Background

Confocal Laser Endomicroscope (CLE)

Laboratory Investigations

Clinical Investigations

Diagnostic accuracy

Intraoperative Results

CLE Digital Biopsies with Fluorescein sodium (FNa): Image Features

The Utilization of Intraoperative Confocal Laser Endomicroscopy During the

Fluorescence Guided Surgery for Brain Tumors Evgenii Belykh, MD, Claudio Cavallo, MD, Xiaochun Zhao, MD, Eric J. Miller, MS, Arpan A. Patel, MS, Nikolay L. Martirosyan, MD, PhD, Vadim A. Byvaltsev, MD, PhD, Robert F. Spetzler, MD, Jennifer Eschbacher, MD, Peter Nakaji, MD, Mark C. Preul, MD

Departments of Neurosurgery and Neuropathology, Barrow Neurological Institute, St. Joseph’s Hospital and Medical Center, Phoenix, Arizona; Department of Neurosurgery, Irkutsk State Medical University Irkutsk, Russia

1. Belykh E, Patel AA, Miller EJ, Bozkurt B, Yağmurlu K, Woolf EC, Scheck AC, Eschbacher JM, Nakaji P, Preul MC. Probe-based three-

dimensional confocal laser endomicroscopy of brain tumors: technical note. Cancer Manag Res. 2018 Aug 30;10:3109-3123. doi: 10.2147/CMAR.S165980. eCollection 2018. PubMed PMID: 30214304; PubMed Central PMCID: PMC6124793.

2. Belykh E, Cavallo C, Gandhi S, Zhao X, Veljanoski D, Izady Yazdanabadi M, Martirosyan NL, Byvaltsev VA, Eschbacher J, Preul MC, Nakaji P. Utilization of intraoperative confocal laser endomicroscopy in brain tumor surgery. J Neurosurg Sci. 2018 Aug 28. doi: 10.23736/S0390-5616.18.04553-8. [Epub ahead of print] PubMed PMID: 30160080.

3. Belykh E, Miller EJ, Patel AA, Yazdanabadi MI, Martirosyan NL, Yağmurlu K, Bozkurt B, Byvaltsev VA, Eschbacher JM, Nakaji P, Preul MC. Diagnostic Accuracy of a Confocal Laser Endomicroscope for In Vivo Differentiation Between Normal Injured And Tumor

Tissue During Fluorescein-Guided Glioma Resection: Laboratory Investigation. World Neurosurg. 2018 Jul;115:e337-e348. doi: 10.1016/j.wneu.2018.04.048. Epub 2018 Apr 17. PubMed PMID: 29673821.

4. Martirosyan NL, Eschbacher JM, Kalani MY, Turner JD, Belykh E, Spetzler RF, Nakaji P, Preul MC. Prospective evaluation of the utility

of intraoperative confocal laser endomicroscopy in patients with brain neoplasms using fluorescein sodium: experience with 74

cases. Neurosurg Focus. 2016 Mar;40(3):E11. doi: 10.3171/2016.1.FOCUS15559. PubMed PMID: 26926051. 5. Belykh E, Martirosyan NL, Yagmurlu K, Miller EJ, Eschbacher JM, Izadyyazdanabadi M, Bardonova LA, Byvaltsev VA, Nakaji P, Preul MC.

Intraoperative Fluorescence Imaging for Personalized Brain Tumor Resection: Current State and Future Directions. Front Surg. 2016 Oct 17;3:55. eCollection 2016. Review. PubMed PMID: 27800481; PubMed Central PMCID: PMC5066076.

Glioma Meningioma

Frozen section

Specificity 96% 98%

Sensitivity 95% 97%

Confocal laser endomicroscpoy

Specificity 94% 93%

Sensitivity 91% 97%

Tumor Microvasculature FNa Dosages

Identification of Nuclear Features

Intraoperative diagnosis in neurosurgery has traditionally relied on frozen and formalin fixed-paraffin-embedded section analysis of biopsied tissue samples. Although this technique still considered as the “gold standard” for establishing a histopathologic diagnosis, it entails a number of significant limitations including the time required for transferring, processing and interpreting the tissue, the presence of artifacts and sampling errors, as well as the differences present when comparing frozen and permanent section that may lead to misdiagnosis. Rapid intraoperative diagnosis has become possible with refinement and miniaturization of the research-type confocal laser scanning microscope into a handheld confocal laser endomicroscope (CLE). We report the evolution of this technology and present analysis of recent use of the updated CLE tool on patients during fluorescein sodium-guided brain tumor surgeries.

