LIBS based Tissue Differentiation for Er:YAG Surgical Laser

Fanuel Mehari

1,2

, Benjamin Lengenfelder

1,2

, Robert Figura

, Florian Kl

ampﬂ

and Michael Schmidt

1,2

Friedrich-Alexander-Universit

at Erlangen-N

urnberg (FAU), Institute of Photonic Technologies (LPT),

Konrad-Zuse-Straße 3/5, 91052 Erlangen, Germany

Erlangen Graduate School in Advanced Optical Technologies (SAOT), Paul-Gordan-Straße 6, 91052 Erlangen, Germany

Keywords:

LIBS (Laser-induced Breakdown Spectroscopy), Biophotonics, Laser Surgery, Tissue Differentiation, Er:YAG

Laser.

Abstract:

The analytical technique laser-induced breakdown spectroscopy (LIBS) is becoming an attractive technique in

the ﬁeld of medicine. One emerging application is the differentiation of biological tissues in real-time during

laser surgery. This work attempts to further investigate the use of LIBS together with a surgical Er:YAG laser.

The main goal here is to investigate the effect of body ﬂuids as potential contaminants during LIBS based

differentiation of soft tissues. Furthermore, the work attempts to exploit the use of the surgical laser as a

cleaning laser and compares the effect when only the LIBS laser is used for tissue differentiation. The study

shows that body ﬂuids have a signiﬁcant inﬂuence on the LIBS spectra and that a surgical laser might serve as

an in-vivo cleaner leading to improved tissue classiﬁcation during laser surgery.

1 INTRODUCTION

Lasers have become one of the most important tools

used in modern medicine. Due to the unique in-

trinsic properties of laser light, such as monochro-

maticity, focusability and high intensity, lasers are

being used in general surgery (Walter, 1999) as well

as in specialized disciplines like dermatology (Gold-

berg., 2013), and ophtalmology (Solomon, 2009).

One of the most common lasers used as laser scalpel

is the Er:YAG since it offers very high absorption

in water and thus efﬁcient processing for all tissue

types. The advantages laser surgery brings about

include providing a sterile surgery technique along

with highly precise treatment areas with little to no

thermic damage. However, the laser scalpel lacks any

feedback including tactile information and it becomes

difﬁcult to operate without inﬂicting damage to crit-

ical tissues such as muscle or nerves. This problem

leads to the investigation of various feedback mod-

alities providing tissue discrimination for laser sur-

gery. As a consequence, optical based techniques

for tissue discrimination have gained signiﬁcant in-

terest. One of the optical techniques that is drawing

attention for a feedback system is LIBS. LIBS uses

a pulsed laser to generate a plasma plume which va-

porizes a small volume of a given specimen. Using

a spectrometer, the plasma decay emission spectrum

is analyzed and then used to identify and quantify the

constituent elements the specimen is made of (Cre-

mers and Radziemski., 2013). LIBS has been suc-

cessfully used to analyze various kinds of tissues un-

der ex-vivo conditions, including biological tissues

such as bone and cartilage (Mehari, 2014). Neverthe-

less, the fact that in-vivo conditions profoundly dif-

fer from those outside the body has to be taken into

consideration. Tissue surfaces inside the body can be

contaminated with body ﬂuids containing various ele-

ments that could have an inﬂuence on the obtained

LIBS spectra.

2 OBJECTIVES

In this work, the inﬂuence of simulated in-vivo con-

ditions (by putting a thin saline layer on top of the tis-

sue samples) on the LIBS classiﬁcation is ﬁrst invest-

igated without the usage of a surgical laser (single-

pulse). In addition, it considers the use of a con-

ventional surgical Er:YAG laser as a cleaning instru-

ment to free the tissue’s surface from liquid contam-

inants prior to LIBS analysis as one possible way to

minimize the inﬂuence of body ﬂuids (double-pulse).

First, the LIBS spectra of contaminated and dry fat

and muscle tissue is evaluated without the usage of

Mehari, F., Lengenfelder, B., Figura, R., Klämpﬂ, F. and Schmidt, M.

LIBS based Tissue Differentiation for Er:YAG Surgical Laser.

DOI: 10.5220/0006637402470251

In Proceedings of the 6th International Conference on Photonics, Optics and Laser Technology (PHOTOPTICS 2018), pages 247-251

ISBN: 978-989-758-286-8

247

a Er:YAG laser. Second, a conventional surgical

Er:YAG laser is used in combination with LIBS ana-

lysis for contaminated and dry fat and muscle tissue.

