DEVELOPMENT OF AN EX VIVO QUANTITATIVE

SPECTROSCOPIC SCANNER

D. S. Ferreira

, N. Lue

, G. Minas

, M. S. Feld

, K. Badizadegan

3,4

and C. Yu

Centro Algoritmi, Universidade do Minho, Campus de Azurém, 4800-058, Guimarães, Portugal

GR Harrison Spectroscopy Laboratory, Massachusetts Institute of Technology

77 Massachusetts Avenue, Cambridge, MA 02139, U.S.A.

Departments of Pathology and Health Sciences and Technology, Harvard Medical School, Boston, MA 02114, U.S.A.

James Homer Wright Pathology Laboratories, Massachusetts General Hospital

WRN219, 55 Fruit St, Boston, MA 02114, U.S.A.

Optics Research Laboratory, Research & Development Division, Canon U.S.A., Inc.

9030 South Rita Road, Suite 302, Tucson, AZ 85747, U.S.A.

Keywords: Spectroscopic scanner, Diffuse reflectance, Fluorescence, Spectral imaging.

Abstract: We describe an ex vivo quantitative spectroscopy (QS) scanning platform which enables integration of

different optical modalities for the assessment of ex vivo tissue properties. As a first implementation, the QS

scanner combines diffuse reflectance spectroscopy (DRS) and intrinsic fluorescence spectroscopy (IFS) to

provide a multidimensional image of tissue structural and biochemical properties. The wide area coverage is

achieved by mechanically scanning of the optical probe. The spectroscopic data is taken one grid at a time

with variable grid-to-grid (GTG) distance and field of view (FOV). The ex vivo tissue surface under

examination can have a variable size since both GTG distance and FOV can be controlled. We demonstrate

the clinical utility of this system using an ex vivo tissue model with ultimate goal of imaging excised tissue

margins.

1 INTRODUCTION

Fast and reliable intra-operative diagnosis is critical

for the success of oncological surgery in a variety of

organ systems. After any cancer tissue resection it

has to be ensured that all malignant tissue was

removed and for that a surgical pathologist has to

examine the tissue margins (Haka, 2006). Current

clinical standards include visual inspection of the

tissue, followed by selective assessment of any

suspicious sites by frozen sectioning and rapid

histological evaluation. This procedure is still not

very efficient since according to a study by the

College of American Pathologists (Novis, 1997) a

significant number of hospitals do not routinely

provide intra-operative feedback to the surgeon

within 20 minutes of tissue delivery, adding costs

and an increased risk of morbidity associated with

extra time spent in the operating room. Additionally,

frozen section diagnoses are almost always

performed on a few “representative” portions of

tissue, resulting in potential discrepancies between

the frozen section assessment and the definitive

margin status which becomes only available once

the entire tissue has been processed post-operatively.

Therefore, there is a significant technological and

clinical need for methods capable of rapid and

reliable evaluation of excised tissues in real time.

Ex vivo imaging strategies have already been

proposed as potential tools for surgical margin

assessment. Mahadevan-Jansen and coworkers have

successfully applied contact probe autofluorescence

and diffuse reflectance spectroscopy, and a spectral

imaging to classify positive and negative margins of

excised breast specimens with high sensitivity and

specificity (Keller, 2010). However, a quantitative

analysis using images was not performed. Pogue and

co-workers used confocal reflectance microscopy

and spectrograph to raster-scan ex vivo tumors

margins and obtained mainly quantitative scattering

parameters associated with tissue morphology

(Krishnaswamy, 2009). Similar to previous wide

area spectroscopic imaging, the non-probe method

required some correction to deal with lines shape

spectra if the technique was used to acquire

quantitative absorption parameters (Keller, 2010). A

274

S. Ferreira D., Lue N., Minas G., S. Feld M., Badizadegan K. and Yu C..

DEVELOPMENT OF AN EX VIVO QUANTITATIVE SPECTROSCOPIC SCANNER.

