Reverse Translational Research

How Clinical Trials on Fluorescence Imaging for Vocal Cord Cancer Fuels

Fundamental Research

Olivier Gaiffe

, Christian Pieralli

, Laurent Tavernier

, Lionel Pazart

and Bruno Wacogne

1,2

INSERM CIT 808, Besançon University Hospital, Place Saint Jacques, 25030 Besançon cedex, France

FEMTO-ST Institute, UMR 6174, 16 Route de Gray, 25030 Besançon cedex, France

Besançon University Hospital ENT Unit, Bd Flemming, 25030 Besançon cedex, France

Keywords: Translational Research, Vocal Cords, Fluorescence Imaging.

Abstract: Translational research consists in translating fundamental research results as closely as possible to patients.

Researchers sometimes underestimate these studies because it is thought that, although essential for setting

up new investigation tools, they do not deepen fundamental knowledge. However, users face specific

difficulties due to the variability of the biological systems under study. Variability is easily understood from

one patient to another, but there is also variability in a single patient whose metabolism evolves together

with therapeutic actions. Results obtained in translational research often depend on this variability, and new

questions and scientific obstacles arise when research is applied to the real world. In order to address these

new challenges, reverse translational research is required. Fundamental research is fuelled by the results of

translational research. In this position paper, we consider vocal cord fluorescence imaging as an example of

bi-directional translational research. First, we briefly recall the basics of fluorescence imaging, and we

explain why commercial fluorescence systems lead to variable estimations of their efficiency by end-users.

Second, we describe solutions intended to improve fluorescence techniques. This position paper will then

make conclusions.

1 INTRODUCTION

Translational research is a fairly new and rapidly

evolving concept. The general idea is to translate

fundamental research results as closely as possible to

patients via pre-clinical and clinical trials. In other

words, it consists in taking research from bench to

bedside. Sometimes researchers underestimate these

studies because it is thought that, although essential

for setting up new investigation tools, they do not

deepen fundamental knowledge. However, users

face specific difficulties due to the variability of the

biological systems under study. Variability is easily

understood from one patient to another, but there is

also variability in a single patient whose metabolism

evolves together with therapeutic actions. Results

obtained in translational research often depend on

this variability, and new questions and scientific

obstacles arise when research is applied to the real

world. In order to address these new challenges,

reverse translational research is required.

Fundamental research is then fuelled by the results

of translational research.

Consequently, the concept of bidirectional

translational research is emerging: forward from

researchers to end-users and then back to research to

answer new questions or improve current results. In

this position paper, we illustrate this idea through

examples concerning fluorescence optical diagnosis,

and more precisely tools that could be used to

diagnose vocal fold disease.

In part two of this paper, we briefly recall the

basics of fluorescence and we highlight the fact that

both the excitation wavelength and observation

wavelength window must be carefully chosen to

efficiently assess the composition of the tissue under

examination. This section illustrates what is called

translational research.

In part three, we explain why end-user teams

often disagree on the performances of current

commercial fluorescence systems. Here, we will see

that multimodality may offer versatile systems

intended for most end-users. This section illustrates

how translational research can put to evidence new

282

Gaiffe O., Pieralli C., Tavernier L., Pazart L. and Wacogne B..

Reverse Translational Research - How Clinical Trials on Fluorescence Imaging for Vocal Cord Cancer Fuels Fundamental Research.

DOI: 10.5220/0004323802820287

In Proceedings of the International Conference on Biomedical Electronics and Devices (BIODEVICES-2013), pages 282-287

ISBN: 978-989-8565-34-1

 2013 SCITEPRESS (Science and Technology Publications, Lda.)

scientific obstacles.

Part four is devoted to the description of

hyperspectral fluorescence techniques that may be

used to improve diagnosis efficiency. We also

propose architectures based on multimodality. Early

experimental results will be presented for these

techniques. This section illustrates how translational

research fuels fundamental research by means of

what we call reverse translational research.

This paper will then make conclusions.