3 MATERIALS AND METHODS

3.1 Single-pulse LIBS

Figure 1 shows a schematic overview of the setup

used for single-pulse. The experimental setup essen-

tially consists of three major components and sev-

eral auxiliary devices. The centerpiece is a Nd:YAG

(532 nm) laser whose laser beam is focused onto the

tissue samples. Light signals from the LIBS plasma

are collected using an array of convex lenses and

an optical ﬁber cable that is connected to a spectro-

meter for detecting and analyzing the signals. The

array of convex lenses for LIBS signal collection

is focused onto the tissue surface and aligned with

the laser focus. For signal transport to the spectro-

meter a multimode UV grade ﬁber cable was used.

Spectral analysis is performed with an Echelle spec-

trograph (Andor Mechelle 5000) of resolving power

∆λ

= 6000. The tissue samples were placed in a petri

dish that was put on a XYZ-translation stage to be

able to move around the sample in 3D.

Figure 1: Optical setup for single-pulse LIBS.

3.2 Double-pulse LIBS

An adjusted setup is implemented for LIBS experi-

ments following a laser pulse of a surgical Er:YAG

laser (2.94 µm, Glissando, WaveLight(T M)), Erlan-

gen, Germany). The Nd:YAG laser in the single-pulse

experiments was used. The Er:YAG laser is operated

at a repetition rate of 10 Hz with a pulse duration of

350 µs and a pulse energy of 200 mJ. To use it for

LIBS experiments in this study, the laser head is ﬁxed

steady next to the moving stage and directed onto the

tissue sample in a 45

◦

angle. The Er:YAG laser’s fo-

cus lies slightly above the sample surface. Moreover,

it is aligned with the focus of the Nd:YAG laser to

ensure that both lasers hit the same area on the tissue

surface for one measurement. In order to synchron-

ize the two lasers, a pulse detector for infrared light

(DoroTek Lab Bench Detector, 2 − 12µm) is used to

detect the reﬂected light of the Er:YAG laser beam in-

cident on the sample’s surface. Similarly to the ﬁrst

setup, this signal is then used to trigger the pulse gen-

erator. Subsequently, the pulse generator triggers the

Nd:YAG laser with a temporal delay of 100 µs. This

delay combined with an internal processing delay of

approximately 300 µs from ﬂashlamp triggering to Q-

switch activation, ensures that the Nd:YAG laser al-

ways ﬁres after the Er:YAG laser’s pulse has ceased.

3.3 Sample Preparation and Data

Analysis

Fat and muscle tissue samples were extracted from

bisected ex-vivo pig heads at the Department of Oral

and Maxillofacial Surgery University Hospital Erlan-

gen. Using a knife, the tissues were cut into nearly

rectangular pieces of about 5 − 8 mm thickness for

LIBS measurements.

An isotonic saline solution (9 g NaCl per liter) is used

to create a thin liquid layer on the tissue surface to

partly mimic a layer of body ﬂuids present at in-vivo

conditions, for example, during general surgery. Us-

ing a small medical syringe, a few droplets of sa-

line solution are applied onto the tissue surface to be

able to spread the liquid in such a manner to establish

a consistent liquid layer thickness of approximately

20 − 30µm (measured by OCT) that stays steady over

the course of the double-pulse LIBS experiments.

For the LIBS experiments, 6 tissue samples of each,

muscle and fat tissue, are measured in both LIBS

setups. Each tissue sample is measured 100 times,

where one half of the measurement set is obtained

under dry conditions and the other one with a sa-

line solution. Therefore, 600 LIBS measurements are

taken in each of the experimental setups.

The experimental data obtained from LIBS measure-

ments of the tissue samples is subsequently prepared

to undergo different statistical analysis techniques to

investigate the effect of the saline solution and the

cleaning effect of the surgical laser. Here, statistical

analysis is performed to observe the similarity and

differences among the tissues investigated under dif-

ferent conditions. As a ﬁrst step, Principal Compon-

ent Analysis (PCA) is used to reduce the high dimen-

sionality of the data before Linear Discriminant Ana-

lysis (LDA) is performed on the data to classify each

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

248

data set into groups. As a ﬁnal step, the method of

Receiver Operating Characteristics (ROC) determines

the performance of the classiﬁcation by LDA. Import-

ant values for performance assessment of ROC ana-

lysis are the sensitivity (true positive rate), speciﬁcity

(false negative rate).