DOI: 10.5220/0003888402740279

In Proceedings of the International Conference on Biomedical Electronics and Devices (BIODEVICES-2012), pages 274-279

ISBN: 978-989-8425-91-1

 2012 SCITEPRESS (Science and Technology Publications, Lda.)

quantitative optical imaging device to assess breast

tumor margins was developed by Ramanujam et al.

(2009) using diffuse reflectance spectroscopy for the

extraction of scattering and absorption information.

In this work the margin surface is obtained by

manually translating an imaging probe that contacts

the specimen in a container through pre-drilled holes

with 5 mm center to center spacing.

Spectral imaging devices for the assessment of

excised tissue have two major advantages over

traditional practices of surgical margin assessment:

(1) real time analysis, with the benefit of reducing

patient anxiety and avoid potential follow-up

surgery; (2) whole area assessment, with the benefit

of reducing the probability of missing a lesion.

These devices can have a huge value in breast-

conserving surgery or in endoscopic mucosal

resection (EMR). Several studies have suggested

that in breast-conserving excision of cancer re-

operation for positive surgical margins discovered

after the surgery may be required in up to 50% of all

cases (Allweis, 2008), and that EMR achieves

complete excision of dysplastic lesions in only 4%

of the cases (Mino-Kenudson, 2005).

In this paper, we present a new and

complementary strategy to enable real-time

comprehensive assessment of surgical margins in

excised tissues. We have developed an ex vivo

spatial high-resolution quantitative spectroscopy

(QS) scanning platform which enables integration of

different optical modalities to provide quantitative

tissue information that correlate to disease state of

the surrounding tissue. Wide area imaging of

excised tissue is achieved by mechanically scanning

an optical probe, with variable spatial resolution.

Tissue samples are placed for analysis in a flat

platform, which enables an equal pressure across

time and an equal distance between probe and tissue

throughout the analysis.

The developed scanning platform currently

integrates diffuse reflectance spectroscopy (DRS)

and intrinsic fluorescence spectroscopy (IFS) for the

extraction of several spectroscopic parameters, but it

is adaptable to assemble many other optical

modalities. DRS and IFS have been first

implemented on the scanning platform since they

have shown great ability for the detection of

neoplastic diseases by assessing different spectral

features associated with normal and cancerous

tissues (Georgakoudi, 2001); (Tunnell, 2003); (Yu,

2008). These modalities provide quantitative

information about biochemical and structural tissue

attributes, from which diagnostic algorithms can be

developed.

Diffuse reflectance spectra from tissues are used

to extract information about hemoglobin

concentration and saturation, and light scattering

parameters using a well-developed model based on

the diffusion approximation of light propagation in

tissue. DRS provides information about the

morphology and biochemistry of the bulk tissue

(Zonios, 1999). Intrinsic fluorescence is the

fluorescence unaffected by tissue scattering and

absorption, and is obtained using the diffusely

reflected light to remove spectral distortions. The

relative contributions of the endogenous tissue

fluorophores (e.g., NADH and collagen) can be

extracted from the intrinsic fluorescence (Müller,

2001). Several studies of reflectance and

fluorescence for tissue diagnosis using optical fiber

contact probes for light delivery and collection have

been performed in different anatomic sites (Bard,

2006); (Chang, 2005); (Georgakoudi, 2001);

(Müller, 2003). Despite their potential, contact probe

techniques commonly suffer from undersampling.

The proposed ex vivo QS scanning platform

overcomes this drawback since it extends spectral

diagnosis to the imaging mode, enabling wide area

surveillance of tissue ex vivo.

This paper describes the design and feasibility

studies of a multi-modal scanning platform for an

intra-operative medical device that is able to perform

a rapid, real-time, detailed, and reliable quantitative

spectroscopic analysis of tissue surfaces. The major

benefit of this “adaptable scanning platform”

concept is that it is not limited to only one optical

modality, enabling the selection of the appropriate

technique for each margin assessment, or to use a

combination of different techniques.