2 BASICS OF FLUORESCENCE:

TRANSLATION

One way to discriminate unhealthy tissues from

healthy ones is by detecting an abnormal

concentration of particular proteins by means of

fluorescence measurements. Fluorescence can be

explained with the help of the Jablonski diagram

(figure 1). By illuminating the molecules, electrons

absorb light and “go upward” from energy state U0

to energy state U1. The molecules then relax, first by

a non-radiative loss of energy (from U1 to U2) and

finally down to energy state U0. This last step is a

radiative process and the light emitted at this stage is

called fluorescence.

Figure 1: Jablonski diagram.

Two different methods can be used to detect

proteins. The first deals with exogen fluorescence

which consists in applying a mixture of mono- or

polyclonal antibodies functionalized with

fluorophores. This method requires foreign

substances to be injected into the body, however.

The other way of detecting the target proteins is

to use endogen fluorescence, also termed

autofluorescence. It consists in studying the natural

fluorescence of the target proteins. In this case, once

the target proteins are determined, an illumination

source suited to their absorption spectrum must be

chosen. The first step of the study is therefore to

determine the target proteins adequately. For

instance, for vocal folds, we can name collagen,

elastin, NADH, flavins or porphyrins. Figure 2

shows the absorption and emission spectra of flavins

(Wagnieres 1998 – Richards-Kortum 1996).

Figure 2: Absorption and emission spectra of flavins.

2.1 Influence of the Excitation

Wavelength

Fluorescence intensity depends on different factors

such as the quantum properties of the fluorescent

protein, its concentration in the tissue and the

absorption of both excitation and emission

wavelengths by the tissue. These parameters are not

easily controllable. However, fluorescence intensity

also depends on the excitation wavelength used.

Figure 3 illustrates this with flavins. The shape of

the emission spectrum remains constant but its

amplitude depends on the excitation wavelength.

Figure 3: Effect of the excitation wavelength on

fluorescence intensity.

This illustrates the importance of the choice of

excitation wavelength. Note that the fluorescence

intensity also depends on the spectral width of the

excitation around the central excitation wavelength.

2.2 Influence of the Observation

Wavelength Window

Several fluorescent proteins are usually present in

tissue. A single excitation wavelength often induces

fluorescence in different proteins even if the

excitation wavelength is carefully chosen. This does

not necessarily mean that observation provides

average information on each individual protein

ReverseTranslationalResearch-HowClinicalTrialsonFluorescenceImagingforVocalCordCancerFuelsFundamental

Research

283

contribution. Since different proteins emit

fluorescent light according to their own emission

spectra, the choice of the observation wavelength

window may help differentiate between the

information from each protein. This is illustrated in

figure 4 in the case where both flavins and

porphyrins are considered (Wagnieres 1998 –

Richards-Kortum 1996).

Figure 4: Influence of the observation window.

As can be seen, when higher observation

wavelengths are considered, the greatest

contribution comes from porphyrins. Conversely,

when shorter wavelengths are considered, the

greatest contribution comes from flavins.

3 TRANSLATIONAL RESEARCH:

CONFRONTATION WITH

REAL SITUATIONS

In real life, the situation is much more complicated

due to the large number of fluorescent proteins

present in the tissues. The figure below shows the

excitation and emission spectra of several proteins

(Wagnieres, 1998); (Richards-Kortum, 1996).

It can be seen that neither the choice of the

specific excitation wavelength nor the choice of the

observation window can help to dissociate the

fluorescence contribution of each individual protein.

Concerning translational research, a large

number of studies have been conducted in order to

assess the efficiency of experimental or commercial

devices. Highly interesting reviews have been

published (Piazza, 2011); (Kraft, 2010); (Shin,

2010); (Mehrotra, 2010); (Rethman, 2010).

Figure 5: Absorption and emission spectra of some

proteins.