4 RESULTS AND DISCUSSION

4.1 Single-pulse LIBS

Figure 2 shows the average LIBS spectrum for dry

fat and muscle tissue, which are compared. Due to

the fact that fat tissue contains a very high amount of

carbon (Woodard and White, 1986), the maximum in-

tensity peak in the LIBS spectrum can be observed at

247.85 nm which corresponds to the emission line of

carbon species (Kramida and Ralchenko, ). Further

we observe a smaller peak at 279.55 nm that corres-

ponds to the emission line of magnesium (Mg) which

is contained in fat tissue in a small amount (Wood-

ard and White, 1986). Comparing these values to the

average spectrum of muscle tissue, the carbon peak

obtained from muscle is signiﬁcantly smaller which

is expected due to the relatively much lower carbon

amount in this tissue type. The magnesium peak is

stronger in the muscle spectrum. On the other hand,

the LIBS spectrum of fat shows several signiﬁcant

peaks in the spectral region 385 − 390nm which are

observed to a much lower extent for muscle tissue.

This peak region corresponds to carbon-related mo-

lecular emissions from CN (with band head at ap-

proximately 388.42 nm) which is related to the higher

carbon content of fat tissue. Other peaks of interest

for successful tissue differentiation are the sodium

peak (Na) at 589 nm, the hydrogen peak at 656.3 nm

and the oxygen peak (O) at 777.4 nm. While the so-

dium and the hydrogen peaks are very comparable for

muscle and fat tissue regarding the intensity, the oxy-

gen peak is much stronger observed for muscle tissue.

Next step is to compare the spectra of the tissues with

saline layer (Figure 3).

The most prominent change in the spectrum is

easily observed in the carbon peak at 247.85 nm for

muscle and fat tissue. Due to the liquid on the tissue

surface, the LIBS laser does not ablate as much tis-

sue material as without the liquid, because the laser

energy is now absorbed by the tissue as well as the li-

quid leading to a smaller amount of energy deposited

into the tissue material. Hence, there are less carbon

species ionized in the plasma plume. In the same way,

a decrease in the intensity values of the magnesium

and the CN peak can be observed. At the same time,

Figure 2: Measured single-pulse LIBS spectra for dry fat

and muscle tissue.

Figure 3: Measured single-pulse LIBS spectra for contam-

inated fat and muscle tissue.

the sodium peak at 589 nm increases, which can be

explained reliably by the use of saline solution that

contains 0.9 % of sodium chloride. The increased so-

LIBS based Tissue Differentiation for Er:YAG Surgical Laser

249

dium peak proves the assumption that the elements

contained in the saline solution will have an impact

on the obtained LIBS spectra of the tissue.

4.2 Double-pulse LIBS using Er:YAG

Laser

Analyzing the average LIBS spectrum of dry (Fig-

ure 4) and wet fat tissue (Figure 5) ﬁrst, it can be ob-

served that the C and CN peaks signiﬁcantly change.

This is due to the saline solution layer on the tis-

sue surface preventing carbonization of the underly-

ing tissue, so that less carbon species were ablated and

detected in the plasma plume. In addition to that, the

sodium (Na) peak at 589 nm increased slightly due to

the sodium content of the saline solution, which oc-

cured to a similar extent in the spectra of the single-

pulse setup.

Figure 4: Measured double-pulse LIBS spectra for dry fat

and muscle tissue.

Similar to the observations made for fat tissue,

the C and CN peaks signiﬁcantly decreased in com-

parison to the average spectrum of dry muscle tissue

taken in this setup. This can be explained again by

the saline solution layer on the tissue surface lessen-

ing the process of carbonization by the Er:YAG laser.

Moreover, an increase of the sodium, hydrogen and

oxygen peaks is observed (Figure 5). Although this

may indicate that the Er:YAG laser did not ablate all

of the saline solution prior to LIBS analysis, there is

Figure 5: Measured double-pulse LIBS spectra for contam-

inated fat and muscle tissue.

another scenario potentially contributing to this be-

havior. Since the Er:YAG laser pulse is incident

shortly before the LIBS laser, ablated material by

the ﬁrst laser pulse is ejected perpendicularly to the

surface and could interfere with the following LIBS

laser. The Er:YAG laser could ablate the saline solu-

tion, whose ablated particles are then ionized by the

Nd:YAG laser and therefore detected in the plasma

plume. This would explain that the Er:YAG laser in-

deed is able to clean the tissue surface, however, the

ablated material is also detected.

