Identifying Isolated Glioblastoma Tissues in Human Patients through

Their Optical and Spectral Properties

Hussein Mehidine

1,5

, Fanny Poulon

, Ali Ibrahim

, Marjorie Juchaux

, Pascale Varlet

2,3

Bertrand Devaux

, Johan Pallud

3,4

and Darine Abi Haidar

1,5

Laboratory, UMR 8165-CNRS/IN2P3, Paris-Saclay University, 91405 Orsay, France

Neuropathology Department, Sainte-Anne Hospital, 75014 Paris, France

IMA BRAIN, INSERMU894, Centre de Psychiatrie et de Neurosciences, Paris, France

Neurosurgery Department, Sainte-Anne Hospital, 75014 Paris, France

Université Paris Diderot, Sorbonne Paris Cité, F-75013 Paris, France

Keywords: Glioblastoma, Scattering Coefficient, Absorption Coefficient, Endogenous Fluorescence.

Abstract: Survival rates and health-related quality of life of adult patients suffering from glioblastoma depend

significantly on the extent (no residual tumor tissue) and precision (no collateral damage) of the surgical

resection. Assistance in defining the borders of the infiltrating component of the glioblastoma would be

valuable to improve outcomes. A tissue can be defined by its optical properties : absorption, scattering,

intensity of fluorescence, that will give a unique signature. In this work we look at the absorption and

scattering coefficients of glioblastoma and control tissues from adult patients using an integrating sphere,

spectral measurements were also took on the samples using a fiber endoscope. The preliminary results show

the potential of using endogenous fluorescence for intraoperative identification of residual glioblastoma

tissue in the wall of the surgical cavity of resection.

1 INTRODUCTION

Glioblastoma (GBM) is the most common and most

aggressive malignant primary brain tumor in adults.

Following a magnetic resonance imaging (MRI)

analysis, its oncological treatment comprises 1) a

maximal safe resection encompassing the contrast-

enhanced tumor tissue, whenever feasible, 2) an

adjuvant treatment with combined radiotherapy and

concomittant chemotherapy followed by adjuvant

chemotherapy. The surgical resection technique is

limited by the difficulty to discriminate

intraoperatively between healthy tissue and tissue

infiltrated by isolated glioblastoma cells, mainly in

the wall of the surgical resection cavity, which

contain the infiltrated boundaries of the tumor. The

clinical concern is that, because glioblastoma

infiltration is not completly resected, the tumor

recure systematically and the patient will have to

return for a new operation which could lead to more

constraining side effects and reduce chances of

survival. The median overall survival is less than 18

months.

Indeed, glioblastoma is one of the most

infiltrating tumor, even if the solid tumor area is

easily detectable on MRI as a contrast-enhanced

tumor tissue, the infiltrated regions, which contains

active and isolated glioblastoma cells, are not

contrasted from healthy regions. These infiltrated

area have the same visual appearance as the healthy

ones, which makes them difficult to delineate.

Nowadays, the only technique that gives accurate

information on infiltrated area is the

histolopathogical analysis of a biopsy sample, which

is done ex-vivo and takes several days.

Recently, several intraoperative techniques have

been proposed to solve this problem, such as the

linear endomicroscope commercialized by Mauna

Kea technologies. This tool presents the advantages

of in-vivo and real time imaging. However, this

technique still not able to provide a multimodality of

measurements, which seems necessary to

discriminate healthy from tumoral tissues with a

strong relialability.

Our team develops a non-limear optical

endomicroscope, which is a new tool designed to the

surgical room. This intraoperative system will allow

136

Mehidine, H., Poulon, F., Ibrahim, A., Juchaux, M., Varlet, P., Devaux, B., Pallud, J. and Haidar, D.

Identifying Isolated Glioblastoma Tissues in Human Patients through Their Optical and Spectral Properties.

DOI: 10.5220/0006531501360140

In Proceedings of the 6th International Conference on Photonics, Optics and Laser Technology (PHOTOPTICS 2018), pages 136-140

ISBN: 978-989-758-286-8

the neurosurgeon to have a fast, reliable and

reproducible response on the nature of surrounding

tumor tissues, using several endogenous contrasts.

The multimodaly, regrouping Two Photons

Excitation (TPEF), Second Harmonic Generation

(SHG), spectral and fluorescence lifetime

measurements, was already proved as essential in

glioma tissue detection (Zanello et al., 2017).

In this context, it is necessary to identify all optical

properties of glioblastoma tumor, and compare them

with those of healthy tissues. In this work, many

optical parameters of frozen samples were found,

absorption and scattering coefficients using the

integrated sphere technique under 430 nm. The

spectral signature of endogenous fluorophores, using

linear excitation at 375 nm and 405 nm excitation

wavelength, were recorded.