2 METHODS

Given that ex vivo tissue analysis does not have

typical restrictions of in vivo imaging such as

imaging geometry, surface contour, patient motion

and the like, spectroscopic mapping of an arbitrarily

wide area is achieved by mechanically scanning an

optical probe in an inverted geometry. Spectroscopic

data are taken one grid at a time with variable grid-

to-grid (GTG) distance and field of view (FOV). In

our instrument, the ex vivo tissue surface under

examination can have a variable size from 2

square mm to 4 square cm, as well as variable

resolution, which can be as high as a quarter of the

spot size since both GTG distance and FOV are

controllable. Still, if necessary, the area for tissue

analysis can be easily increased in the future by

DEVELOPMENT OF AN EX VIVO QUANTITATIVE SPECTROSCOPIC SCANNER

275

including a larger sample holder to the scanning

platform. Using model-based diagnostic algorithms,

this instrument will be able to correlate

spectroscopic parameters with disease status in real

time.

The instrument’s contact probe, i.e. the FastEEM

probe (Tunnell, 2003), consists of a single light

delivery fiber surrounded by six collection fibers

that collect light from tissue and deliver it back to

the spectrograph (all seven fibers with 200 µm core

and NA = 0.22). All fibers are fused together at the

tip and polished at 17 degree angle to provide the

overlapping of detector and collector optical cones.

For spectroscopic scanning, we use only one of the

collection fibers to collect tissue reflectance and

fluorescence from a spot size of approximately 500

µm. The probe parameters are incorporated in our

reflectance (Zonios, 1999)

and fluorescence (Müller,

2001)

models. Wide area coverage is achieved by

scanning the light spot over the tissue using XY

mechanical scanning.

Figure 1 depicts the schematic diagram of the

instrument. To perform DRS measurements, white

light from a 75W CW xenon arc lamp (Oriel

Instruments, USA) is coupled via the delivery fiber,

to illuminate a “diagnostic spot” of ~0.5 mm in

diameter on the tissue sample. DRS signal from the

sample is collected, with adjacent collection fibers,

and coupled to a spectrometer (USB 2000+, Ocean

Optics, USA). A personal computer equipped with

Labview 8.5.0 software and DAQ data acquisition

board NI PCI-6221 (National Instruments, USA) are

used to control and coordinate the various

components, including the GTG distance and FOV

of the 2-D mechanical scanning (M-605.1DD and

M-126.DG1, Physik Instrument, Germany).

Spectroscopic data from the spectrometer and 2D

stage positions are acquired and analyzed. Same

resources, data handling and data acquisition are

utilized for IFS measurements, except the light

source is a pulsed diode pumped solid state laser that

delivers 355 nm light pulses of duration ~0.6 ns and

energy ~0.26 μJ at ~38 kHz (SNV-40F-000, Teem

Photonics, France). Note that different

measurements are accomplished through switching

the excitation sources with an installed flipping

mirror. Without any significant changes to the

scanning engine, the integration of other optical

modalities, such as hyperspectral, infrared or

Raman, would only require add-on and data

acquisition to the platform. This all-in-one device

could be a powerfull tool in clinical tissue

diagnostic. For the measurements, liquid phantoms

and tissue samples were placed in removable

ultraviolet glass Petri dishes, mounted on a custom

sample holder.

Figure 1: Schematic layout of the ex vivo QS scanning

platform.

3 RESULTS

3.1 System Calibration

Calibration was performed using liquid phantoms

with known scattering, absorption and fluorescence

properties. These phantoms were constructed by a

mixture of water based intralipid - scatterer -

(Fresenius Kabi AG), hemoglobin - absorber -

(Sigma Aldrich Co.) and furan - fluorophore -

(Lambda Physik) at various concentrations. This

fluorescent dye was selected because it has an

excitation and emission spectra similar to that of

collagen, which is an endogenous tissue fluorophore

important for diagnosis. All the measurements were

performed using a wavelength range from 350 to

700 nm for DRS and 380 to 700 nm for IFS.