References

GL261-luc

tumor model;

Monitor growth with IVIS

Experimental Surgeries after IV FNa Injection

Investigation of CLE Optimal Imaging Parameters

(A) In vivo and rapid ex vivo confocal laser endomicroscopy (CLE) imaging. (B) Wide-field operative microscope imaging with Yellow 560 filter. (C) Generation 1 (Gen1) CLE image acquisition. Scale bar is 50 μm. (D, E) Benchtop laser scanning microscopy. Scale bar is 50 μm. (F) Z-stack volumetric reconstruction. (G) Histologic stained coronal brain slice.

Comparison of the CLE images taken with green bandpass (A, D), green longpass (B, E), and red (F) optical filters. Thick arrows signify postprocessing in FIJI, which includes despeckling and maximum brightness adjustment. (C) Matching regions of interest (ROI) were selected in FIJI to compare contrast between various structures in longpass and bandpass filters. ROIs were manually drawn over the brightest structures (n=5), over the surrounding gray background (n=3), and on the dark round structures (n=5).

This research was supported by the Newsome Chair in Neurosurgery Research held by Dr. Preul and by the Barrow Neurological Foundation. This project received materials support from Carl Zeiss AG, Oberkochen, Germany. Dr. Belykh received scholarship support SP-2240.2018.4. The authors thank the staff of Neuroscience Publications at Barrow Neurological Institute for assistance with poster preparation.

Acknowledgments

(A) Comparison of the tumor cellular features visible on CLE images with FNa and acridine orange (AO). (B)

Gradient of FNa distribution delineates cell contours with bright gray, while cell nuclei may be visible in some of the cells in a darker gray. AO staining of the same specimen shows true nuclei size. (C) Comparison of the nuclei size measured on the images with AO staining to the nuclei size measured on the images with the FNa staining (n = 106 cells measured with AO stain; n=53 nuclei and n = 52 cell diameters were measured with FNa stain). The average nucleus size determined on the FNa images was similar to the true average nucleus size based on the AO staining. Gen2 CLE image (B) shows anisocytosis, similar to (D) histologic findings.

Higher concentrations of FNa produces images with less noise on a generation 2 confocal laser endomicroscope. Images acquired with a low dose of FNa (A) could be postprocessed to improve brightness (B), so they appear similar to higher FNa dose images (C). Comparison of the enlarged regions in A, B, and C shows that cell contours are distinguished on all images; however, noise level (diagrams in A, B, and C) is less with a higher FNa concentration. (D) Subjective grading of overall CLE imaging quality is presented as a mean of all grades from 5 independent raters. The image quality is significantly better with higher FNa dosages (PANOVA=0.002).

CLE Tumor Margin Identification

Tumor border after intravenous FNa injection. (A) Generation 2 (Gen2) and (C) generation 1 (Gen1) confocal laser endomicroscopy images show a similar cellular architecture pattern of the GL261 glioma border region. (B) Hematoxylin and eosin (H&E) image of the tumor border from a matched sample. (D) Larger field of view ex vivo benchtop confocal image shows gradient of fluorescein diffusion from the tumor to the normal brain. Scale bar is 50 μm. LUT = look-up table.

Abnormal tumor vessels visualized with (A) generation 2 (Gen2) and (B) generation 1 (Gen1) confocal laser endoscopes (CLE) after fluorescein sodium injection. CLE images show similar tissue architecture pattern. (C) Hematoxylin and eosin (H&E)-stained tumor tissue shows tumor vasculature.

Z-stack imaging

Seventy-four consecutive adult patients (31 male and 43 female) were prospectively enrolled in the diagnostic accuracy study. There were 21 gliomas and 30 meningiomas. Multiple locations within the resection bed were imaged with CLE. in vivo The CLE specificity and sensitivity for gliomas were 94% and 91%, respectively. The frozen section specificity and sensitivity for gliomas were 96% and 95%, respectively. Both falce negative and false positive values were evident in situations of attempting to distinguish reactive changes and recurrent glioma. FN and FP values for meningiomas are associated only with distinguishing between meningioma and schwannoma (Martirosyan et al., 2016)

FNa IV 2-5 mg/kg

At the start of anesthesia

Tumor resection Yellow mode CLE digital biopsy

Fusiform tissue architecture, increased cell density, cells arranged in pseudocolumnar fashion. Cells appear anaplastic with pleomorphic nuclei.

Poorly demarcated nests without septal separation; The tissue appeared densely granulated on the CLE images with monomorphic cells that had irregular staining with fluorescein.

Hypercellular tissue with markedly atypical elongated tumor cells that appeared dark against a bright fluorescent background.