Commercial apparatuses relying on this principle

are readily available but the clinical trials performed

to date do not present sufficient evidence of their

ability to provide a reliable diagnosis (Piazza, 2011);

(Rethman, 2010). These systems suffer from the fact

that the excitation is made over a large wavelength

band (Arens, 2007); (Mehrotra, 2010). In other

words, many proteins are excited: the useful signal is

buried in the various fluorescence signals emanating

from non-relevant proteins, thus preventing the

target protein from being detected. The same is true

of their detection principle, as the observation

wavelength range is also quite wide. Therefore,

superimposition of information from a large number

of proteins is observed and dilutes useful signal into

what can be considered as noise. Figure 6 shows the

excitation and observation wavelength windows

commonly used in commercial systems.

Although autofluorescence is considered as

highly effective in the early diagnosis of laryngeal

cancer and its precursor lesions (Kraft, 2010),

clinical trials lead to varying specificity and

sensitivity results even with the same commercial

system: 98% sensitivity and 100% specificity (Lane,

2006), 50% sensitivity and 38.9% specificity

(Mehrotra, 2010) and 97% sensitivity and 94%

specificity (Poh, 2006).

Figure 6: Wavelength windows commonly used in clinical

trials.

BIODEVICES2013-InternationalConferenceonBiomedicalElectronicsandDevices

284

In fact, in the case of upper aerodigestive tract

cancer evaluation, optical techniques such as tissue

staining, chemiluminescence, autofluorescence and

tissue reflectance analysis have given unconvincing

results. For the latter two techniques, the following

description was reported (Rethman, 2010):

- "There is insufficient evidence that commercial

devices based on autofluorescence enhance visual

detection of potentially malignant lesions beyond

that achieved through a conventional visual and

tactile examination (Patton, 2008).

- There is insufficient evidence that commercial

devices based on tissue reflectance enhance visual

detection of potentially malignant lesions beyond

that achieved through a conventional visual and

tactile examination (Patton, 2008)."

The conclusion to be made regarding these

translational research results is that none of the

techniques is entirely satisfactory. Because these

techniques can be greatly enhanced in terms of

sensitivity and specificity, they often require

subjective interpretation and depend on the visual

recognition skills of the examiner (Shin 2010). The

same authors also explain that "the combination of

wide-field and high-resolution fluorescence imaging

systems with automated image analysis should be

investigated to maximize overall diagnostic

performance".

At this stage, forward translational research was

conducted. Research on fluorescence led to the

testing of advanced systems in clinical trials. These

trials concluded that systems are largely improvable

in terms of specificity and/or sensitivity. We have

explained the reasons for information loss in the

present paper. Reverse translational research should

now be envisaged to explore the performance of

advanced fluorescence techniques, possibly coupled

with other optical investigations of tissue properties.

4 REVERSE TRANSLATIONAL

RESEARCH: HOW TO

ANSWER NEW OBSTACLES

HIGHLIGHTED BY CLINICAL

TRIALS

We entirely agree with (Shin, 2010) concerning the

utility of combined techniques. Furthermore, we

believe that other modalities may be included in

advanced devices.

For example, we may consider a device that can

excite target molecules at several specific

wavelengths and also perform hyperspectral signal

detection. We can thus analyze one or more small

spectral bandwidths centered on the fluorescence

wavelengths of several proteins (Muller, 2003);

(Gillenwater, 1998).

4.1 Possibilities with Hyperspectral

Fluorescence

Our basic experimental set-up consists of a fiber

probe containing several optical fibers (figure 7:

left). The central one is used to illuminate the tissue

with a series of monochromatic wavelengths (via

optical switches). Collection fibers are used to

collect the emitted fluorescence. The light is then

launched into a spectrometer. For each pixel of the

image, the whole fluorescence spectrum is recorded

as depicted in figure 7 (right) in the case of a tree

leaf.

Figure 7: Optical fiber probe (left) and hyperspectral

fluorescence image (right).

In our preliminary experiments, the image is

formed by scanning the sample. In a more advanced

system, fiber bundles coupled with hyperspectral

CCD cameras will be employed.