The accuracy and capability of the system for

reflectance measurements was then carried on using

several combinations of various concentrations of

intralipid, furan (0.5 µg/mL) and hemoglobin (0.6, 1,

and 1.5 mg/mL). Reflectance spectra were acquired

from one spot in each sample. All the spectra were

normalized by a reflectance standard (Labsphere

SRS-20) in order to remove spectral distortions and

spatial inhomogeneities related with the instrument’s

spectral and spatial responses.

Figure 2 shows the calibrated DRS spectra, from

the same position on the sample, for different

phantoms. We use DRS for the extraction of

diagnostic information: by fitting the reflectance

spectrum to the diffuse scattering model described

by Zonios et al. (1999) three DRS parameters were

extracted for each pixel: A, the reduced scattering

Delivery fiber

White-Light

Source

Laser Source

Spectrometer

Sample holder

translation

sta ge

White-light

camera

Collection

fiber

Lens 1

Lens 2

Flipa ble

Mirror

BIODEVICES 2012 - International Conference on Biomedical Electronics and Devices

276

coefficient at the reference wavelength; B, related to

the average scatter size; and cHb, the total

concentration of hemoglobin. However, it is

important to notice that in the presented results all

the samples were exposed to air and, thus, their

estimated oxygen saturation is close to 1.0.

Figure 2: Calibrated reflectance spectra (solid lines)

measured on different tissue phantoms. The best fit spectra

are also plotted (dashed lines). The characteristic

absorption bands of hemoglobin at 420nm, 540nm, and

580nm are clearly visible.

Optimal fits were obtained between the measured

and computed spectra from the samples. From the

excellence agreement, the computed spectra give the

correct values of reflectance spectroscopy

parameters, which ensure that the instrument can

accurately measure the scatter and absorber

parameters. The values for parameters A, B, and cHb

are compiled in Table 1.

Table 1: Reflectance parameters (A, in mm

-1

, B, and cHb

in mg/mL) measured from tissue phantoms with different

hemoglobin concentrations.

The accuracy of fluorescence measurements was

assessed using the same set of phantoms. The

fluorescence at each spot is analyzed using IFS:

reflectance measurements are used to correct the

bulk fluorescence spectra (affected by scattering and

absorption) using the model described by Müller et

al. (2001) to extract the IFS spectra. Figure 3 shows

the fluorescence spectrum of pure furan in water

(blue line), and the several bulk and IFS spectra

measured using phantoms with different hemoglobin

concentrations.

Figure 3: Bulk fluorescence spectra (dashed lines)

measured on different tissue phantoms. The corresponding

intrinsic fluorescence spectra (IFS) are also plotted (solid

lines). The blue spectrum is the fluorescence measured

from pure furan in water. Note that the calculated IFS

spectra, which are independent to the absorbers and

scatterers and the raw spectrum of pure furan are well

overlapped.

As expected, the data confirms that bulk

fluorescence spectra vary considerably with

hemoglobin concentration in opposition to the IFS

spectra. The excellent agreement between the IFS

spectra and the spectrum of pure furan in water

indicates that IFS method can be used to remove the

distortions caused by tissue scattering and

absorption. These data provide evidence for accurate

calibration of our QS scanning platform.

3.2 Brain Tissue Imaging

The performance characteristics of the ex vivo QS

system were demonstrated using an inherently high

contrast sample with sharp regional boundaries (a

section of formalin-fixed human brain cortex with

gray and white matter, with an approximate size of 2

cm by 2 cm). A diffuse reflectance map of the brain

is shown in Figure 4(a), obtained using a step size of

125 μm, and an integration time of 3 ms at each

point. In this image, fine detail and high contrast

between the gray and white matter of the brain

cortex is clearly visible. For demonstration, high-

resolution spectral maps of the scattering parameter,

A, and measured hemoglobin concentration are

shown in Figure 4(b) and Figure 4(c), respectively.

As expected, it is revealed a higher hemoglobin

concentration on gray matter (related with higher

blood volume) (Hamberg, 1996). These results

demonstrate the ability of ex vivo QS scanner to

provide spectral contrast based on tissue parameters.

