other than neurosurgery to rapidly assess biopsy specimens and guide surgical deci-sion making.2,6,7,14,15

In this study, we evaluated the utility of CRM for rapid assessment of pituitary specimens in the immediate ex-vi-vo period as a proof-of-principle study. To our knowledge, this is the first description of CRM for this application in the medical literature to date.

MethodsBiopsy Specimens

Patients with CNS tumors undergoing surgery at our institution between February 2014 and November 2015 were prospectively enrolled in our study. This study was approved by the IRB of St. Joseph’s Hospital and Medical Center, in Phoenix, Arizona. Fresh human biopsy speci-mens were collected in the operating room and transport-ed to the Department of Pathology, where both CRM and diagnostic frozen sections were performed, when appli-cable. CRM images were collected by team members with training in confocal microscopy (K.Y.G., J.G., and M.I.Y.). Eleven patients who underwent transnasal transsphenoi-dal surgery for suspected pituitary adenoma during this period were included in the study.

CRM ImagingFor CRM imaging, our methodology has been previ-

ously described.5 Briefly, biopsy specimens were placed in ice-cold saline in uncoated no. 1.5 glass-bottom dishes (MatTek Corp.). No tissue processing or staining was per-formed prior to CRM imaging. CRM imaging was per-formed with a Zeiss inverted LSM 710 laser-scanning confocal microscope (Carl Zeiss AG). Images were ac-quired with a magnification of 20×/0.8 NA (numerical aperture) air objective and a magnification of 40×/1.2 NA water immersion objective. CRM images were acquired by raster-scanning the sample with a 633-nm diode laser and collecting reflected photons of the same incidence wavelength. The confocal aperture was set to 1 Airy unit for all imaging. The laser and gain values were set to fill the dynamic range of the photomultiplier tube, and the frame size was set to sample at the Nyquist sampling rate (i.e., cycles per unit distance). Images were collected in an 8-bit format. Simulated low-magnification images were generated by using the Zeiss tiling function, which pro-duces a large-volume image by stitching together multiple highly resolved and magnified individual images. After CRM, biopsies were marked and processed for standard histopathological assessment with H & E staining. Frozen sections were performed when an intraoperative diagno-sis was required. CRM images were assessed for pathog-nomonic features and compared with the corresponding H & E images. All image processing was performed using linear functions within ImageJ (US NIH).

Specimen AnalysisAll biopsy specimens from the 11 patients were com-

pared on the basis of their CRM, frozen section, and

H & E–stained features by a neuropathologist (J.M.E.), who noted distinguishing features for descriptive purpos-es. For the blinded portion of the study, 16 representative CRM images from the 11 cases were selected by the neu-ropathologist (J.M.E.): 7 images of normal adenohypophy-sis and 9 images of pituitary adenoma. These images were presented in a blinded fashion to a second dedicated neu-ropathologist (S.W.C.), who had no prior knowledge of the cases.

ResultsDescriptive Features

CRM images of pituitary adenoma demonstrated sheets of cells with prominent nuclei and brightly reflec-tive cytoplasm. Increased cellularity, tissue architecture, vascularity, and stroma were all readily visualized and were similar to the matched H & E–stained slides (Fig. 1). CRM images of normal pituitary gland showed smaller, more uniform nuclei, as well as scattered small lobules of pituitary epithelial cells, consistent with matched H & E–stained images of normal gland (Fig. 2). Pituitary adeno-ma subtypes included 4 gonadotroph adenomas, 2 corti-cotroph adenomas, 1 prolactinoma, and 1 alpha-subunit secreting adenoma; 3 biopsy specimens were from normal pituitary gland.

Blinded ReviewFifteen (94%) of the 16 CRM images were correctly

identified as either normal pituitary adenohypophysis or pituitary adenoma by the second neuropathologist, who had no prior knowledge of the cases. This neuropatholo-gist incorrectly labeled 1 image of normal gland as ad-enoma.

DiscussionWith the increasing trend toward personalized medi-

cine, the importance of patient-specific tumor analysis and molecular characterization of biopsy specimens will continue to expand. This movement requires not only op-timized tumor specimens for analysis but also research dedicated to the genetic and molecular characteristics of individual patient biopsy specimens. Indeed, the charac-terization of pituitary adenomas on the basis of genetic and molecular characteristics, and the link between these markers and clinical behavior, has grown exponentially within the past decade.3,8,11,18 Given the recent advances in our knowledge in these fields, biopsy specimen pres-ervation and biobanking for research studies are arguably more important now than at any time in the past.

The current standard for intraoperative diagnosis is frozen section analysis, which is limited by prolonged processing time, freezing and sectioning artifact, and deg-radation or loss of tissue for permanent specimen stud-ies.16,17 When an intraoperative diagnosis is required and only a small biopsy specimen is available, such as in the case of a pituitary microadenoma, frozen sectioning may jeopardize the ability to perform definitive molecular or genetic evaluation in the future. The ability to obtain an

Confocal reflectance microscopy for pituitary adenoma

J Neurosurg May 26, 2017 3

intraoperative diagnosis with CRM without sacrificing a biopsy specimen circumvents this problem and facilitates definitive diagnostic studies and further molecular char-acterization.