2 MATERIALS AND METHODS

2.1 Samples

Samples were carried from the neurosurgery

department to the neuropathology laboratory at

Sainte-Anne hospital, Paris. The delay between the

end of the resection operation and the reception of

tissue was about fifteen minutes. After that, the

samples were stored at -80°C for few days. Then

they were cut at -18°C into 600 µm thickness

sections using a cryostat (Leica Microsystems) and

fixed with a 100% alcohol solution. Six

glioblastomas and seven healthy cortex samples

were analysed in this preliminary study.

2.2 Spectral, Transmittance and

Reflectance Study

The set-up shown in figure 1, is placed in the

neuropathology department at Sainte-Anne Hospital,

Paris-France. The setup is based on a linear

excitation fiber endoscope with three pulsed diode

lasers.

For integrating sphere measurements, the pulsed

diode laser emitting at 430 nm (Picoquant Germany,

LDH-P-C-430B, FWHM 70 ps) excitation

wavelength was used. Transmittance and reflectance

measurements of each sample were achieved using

an integrating sphere (Thorlabs, IS200-4). Its inner

surface is covered by a 99% reflective teflon, it have

four ports with a 12.7 mm diameter, and a fifth port

with a 3mm diameter to transfer the signal from the

integrating sphere to the spectrometer (Ocean optics,

HR2000) using an optical fiber (Ocean optics).

The average laser power used for this study was

around 5 mW. A diaphragm was placed after the

laser source to reduce the beam diameter to 1 mm

and the laser beam was pointed to five differents

Region Of Interest (ROI) in each specimen.

Transmittance and reflectance were measured for

each ROI, and finally the average of these five

values was calculated.

For spectral measurements, the excitation is

performed with the two others pulsed diode lasers

emitting at 405 nm (Picoquant-Germany, LDH-P-C-

405B, FWHM 60 ps,) and 375 nm (PicoQuant-

Germany, LDH-P-C-375B, FWHM 45ps) with a

maximal power of 1.1 mW. These diodes were

controlled with a driver (PicoQuant-Germany, PDL-

808 “Sepia”).

A specific Photonic Crystal Double-Clad Fiber

(PC-DCF) was used to excite and collect the

fluorescence signal (Ibrahim et al., 2016a, 2016b).

Collected signal is lead to the spectrometer (Ocean

Optics, QEPro 6500) through a long pass filter

(Semrock, SR420) to eliminate laser reflection.

Figure 1: Schematic representing the implemented setup,

including the integrating sphere and the spectral analysis.

3 DATA ANALYSIS

3.1 Optical Coefficients

The Inverse Adding Doubling (IAD) algorithm (S.

A. Prahl et al., 1993) was used to find the absorption

coefficient (µ

) and the reduced scattering

Identifying Isolated Glioblastoma Tissues in Human Patients through Their Optical and Spectral Properties

137

coefficient (µ’

) by refearing to the measured values

of the transmittance and the reflectance.

The scattering coefficient µ

is deduced using the

equation (1)

= µ’

/(1-g)

(1)

Where g is the anisotropy factor of the sample,

so we consider that g=0.89 for glioblastoma samples

and g=0.86 for cortex samples (Poulon et al., 2017).

This algorithm solves iteratively the radiative

transport equation until the numerical adjustment

and the experimental values of reflectance and

transmittance matches (Prahl et al., 1993). The

refractive index of the samples was not measured:

we consider that the refractive index n = 1.44 is the

same for all samples examinated by refering to the

literature (Bevilacqua et al., 1999).

3.2 Spectral Analysis

A Matlab code developed by our team and already

used in previous publications (Haidar et al., 2015)

(Zanello et al., 2017) was used to process spectral

data.

During spectral measurements, we used two

excitation wavelengths: (i) 405 nm to excite

efficiently five different endogenous fluorophores:

Nicotinamide Adenine Dinucleotide (NADH),

Flavins (FAD), lipopigments, porphyrins and

chlorines, and (ii) 375 nm to excite effectively the

NADH and FAD.

4 RESULTS

4.1 Optical Coefficients

The distribution of all scattering coefficient µ

values obtained in our measurements are shown in

Figure 2. The glioblastoma samples presented a

higher scattering coefficient than the cortex samples

and the glioblastoma samples had a higher

absorption coefficient than the cortex samples.

Figure 2: Distribution of the scattering coefficient values

(mm

-1

)(a) and the absorption coefficient values (mm

-1

) (b)

for glioblastoma and cortex tissues excited with 430 nm.

4.2 Spectral Measurements

In Figure 3, we compared the total fluorescence

intensity of glioblastoma and cortex samples, excited

with 405 nm (a) and 375 nm (b).

The difference in the spectral intensity was well

underlined between glioblastoma and cortex

samples. The emitted fluorescence from cortex

samples was 2.5 times higher than for glioblastoma

samples using a 405 nm excitation wavelength, and

2 times higher using 375 nm. We observed that

using 375 nm excitation wavelength we optimized

the NADH excitation and so the emitted

fluorescence signal was higher at this wavelength.

Using 405 excitation molecule we excited five

endogenous molecules (Zanello et al., 2016) as

shown in the spectral shape of figure 3.a.