Parameters [Hb] = 0.6 [Hb] = 1.0 [Hb] = 1.5

A 1.074 1.074 1.070

B 0.289 0.278 0.230

cHb 0.619 1.002 1.472

DEVELOPMENT OF AN EX VIVO QUANTITATIVE SPECTROSCOPIC SCANNER

277

Figure 4: Representative QS images of brain cortex taken

with the proposed spatial high-resolution scanner: a) total

reflectance maps (in arbitrary units); b) quantitative map

of the scattering parameter A (mm

-1

); c) quantitative map

of hemoglobin concentration (mg/dl). Each box represents

approximately a 1 x 1 cm scanning area.

The concept has thus been demonstrated with the

developed bench-top platform using DRS and a

small biological sample. However, as previously

mentioned, the system is not limited to DRS and

IFS. Other modalities, such as Raman and infrared

spectroscopy can be readily integrated on our

platform to provide additional tissue information.

4 CONCLUSIONS

A quantitative multi-modal spectroscopy scanning

platform was constructed for assessing ex vivo tissue

biochemical and morphological information. This

newly developed instrument is ideal for

characterization of surgically excised tissue margins

and provides two major benefits over the current

practice: (1) reduce patient anxiety and avoid

follow-up surgery because on-the-fly real time data

analysis can be performed; (2) reduce the probability

of missing a lesion because the whole ex vivo tissue

area can be assessed. The proof of principle has been

demonstrated in this study with bench-top prototype

for quantitative spectroscopic scanning of a

biological sample while the construction of the

compact clinical unit is in progress. For the clinical

system several instrumentation and software

advances are needed: reduction of system size,

increase in collection and analysis speed, and

improved user interface with diagnostic algorithms.

Further investigation is also needed to address

the effect of excision in quantitative hemoglobin and

fluorophores measurements. For instance, excised

specimens contaminated by the presence of surface

blood may absorb the majority of reflected light,

significantly reducing the reflectance signal

(Volynskaya, 2008). In addition, some fluorophores

(e.g. collagen) might be stable in excised tissue,

whereas others (e.g. NADH) might degrade over

time, precluding an accurate extraction of its

concentration.

The extraction of quantitative optical parameters,

such as hemoglobin and collagen, has proven to be

helpful for the differentiation of normal and

malignant tissues. A recent study from Volynskaya

et al. (2008) has successfully demonstrated that

higher hemoglobin concentrations and higher

collagen values were more likely to be found in

ductal carcinoma of breast tissue than in normal

breast tissue, and thus could be used as diagnostic

parameters.

The ex vivo spectroscopic scanning platform

concept is not restricted to only DRS and IFS and

should be extended in the future to other optical

modalities in order to gather additional and

complementary diagnostic information.

ACKNOWLEDGEMENTS

This research was supported by the National

Institute of Health (grants P41-RR02594 and R01-

CA97966) and the Portuguese Foundation for

Science and Technology under the MIT|Portugal

Program (SFRH/BD/38978/2007). The authors

gratefully acknowledge Ramachandra R. Dasari for

all his support during this research.

REFERENCES

Allweis, T. M., Kaufman, Z., Lelcuk, S., Pappo, I., Karni,

T., Schneebaum, S., Spector, R., Schindel, A.,

Hershko, D., Zilberman, M., Sayfan, J., Berlin, Y.,

0.5

1.5

2.5

3.5

4.5

BIODEVICES 2012 - International Conference on Biomedical Electronics and Devices

278

Hadary, A., Olsha, O., Paran, H., Gutman, M.,

Carmon, M., (2008). A prospective, randomized,

controlled, multicenter study of a real-time,

intraoperative probe for positive margin detection in

breast-conserving surgery. Am J Surg 196:483-489

Bard, M. P. L., Amelink, A., Skurichina M., Noordhoek,

Hegt

, V., Duin, R. P., Sterenborg, H. J., Hoogsteden,

H. C., Aerts, J. G., (2006). Optical spectroscopy for

the classification of malignant lesions of the bronchial

tree. Chest 129:995-1001

Chang, S. K., Mirabal, Y. N., Atkinson, E. N., Cox, D.,

Malpica, A., Follen, M., Richards-Kortum, R., (2005).