Confocal microscopy is an imaging modality that gen-erates high-resolution images by spatially filtering visual-ized photons through a pinhole.4 This technique allows for the “optical sectioning” of a specimen, wherein various depths of a biopsy sample can be visualized without its physical manipulation. The use of confocal microscopy has also been coupled with the use of fluorescent dyes to achieve rapid histopathological diagnosis in numerous fields, including gastroenterology, dermatology, gynecol-ogy, and ophthalmology.2,6,7,14,15 Furthermore, confocal microscopy has been used intraoperatively with intrave-nous dyes to visualize tumors of the CNS in vivo.1,9,10, 12,13 In contrast, CRM utilizes incident light, rather than high-energy fluorescence, and it provides a label-free assess-ment of biopsy specimens without tissue manipulation. We recently reported on the utility of CRM for assessment of the cellularity of high-grade glioma specimens prior to biobanking, as well as on the preservation of molecular in-tegrity within biopsy specimens with this technique.5 The goal of the current study was to expand upon the utility of this emerging technology and to outline its particular advantages for use in the diagnosis of pituitary adenoma specimens.

Several key aspects of pituitary adenoma histopathol-ogy allow for its definitive diagnosis on frozen-section and permanent histology. Pituitary adenomas grow as sheets of proliferative cells and demonstrate increased cellu-larity compared with sections of normal pituitary gland (Figs. 1 and 2). These key features were readily observed on CRM after specimens were transferred directly from the operating room. Although CRM does not reach the level of cellular resolution and tissue contrast observed with H & E staining, features necessary for diagnosis of normal pituitary versus adenoma were apparent to both neurosurgeons and neuropathologists. The ability of CRM to reveal the diagnosis rapidly and preserve the entire specimen for further permanent sectioning and molecu-lar analysis prompted us to perform this study and report these findings.

In this proof-of-principle study, we tested a second neu-

ropathologist with no prior knowledge of the cases and limited CRM experience (S.W.C.). He was able to cor-rectly distinguish between pituitary adenoma and nor-mal gland in 15 (94%) of 16 CRM images. One image of normal gland was labeled as adenoma in this portion of the study; however, the remainder of the images were correctly identified. It is important to note that this study did not address the sensitivity and specificity of CRM for pituitary adenoma diagnosis, and future studies should be designed to address these outcomes.

Although our results demonstrate the utility of this novel technology, the study does have limitations. The main limitation is that all 11 specimens were evaluated prospectively by the senior author (J.M.E.) and 16 repre-sentative images were selected for blinded review by the second neuropathologist (S.W.C.). While this study design demonstrates the utility of CRM for intraoperative adeno-ma diagnosis, it does not provide a direct comparison with frozen sectioning. Future blinded studies of CRM com-pared with frozen sectioning for intraoperative pituitary adenoma assessment are warranted.

The study also did not assess the time to diagnosis for intraoperative CRM. Although our recent experience us-ing CRM for this application demonstrates a substantial savings in time compared with frozen sectioning, diagno-sis by CRM is associated with a learning curve, and analy-sis of early specimens required substantially more time than specimens later in our series. This learning curve precluded our ability to include time to diagnosis as an outcome in this study. Future studies of CRM for pituitary adenoma should include this outcome measure when com-paring this technology to frozen sectioning.

ConclusionsCRM is a simple and reliable method for the rapid

evaluation of pituitary adenoma specimens ex vivo and offers several advantages over the current practice of fro-zen sectioning. This technique can accurately distinguish between pituitary adenoma and normal gland, while pre-serving the entire specimen for further permanent analy-sis, immunohistochemical studies, and molecular studies. This technology may have an increasing role in pituitary adenoma surgery as the need for genetic and molecular characterization of individual tumors increases.

FIG. 1. Photomicrographs of a pituitary adenoma. CRM image (left) and H & E stain (right) of the same tumor both demonstrate sheets of cells with prominent nuclei. In the CRM image (left), horizontal streak artifacts at the top and bottom of the image are the result of image “tiling” during composition. Original magnification ×20.

FIG. 2. Photomicrographs of a normal pituitary gland. CRM image (left) showing smaller, more uniform nuclei, as well as scattered small lobules of pituitary epithelial cells, consistent with matched H & E stain (right) of normal gland. Original magnification ×20.

M. A. Mooney et al.

J Neurosurg May 26, 20174

AcknowledgmentsFunding support for this study was provided by the Barrow Neu-rological Foundation.