Histological CLE picture showed typical fibrous elongated cells arranged in whorls and nests. Collagen bands and psammoma bodies were frequently identified

High grade glioma Meningioma Choroid plexus carcinoma Pituitary adenoma

MR

I C

LE

Volu

me

vie

w

H&E

Low CLE contrast

CLE digital biopsy

Tumor resection

5 min

CLE digital biopsy

2011-2016 Initial

Protocol

2016-2018 Optimized Protocol

Good CLE quality

FNa IV 5 mg/kg

CLE digital biopsy

FNa IV 5 mg/kg

H&E

Analysis, Comparison

Our clinical experience with CLE includes 237 patients with gliomas, meningiomas and other CNS pathologies examined ex vivo and in vivo using a Gen1 CLE and 48 patients using a Gen2 CLE (since 2016: 19 HGG, 3 LGG, 11 pituitary adenomas, 2 craniopharyngiomas, 2 metastases, 2 schwannomas, 4 meningiomas, 2 treatment effects, 1 focal cortical dysplasia, 2 hemangioblastomas) examined ex vivo. The protocol was approved by the IRB of the Barrow Neurological Institute, St. Joseph’s Hospital and Medical Center, Phoenix, Arizona. Images were reviewed by a neuropathologist and 3 neurosurgeons. Since 2016 we optimized the protocol to work in combination with fluorescein-guided surgeries.

BACKGROUND

Fluorescence guidance with 5-aminolevulinic acid (5-ALA) increases resection extent and may benefit surgical outcomes in high-grade gliomas. We investigated if a combination of various fluorescence guidance techniques can provide further advantage. We systematically assessed and compared the fluorescent patterns of 5-ALA-induced protoporphyrin IX (PpIX), fluorescein sodium (FNa) and indocyanine green (ICG) to identify GL261 gliomas in mice and RFP-U251 gliomas in rats. METHODS

5-ALA (5mg), FNa (5 mg/kg) and ICG (20 mg/kg) were administered perioperatively. Fluorescence patterns were recorded with operative microscope, laser scanning confocal microscope, and confocal laser endomicroscope. Fluorescence was assessed quantitatively as a surface area of fluorescent positive tissue and qualitatively(false/true, positive/negative) as compared to HE-stained histology. RESULTS

FNa highlighted a larger surface area(18.0mm2) than white light (16.9mm2, p=0.016) or PpIX (16.0mm2, p=0.035). Both 5-ALA and FNa showed inhomogeneous staining patterns: multiple areas of equal staining, when PpIX was present and FNa was not, and vice versa. ICG was visible in 8/31 tumor samples, all immediately after injection. ICG did not reveal clear tumor margins, but stained non-tumor tissue and nearby vasculature. FNa signal was stronger (tumor to background ratio (TBR)1.93±0.56) compared to 5-ALA (1.52±0.31; p<0.002). ICG TBR (2.75±0.87) was higher than Fna (p<0.006) or 5-ALA (p<0.001). Cell-level imaging revealed PpIX-positive, FNa-negative invading tumor cell groupings and areas. However, some tumor parts were PpIX-negative and FNa-positive, or PpIX/FNa-negative. Differences in FNa-negative (8/30, 27%) and PpIX-negative (12/30, 40%) tumor areas were insignificant, p=0.27. False positive FNa fluorescence (13/30, 43%) was more frequent than 5-ALA (4/30, 13%; p<0.01). CONCLUSION

ICG highlighted tumor only within the first few minutes and stained mostly hypervascularized areas. Confocal and surgical imaging revealed inhomogeneous tumor border staining with PpIX/FNa. Simultaneous administration of 5-ALA and FNA may provide additional benefit. Neither ICG, 5-ALA or FNa worked perfectly, emphasizing the need for more specific markers for fluorescence-guided brain tumor resection.

Metabolic and Blood-brain Barrier Markers for Fluorescence-Guided Surgery: Systematic

High-Resolution Microscopy Investigation in Human Relevant Experimental Gliomas Evgenii Belykh, MD1,5, Alessandro Carotenuto, BA1, Laeth L. George, BS1, Eric J. Miller, BS1, Debbie R. Healey, BS4, Ernesto Luna Melendez, BS3,

Vadim A. Byvaltsev., MD, PhD5, Michael Staren1, Adrienne C. Scheck, PhD2, Chad Quarles, PhD4, Shwetal Mehta, PhD3, Peter Nakaji, MD1, Mark C. Preul, MD1 Departments of 1Neurosurgery, 2Neurooncology Research, 3Brain Tumor Research Center, and 4Imaging Research, Barrow Neurological Institute, St. Joseph’s Hospital and Medical Center, Phoenix, Arizona; 5Department of Neurosurgery, Irkutsk State Medical University Irkutsk, Russia

Methods Animals

• 10-12 week old Female B6(Cg)-Tyrc-2J/J mice • 10 week old nu/nu mice (Jackson Laboratory)

Mouse glioma model

• GL261 cells were made bioluminescent via Lentiphos HT System with Lenti-X HT Packaging Mix (Clontech Laboratories, Inc.) and FUW-GL plasmid (Dr J.B. Rubin, Washington University of St Louis, MO).