The image obtained in this case consists of a 4D

hyperspectral cube. There are different ways of

investigating this hyperspectral cube. Figure 8

illustrates this. On the one hand we can select

different observation wavelength windows (top

right). It can be clearly seen that different pieces of

information appear depending on the observation

window. On the other hand, we can try to define

hypervolumes in the hyperspectral cube. These

hypervolumes can be specific to the possible

pathological nature of the tissue. This can be

achieved through the use of data processing

algorithms, such as the Kernel Principal Component

Analysis or the Support Vector Machine (Diaz-Ayil,

2007); (Adbat, 2012).

ReverseTranslationalResearch-HowClinicalTrialsonFluorescenceImagingforVocalCordCancerFuelsFundamental

Research

285

Figure 8: Different ways of exploiting hyperspectral

fluorescence.

This hyperspectral technique can be further

developed using images obtained with different

excitation wavelengths. For each supplementary

excitation wavelength, four dimensions are added to

the hyperspectral cube as depicted in figure 9.

Figure 9: Illustration of multiple excitations and

hyperspectral fluorescence imaging.

Note that fluorescence lifetime imaging could

also be included in this type of measurement.

4.2 Other Possibilities with Diffusion

This autofluorescence measurement can also be

coupled with an analysis of the diffuse reflectance of

the tissues, i.e. on the photon elastic scattering

process occurring in the biological sample. The

bimodality of the system obtained can then

significantly improve the sensitivity and specificity

of the diagnosis (Diaz-Ayil, 2007). Beyond sheer

data acquisition and direct analysis, further

information can be obtained from the clinical trials

by coupling and crossing the fluorescence and

scattering measurements, allowing us to retrieve the

intrinsic fluorescence characteristics of the sample

(Wu, 1993); (Muller, 2001).

Figure 10 illustrates the possibilities offered by

spectroscopic reflectance measurements. We can see

that different details can be observed depending on

the spectral observation window used to investigate

diffusion properties of tissues. Note that these

images were obtained with the same fiber probe as

the one used in the previous sub-section, thus

demonstrating that multimodality can be obtained

relatively easily.

Figure 10: Example of spectral analysis of tissue diffusion.

Now let us go back to the discussion of

translational research. In the previous section, we

have seen that translational research highlights

obstacles and questions that would not have arisen if

they had not been applied to real clinical situations.

In this section, we see that it is the role of reverse

translational research to go back to more

fundamental studies and to work at innovative

solutions in order to address these new questions.

5 CONCLUSIONS

In this position paper we look at optical fluorescence

imaging of the vocal folds as an example of bi-

directional translational research. As mentioned in

the introduction, translational research is sometimes

underestimated because it is thought that it does not

deepen fundamental knowledge.

However, users face specific difficulties due to

the variability of the biological systems under study.

Results obtained in translational research often

depend on this variability and new questions and

scientific obstacles arise when research is applied to

the real world. In order to address these new

challenges, reverse translational research is required.

Fundamental research is then fuelled by the results

of translational research and the latter should be

considered essential to fully understand the

biological system under study.

We focus herein on advanced fluorescence

techniques in order to illustrate bidirectional

translational research. The optical methods that are

currently used are in need of improvement. We thus

propose to develop a standalone device able to

assess the possible pathological nature of vocal

folds. Pre-clinical and clinical trials will then be

conducted in order to transform the expected

research results into new optical diagnosis tools.

To conclude, we believe that translational

research should not be underestimated because it

BIODEVICES2013-InternationalConferenceonBiomedicalElectronicsandDevices

286

fuels fundamental studies when new questions or

scientific obstacles have been highlighted by pre-

clinical or clinical trials. Through the example of

vocal cord fluorescence imaging, this position paper

illustrates that the concept of bi-directional

translational research should be applied to all work

aiming to develop new medical devices.

ACKNOWLEDGEMENTS

The studies presented in this paper are conducted in

the frame of the STREP project "µRALP" funded by

the European Commission's 7th Framework

Program. It addresses the FP7-ICT-2011-7 call for

cognitive and robotic systems.