References 1. Behbahaninia M, Martirosyan NL, Georges J, Udovich JA,

Kalani MY, Feuerstein BG, et al: Intraoperative fluorescent imaging of intracranial tumors: a review. Clin Neurol Neu-rosurg 115:517–528, 2013

2. Benati E, Argenziano G, Kyrgidis A, Moscarella E, Ciardo S, Bassoli S, et al: Melanoma and naevi with a globular pattern: confocal microscopy as an aid for diagnostic differentiation. Br J Dermatol 173:1232–1238, 2015

3. Chen Z, Li Z, Chang Y, Ma L, Xu W, Li M, et al: Relation-ship between NF-kB, MMP-9, and MICA expression in pituitary adenomas reveals a new mechanism of pituitary adenomas immune escape. Neurosci Lett 597:77–83, 2015

4. Fellers TJ, Davidson MW: Introduction to confocal micros-copy. Olympus Microscopy Resource Center. (http://olym-pus.magnet.fsu.edu/primer/techniques/confocal/confocalin-tro.html) [Accessed January 26, 2017]

5. Georges J, Zehri A, Carlson E, Nichols J, Mooney MA, Martirosyan NL, et al: Label-free microscopic assessment of glioblastoma biopsy specimens prior to biobanking. Neuro-surg Focus 36(2):E8, 2014 [Erratum in Neurosurg Focus 36(6):Erratum, 2014]

6. Hartmann D, Ruini C, Mathemeier L, Dietrich A, Ruzicka T, von Braunmühl T: Identification of ex-vivo confocal scan-ning microscopic features and their histological correlates in human skin. J Biophotonics 9:376–387, 2016

7. Kang D, Schlachter SC, Carruth RW, Kim M, Wu T, Taba-tabaei N, et al: Comprehensive confocal endomicroscopy of the esophagus in vivo. Endosc Int Open 2:E135–E140, 2014

8. Manojlovic-Gacic E, Skender-Gazibara M, Popovic V, Sol-datovic I, Boricic N, Raicevic S, et al: Oncogene-induced senescence in pituitary adenomas—an immunohistochemical study. Endocr Pathol 27:1–11, 2016

9. Martirosyan NL, Georges J, Eschbacher JM, Cavalcanti DD, Elhadi AM, Abdelwahab MG, et al: Potential application of a handheld confocal endomicroscope imaging system using a variety of fluorophores in experimental gliomas and normal brain. Neurosurg Focus 36(2):E16, 2014

10. Mooney MA, Zehri AH, Georges JF, Nakaji P: Laser scan-ning confocal endomicroscopy in the neurosurgical operat-ing room: a review and discussion of future applications. Neurosurg Focus 36(2):E9, 2014

11. Saeger W, Petersenn S, Schöfl C, Knappe UJ, Theodoro-poulou M, Buslei R, et al: Emerging histopathological and genetic parameters of pituitary adenomas: clinical impact

and recommendation for future WHO classification. Endocr Pathol 27:115–122, 2016

12. Sanai N, Eschbacher J, Hattendorf G, Coons SW, Preul MC, Smith KA, et al: Intraoperative confocal microscopy for brain tumors: a feasibility analysis in humans. Neurosur-gery 68 (2 Suppl Operative):282–290, 2011

13. Sanai N, Snyder LA, Honea NJ, Coons SW, Eschbacher JM, Smith KA, et al: Intraoperative confocal microscopy in the visualization of 5-aminolevulinic acid fluorescence in low-grade gliomas. J Neurosurg 115:740–748, 2011

14. Su P, Liu Y, Lin S, Xiao K, Chen P, An S, et al: Efficacy of confocal laser endomicroscopy for discriminating colorectal neoplasms from non-neoplasms: a systematic review and meta-analysis. Colorectal Dis 15:e1–e12, 2013

15. Tan J, Quinn MA, Pyman JM, Delaney PM, McLaren WJ: Detection of cervical intraepithelial neoplasia in vivo using confocal endomicroscopy. BJOG 116:1663–1670, 2009

16. Tilgner J, Herr M, Ostertag C, Volk B: Validation of intraop-erative diagnoses using smear preparations from stereotactic brain biopsies: intraoperative versus final diagnosis—influ-ence of clinical factors. Neurosurgery 56:257–265, 2005

17. Uematsu Y, Owai Y, Okita R, Tanaka Y, Itakura T: The use-fulness and problem of intraoperative rapid diagnosis in sur-gical neuropathology. Brain Tumor Pathol 24:47–52, 2007

18. Välimäki N, Demir H, Pitkänen E, Kaasinen E, Karppinen A, Kivipelto L, et al: Whole-genome sequencing of growth hormone (GH)-secreting pituitary adenomas. J Clin Endo-crinol Metab 100:3918–3927, 2015

DisclosuresDr. Eschbacher has received clinical or research support for the study described from the Barrow Neurological Foundation.

Author ContributionsConception and design: Eschbacher, Mooney, Georges, White, Little, Nakaji. Acquisition of data: Eschbacher, Mooney, Georges, Yazdanabadi, Goehring, Coons. Analysis and interpretation of data: Eschbacher, Mooney, Preul, Nakaji. Drafting the article: Eschbacher, Mooney, Georges, Nakaji. Critically revising the article: Nakaji. Reviewed submitted version of manuscript: Esch-bacher. Administrative/technical/material support: Goehring, Preul. Study supervision: Eschbacher.

CorrespondenceJennifer M. Eschbacher, c/o Neuroscience Publications, Barrow Neurological Institute, St. Joseph’s Hospital and Medical Cen-ter, 350 W Thomas Rd., Phoenix, AZ 85013. email: neuropub@dignityhealth.org.