• RFP was cloned into pLXSN (Clonetech, Mountain View, California, USA), a mammalian retroviral expression vector. This was done to generate fluorescent U251 cells. A virus was then generated by using the RFP cDNA cloned into pLXSN and packaged into Phoenix A cells. U251 cells were infected with the virus and selected for RFP expression by fluorescence cell sorting.

• Glioma cells were injected intracranially 0.1 mm posterior to the bregma and 2.3mm right of midline

Animal surgery

• The animals were anesthetized • 5-ALA was injected intraperitoneally 2 hours before surgery (5 mg in 200

ul of 1x phosphate buffered saline). • Fluorescein Sodium (FNa 20 mg/kg) injected intravenously at the start of

procedure • ICG (20 mg/kg) injected intravenously at the start of procedure. • Brains were extracted and imaged fresh with operating microscope in

BLUE400 (for PpIX), YELLOW560 (for FNa) and INFRARED800 (for ICG) modes; The brains were also imaged fresh on the benchtop laser

scanning confocal microscope and handheld confocal laser

endomicroscope.

Results

A

B

488 reflectance PpIX FNa Merge

Figure 1. (A) Confocal laser endomicroscope image of FNa stained background so tumor cells appear dark. (B) Confocal laser scanning microscope image illustrating tumor vs. normal brain tissue border. Top half of each image are tumor cells, and the bottom half is normal brain tissue. PpIX stains cytoplasmic inclusions in both tumor and normal brain, with lower brightness in tumor. FNa provides background stain for the tumor cells, which is not visible in normal brain. This is further illustrated in the merge panel.

A PpIX showed tumor areas not highlighted by FNa

B FNa showed tumor areas not highlighted by PpIX

C FNa and PpIX highlighted similar tumor areas

D Surface Area of Tumor E Tumor to Background Ratio

0

0.5

1

1.5

2

2.5

3

3.5

4

4.5

5-ALA FNa ICG

p=.075

p<0.01

p<0.01

-10

-5

0

5

10

15

20

WL - 5-ALA WL - FNa FNa - 5-ALA

Su

rface a

rea

dif

fere

nce,

mm

2

p = 0.6

p = 0.02

p = 0.07

Figure 2: Top Panel (A-C): Fluorescence imaging of mouse gliomas as seen sequentially on the operative microscope with filters. First column – white light (WL); second column – BLUE400 filter for 5-ALA-induced PpIX; third column – Yellow560 filter for FNa; forth column – overlay of FNa and PpIX images. A. PpIX showed areas not highlighted by FNa (red arrows) B. FNa showed areas not highlighted by PpIX (yellow arrows) C. FNa and PpIX highlighted similar areas (green arrow). D. Difference in average surface of tumor area, as perceived under white light, FNa fluorescence, and PpIX fluorescence. N= 39 images analyzed in each group. Images taken at the same area with different filters of operative microscope. E. Tumor-to-background ratio of PpIX, FNa and ICG as a measure of fluorescence strength. Ratio is a measure of fluorescent intensity of the tumor compared to that of normal brain tissue, analyzed on ImageJ. N=48 images analyzed from operating microscope.

Hoechst PpIX FNa Merge

Tumor

Figure 3. Laser scanning microscopy image. 5-ALA-induced PpIX highlighted brain tumor cells distant to the core, while FNa did not. Top row has highlighted blue and yellow squares, corresponding to similarly outlined blue (second row) and yellow (third row) images, which are enlarged. In the middle row, FNa successfully highlighted the sample taken from the tumor core (blue square), but did not do so in the bottom row, corresponding to the sample distant to the tumor core.

488 reflectance PpIX FNa Merge RFP H&E

488 reflectance FNa Hoechst PpIX Merge RFP

C D

A

B

Figure 4. (A): Image illustrates individual tumor cells invading along vessels, as denoted by yellow arrows. 5-ALA-induced PpIX was able to visualize the tumor cells, which corresponded to the red fluorescent protein expression. (B) Laser scanning microscopy image illustrates tumor border showing red fluorescent protein labeled tumor cells invading normal brain area, and absence of PpIX or FNa fluorescence in/around of tumor cells. (C) Laser scanning microscopy shows invasions of tumor into the normal brain along the vessels highlighted by PpIX (D) H&E stain of representative brain showing invading tumor regions (yellow square) in the normal brain, similar to findings in (C).