REFERENCES

Adbat, F., Amouroux, M., et al., 2012, Hybrid feature

selection and SVM-based classification for mouse skin

precancerous stages diagnosis from bimodal

spectroscopy, Opt. Express, Vol. 20, pp. 228-244.

Arens, C., Reussner, D., et al., 2007, Indirect fluorescence

laryngoscopy in the diagnosis of precancerous and

cancerous laryngeal lesions, Eur. Arch. Oto-Rhino-L.,

Vol. 264, pp. 621–626.

Diaz-Ayil, G., Amouroux, M., et al., 2007, Bimodal

spectroscopic evaluation of ultra violet-irradiated

mouse skin inflammatory and precancerous stages:

instrumentation, spectral feature extraction/selection

and classification (k-NN, LDA and SVM), Eur. Phys.

J-Appl. Phys. Vol. 47, pp. 12707.

Gillenwater, A., Jacob, R., et al., 1998, Noninvasive

Diagnosis of Oral Neoplasia Based on Fluorescence

Spectroscopy and Native Tissue Autofluorescence,

Arch. Otolaryngol., Vol.124, pp. 1251-1258.

Kraft, M., Betz, C. S., et al., 2010, Value of fluorescence

endoscopy for the early diagnosis of laryngeal cancer

and its precursor lesions, Head Neck, Vol. 33

pp. 941-948.

Lane, P.M., Gilhuly, T., et al., 2006, Simple device for the

direct visualization of oral-cavity tissue fluorescence,

J. Biomed. Opt., Vol 11, pp. 024006-1 – 024006-7.

Mehrotra, R., Singh, M., et al., 2010, A cross-sectional

study evaluating chemiluminescence and

autofluorescence in the detection of clinically

innocuous precancerous and cancerous oral lesions, J.

Am. Dent. Assoc., Vol. 141, pp. 151–156.

Muller, M. G., Georgakoudi, I., et al., 2001, Intrinsic

fluorescence spectroscopy in turbid media

disentangling effect of scattering and absorption, Appl.

Optics, Vol. 40, pp. 4633-4646.

Patton, L. L., Epstein, J. B., et al., 2008, Techniques for

oral cancer examination and lesion diagnosis: a

systematic review of the literature, J. of American

Dental Association JADA, Vol. 139, pp. 896-905.

Piazza, C., Bon, F. D., et al., 2011, Biologic endoscopy’:

optimization of upper aerodigestive tract cancer

evaluation, Otolaryng. Head Neck, Vol. 19, pp. 67–76.

Poh, C. F., Zhang, L. et al., 2006, Fluorescence

visualization detection of field alterations in tumor

margins of oral cancer patients, Clin. Cancer Res.,

Vol. 12, pp.6716-6722.

Rethman, M. P., Carpenter, W., et al., 2010, Evidence-

Based Clinical Recommendations Regarding

Screening for Oral Squamous Cell Carcinomas, J. of

American Dental Association JADA, Vol. 141, pp.

509-520.

Richards-Kortum, R; Sevick-Muraca, E., 1996,

Quantitative optical spectroscopy for tissue diagnosis,

Annu. Rev. Phys. Chem. Vol. 47, pp. 555–606.

Shin, D., Vigneswaren, N. et al., 2010, Advances in

fluorescence imaging techniques to detect oral cancer

and its precursors, Future Oncol., Vol. 6, pp. 1143–

1154.

Wagnieres, G. A., Star, W. M., et al., 1998, In vivo

fluorescence spectroscopy and imaging for

oncological applications, Photochem. Photobiol.,

Vol.68, pp.603–632.

Wu, J., Feld, M. S., et al., 1993, Analytical model for

extracting intrinsic fluorescence in turbid media, Appl.

Optics, Vol. 32, pp. 3585-3595.

ReverseTranslationalResearch-HowClinicalTrialsonFluorescenceImagingforVocalCordCancerFuelsFundamental

Research

287