PHOTOPTICS 2018 - 6th International Conference on Photonics, Optics and Laser Technology

138

Figure 3: Fluorescence spectra of glioblastoma and cortex

tissues excited with 405 nm (a) and 375nm (b)

wavelength.

5 DISCUSSION

In this preliminary study, we compared fixed cortex

and glioblastoma samples from adult patients using

two types of measurements. First, we observed that

the scattering and absorption coefficient are higher

in glioblastoma samples than in cortex samples. This

could be related to the fact that glioblastoma tissues

have denser vascularization and more collagen fibres

than cortex. This vascularization could be the source

of light scattering in such tissues. Added to the fact

that glioblastoma tissues contain neovascularization

which could affect the absorption coefficient.

By referring to our previous spectral studies on

rats brain tumors and healthy tissues (Haidar et al.,

2015) and on human brain tissues (Zanello et al.,

2017) using 405 nm excitation wavelength, the five

excited fluorophores are more concentrated in the

healthy tissues than in the glioblastoma tissues. For

that, the total fluorescence signal in cortex tissue is

higher than in glioblastoma. The same trend was

shown using 375 nm excitation. At this excitation

wavelength the absorption efficient section of

NADH and the FAD is higher than at 405nm

excitation wavelength. So, we optimise the

efficiency of excitation of this two endogenous

molecules and specially the excitation of NADH.

We can show also that endogenous fluorescence

from healthy tissues is higher than glioblastoma

tissues. This observation is of paramount important

as spectral response can discriminate healthy from

glioblastoma tissues. This is in accordance with the

literature of spectral analysis on freshly resected

tissues (Zanello et al., 2017).

Here, we show bimodal quantitative

measurements on the same tissues. These

measurements prove the power of optical detection,

based on endogenous fluorescence of frozen tissues

to discriminate healthy from tumoral tissues.

This opens a promising door in the detection of

tumors margins. It proved also how important is the

multimodality to detect with reliability the nature

tissues area, highlighting the importance of our

optical endomicroscope.

In the future studies, we will extend our cohort

by examining more types of human brain tumors and

also extend our study to freshly extracted samples.

ACKNOWLEDGEMENTS

This work is supported by a Plan Cancer with

Physicancer program grant “IMOP,” a “Défi

instrumental” program grant from CNRS, the Institut

National de Physique Nucléaire et de Physique des

Particules (IN2P3), and the “Ligue contre le cancer”.

REFERENCES

Bevilacqua, F., Piguet, D., Marquet, P., Gross, J.D.,

Tromberg, B.J., Depeursinge, C., 1999. In vivo local

determination of tissue optical properties:

applications to human brain. Appl. Opt. 38, 4939–

4950.

Haidar, D.A., Leh, B., Zanello, M., Siebert, R., 2015.

Spectral and lifetime domain measurements of rat

brain tumors. Biomed. Opt. Express 6, 1219–1233.

Ibrahim, A., Poulon, F., Habert, R., Lefort, C., Kudlinski,

A., Haidar, D.A., 2016a. Characterization of fiber

ultrashort pulse delivery for nonlinear

endomicroscopy. Opt. Express 24, 12515–12523.

Ibrahim, A., Poulon, F., Melouki, F., Zanello, M., Varlet,

P., Habert, R., Devaux, B., Kudlinski, A., Haidar,

D.A., 2016b. Spectral and fluorescence lifetime

endoscopic system using a double-clad photonic

crystal fiber. Opt. Lett. 41, 5214–5217.

Poulon, F., Mehidine, H., Juchaux, M., Varlet, P., Devaux,

B., Pallud, J., Abi Haidar, D., 2017. Optical

properties, spectral, and lifetime measurements of

Identifying Isolated Glioblastoma Tissues in Human Patients through Their Optical and Spectral Properties

139

central nervous system tumors in humans. Sci. Rep. 7.

doi:10.1038/s41598-017-14381-1.

Prahl, S.A., van Gemert, M.J., Welch, A.J., 1993.

Determining the optical properties of turbid media by

using the adding–doubling method. Appl. Opt. 32,

559–568.

Zanello, M., Poulon, F., Pallud, J., Varlet, P., Hamzeh, H.,

Lahoud, G.A., Andreiuolo, F., Ibrahim, A., Pages, M.,

Chretien, F., 2017. Multimodal optical analysis

discriminates freshly extracted human sample of

gliomas, metastases and meningiomas from their

appropriate controls. Sci. Rep. 7, 41724.

Zanello, M., Poulon, F., Varlet, P., Chretien, F.,

Andreiuolo, F., Ibrahim, A., Pallud, J., Dezamis, E.,

Abi‐Lahoud, G., Nataf, F., 2016. Multimodal optical

analysis of meningioma and comparison with

histopathology. J. Biophotonics.

PHOTOPTICS 2018 - 6th International Conference on Photonics, Optics and Laser Technology

140