PpIX FNa

Hoechst PpIX Merge

488 reflectance PpIX FNa Merge

488 reflectance PpIX FNa Merge

Figure 5. Image acquired via Laser Scanning Microscopy. The two image panels relate how PpIX accumulation in cells is visible under confocal microscope as time progresses. At early time points (top row), protoporphyrin localizes diffusely to the cytoplasm. At a later time point (middle row), PpIX is trafficked into a more granular distribution, most likely representing cellular mitochondria. Bottom row: thick slice image acquired via Laser Scanning Microscopy. Images illustrates PpIX staining resulting in a heterogeneous degree of PpIX production on the cellular level, as noted in the second image with varying degrees of red intensity in various cells.

488 reflectance Hoechst 5-ALA FNa Merge

Figure 6. Image acquired via Laser Scanning Microscopy. Tumor border pictured above. With a weak 5-ALA signal, the tumor cells are visualized, but normal cells are visualized as well.

7

15

26

18

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

5ala FNA

Overstain Frequency - Confocal

no overstain

overstain

10

18

23

15

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

5ala FNA

Stain Frequency - Confocal

stain

miss

14 11

34 37

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

5ala FNA

Stain Frequency – Wide Field

stain

miss

4

13

44

35

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

5ala FNA

Overstain Frequency – Wide Field

no overstain

overstain

Figure 7. Biopsy-based comparison of the incidence and percentage of successful staining and overstaining in PpIX vs FNa in operating wide field microscope (Kinevo) and confocal microscopy, counted as “stain", "misses", “overstain", “no overstain". Total n=48 images analyzed under Kinevo operative microscope, n=33 for confocal microscope. P=0.485 for Stain, p= 0.016 for overstain with operative microscope; P=0.046 for Stain, p=0.037 for Overstain with confocal microscope.

White light ICG

3 m

in

20 m

in

30 m

in

Figure 8. Representative images of mouse gliomas in white light view and Infrared 800 mode of the operating microscope. Upper row – brain was extracted within 3 minutes after intravenous ICG and FNa injection Inconsistent pattern of tumor labeling with ICG is observed, many vessels are highlighted and some tumor area is not highlighted with ICG, but was clearly fluorescent with FNa and 5-ALA. Middle row – retention of low ICG signal is visible 20 minutes after injection. Lower row – most of the brain slices exhibited no fluorescence when harvested after 30 minutes past ICG injection.

Figure 9. H&E confirmation of tumor cell mass (yellow arrow), distant to the main tumor mass above highlighted with PpIX but not with FNa (Similar to Fig. 3).

A

B

Hoechst PpIX FNa

RFP+ PpIX RFP + Hoechst Merge

RFP

Figure 10. Image acquired via laser scanning microscopy. Images illustrate significant heterogeneity in tumor core with various markers visualized and non-visualized in neighboring cells. A. Isolated laser scanning microscopy images of Hoechst, PpIX, FNa, and red fluorescent protein. B. Merged overlays from panel A.

This research was supported by the Newsome Chair in Neurosurgery Research held by Dr. Preul and by the Barrow Neurological Foundation. This project received materials support from Carl Zeiss AG, Oberkochen, Germany. Dr. Belykh received scholarship SP-2240.2018.4. The authors thank the staff of Neuroscience Publications at Barrow Neurological Institute for assistance with poster preparation.

Acknowledgments

Conclusions • ICG was only visible in tumor within the first few minutes after injection and almost

indistinguishable afterwards. Therefore, ICG is not a reliable marker for precise fluorescence guidance in our tumor models.

• Confocal and wide field imaging showed areas of tumor near the border that were better visible with FNa than PpIX along with areas where FNa was inferior to PpIX.

• Confocal LSM demonstrated that when PpIX had bright signal, the frequency of regions where it outperformed FNa was higher, compared to the frequency of regions where FNa outperformed 5-ALA.

• If aiming for a maximal total resection, simultaneous administration of 5-ALA and FNa may provide additional information on the tumor margin.

• Although PpIX (metabolic marker) and FNa (blood brain barrier marker) highlighted gross tumor bulk, neither worked perfectly in delineating tumor margins on the cellular level, highlighting the need for better markers for fluorescence guided brain tumor resection.

Abstract

Tumor

Normal brain

Operating microscope

Confocal laser endomicroscope Laser scanning microscope