4.3 Classiﬁcation Results

Here, we will only focus on the differentiation of the

same tissue type under the two conditions. This is be-

cause the differentiation between the two tissue types

under the two conditions should not be difﬁcult as

their elemental composition is different and hence the

effect of the saline layer is not expected to interfere

to a level that diminishes their difference. This as-

sumption is in agreement with the results in table 1

which gives an overview on the classiﬁcation results

for single-pulse and double-pulse LIBS. It shows the

sensitivity, speciﬁty and AUC (area under curve for

ROC analysis). Classiﬁcation performance is very

good for single-pulse LIBS which is considered as

disadvantageous since the same tissue type should

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

250

Table 1: Classiﬁcation results for single -and double-pulse LIBS. The sensitivity, speciﬁcity and AUC values are shown.

Sample Combination Sensitivity Speciﬁcity AUC

Single-pulse Fat-dry / Fat-water 0.9967 0.9933 0.9987

LIBS Muscle-dry / Muscle-water 0.9133 0.9400 0.9792

Fat-dry / Muscle-dry 0.9967 0.9933 0.9987

Fat-water / Muscle-water 0.9967 0.9933 0.9995

Double-pulse Fat-dry / Fat-water 0.8000 0.8600 0.8629

LIBS Muscle-dry / Muscle-water 0.8267 0.7567 0.8364

Fat-dry / Muscle-dry 0.9733 0.9767 0.9911

Fat-water / Muscle-water 0.9300 0.9267 0.9613

not be differentiated. Comparing the average val-

ues of sensitivity, speciﬁcity and AUC for both LIBS

setups used, it becomes obvious that the classiﬁca-

tion performance decreases for each classiﬁer when

the Er:YAG laser is used. The cleaning effect of the

Er:YAG laser causes the average AUC level to drop by

around 13 % differentiation for fat and muscle tissue.

5 SUMMARY

The experimental results obtained in this study are ex-

pected to lay the groundwork for future soft tissue dif-

ferentiation under in-vivo conditions after the LIBS

technique has proven in recent studies to have the po-

tential for successful ex-vivo tissue differentiation.

The results in this study have shown promising pro-

spects for future in-vivo tissue differentiation using

LIBS. Considering that different kinds of body ﬂu-

ids or liquids can be present on tissue surfaces during

surgery, it was shown that a thin saline solution layer

applied onto the surface has a signiﬁcant inﬂuence on

LIBS spectra of fat and muscle tissue. Furthermore,

the study indicates that an ER:YAG laser pulse prior

to the LIBS laser pulse can serve as an in-vivo surface

cleaner leading to improved tissue classiﬁcation.

In future, the inﬂuence of other possible body ﬂuids

such as blood needs to be addressed to fully under-

stand and evaluate the potential of LIBS to perform

in-vivo tissue differentiation in the future. In addi-

tion to that, the potential inﬂuence of the liquid layer

thickness has to be investigated and other types of tis-

sue such as nerve and skin tissue have to be taken into

consideration for those investigations.

ACKNOWLEDGEMENTS

The authors gratefully acknowledge the funding of

the Erlangen Graduate School in Advanced Optical

Technologies (SAOT) by the Deutsche Forschungs-

gemeinschaft (German Research Foundation-DFG)

within the framework of the Initiative for Excellence.

REFERENCES

Cremers, D. A. and Radziemski., L. J. (2013). Hand-

book of laser-induced breakdown spectroscopy.

Wiley.

Goldberg., D. (2013). Laser Dermatology. Springer.

Kramida, A. and Ralchenko, Y. NIST Atomic Spectra

Database (version 5.4),.

Mehari, F. (2014). Laser induced breakdown spectro-

scopy for bone and cartilage differentiation - ex

vivo study as a prospect for a laser surgery feed-

back mechanism. Biomedical optics express,,

5(11).

Solomon, K. D. (2009). Lasik world literature review:

quality of life and patient satisfaction. Ophthal-

mology, 116(4).

Walter, M. (1999). Photoablation of bone by excimer

laser radiation. Lasers in Surgery and Medicine,,

25.

Woodard, H. Q. and White, D. R. (1986). The com-

position of body tissues. The British journal of

radiology,, 59(708).

LIBS based Tissue Differentiation for Er:YAG Surgical Laser

251