Combined reflectance and fluorescence spectroscopy

for in vivo detection of cervical pre-cancer. J Biomed

Opt 10:024031

Georgakoudi, I., Jacobson, B. C., Van Dam, J., Backman,

V., Wallace, M. B., Muller, M. G., Zhang, Q.,

Badizadegan, K., Sun, D., Thomas, G. A., Perelman,

L. T., Feld, M. S., (2001). Fluorescence, reflectance,

and light scattering spectroscopy for evaluating

dysplasia in patients with Barrett's esophagus.

Gastroenterology 120:1620-1629

Haka, A. S., Volynskaya, Z., Gardecki, J. A., Nazemi, J.,

Lyons, J., Hicks, D., Fitzmaurice, M., Dasari, R. R.,

Crowe, J. P., Feld, M. S., (2006). In vivo margin

assessment during partial mastectomy breast surgery

using raman spectroscopy. Cancer Res 66:3317-3322

Hamberg, L. M., Hunter, G. J., Kierstead, D., Lo, E. H.,

González, R. G., Wolf, G. L., (1996). Measurement of

Cerebral Blood Volume with Subtraction Three-

dimensional Functional CT. AJNR Am J Neuroradiol

17:1861-1869

Keller, M. D., Majumder, S. K., Kelley, M. C., Meszoely,

I. M., Boulos, F. I., Olivares, G. M., Mahadevan-

Jansen, A., (2010). Autofluorescence and Diffuse

Reflectance Spectroscopy and Spectral Imaging for

Breast Surgical Margin Analysis. Lasers Surg Med

42:15-23

Krishnaswamy, V., Hoopes. P. J., Samkoe, K. S., O’Hara,

J. A., Hasan, T., Pogue, B. W., (2009). Quantitative

imaging of scattering changes associated with

epithelial proliferation, necrosis, and fibrosis in tumors

using microsampling reflectance spectroscopy. J

Biomed Opt 14:014004

Mino-Kenudson, M., Brugge, W. R., Puricelli, W. P.,

Nakatsuka, L. N., Nishioka, N. S., Zukerberg, L. R.,

Misdraji, J., Lauwers, G. Y., (2005). Management of

superficial Barrett's epithelium-related neoplasms by

endoscopic mucosal resection: clinicopathologic

analysis of 27 cases. Am J Surg Pathol 29:680-686

Muller, M. G., Georgakoudi, I., Zhang, Q. G., Wu, J.,

Feld, M. S., (2001). Intrinsic fluorescence

spectroscopy in turbid media: disentangling effects of

scattering and absorption. Appl Opt 40:4633-4646

Muller, M. G., Valdez, T. A., Georgakoudi, I., Backman,

V., Fuentes, C., Kabani, S., Laver, N., Wang, Z.,

Boone, C. W., Dasari, R. R., Shapshay, S. M., Feld,

M. S., (2003). Spectroscopic detection and evaluation

of morphologic and biochemical changes in early

human oral carcinoma. Cancer 97:1681-1692

Novis, D. A., Zarbo, R. J., (1997). Interinstitutional

comparison of frozen section turnaround time. A

College of American Pathologists Q-Probes study of

32868 frozen sections in 700 hospitals. Arch Pathol

Lab Med 121:559-67

Ramanujam, N., Brown, J. Q., Bydlon, T. M., Kennedy, S.

A., Richards, L. M., Junker, M. K., Gallagher, J.,

Barry, W. T., Wilke, L. G., Geradts, J. (2009).

Quantitative Spectral Reflectance Imaging Device for

Intraoperative Breast Tumor Margin Assessment. In:

31st Annual International Conference of the IEEE

Engineering in Medicine and Biology Society.

EMBS’09, Minneapolis, Minnesota, USA, Sept. 2-6,

2009, pp 6554-6556

Tunnell, J. W., Desjardins, A. E., Galindo, L.,

Georgakoudi, I., McGee, S. A., Mirkovic, J., Mueller,

M. G., Nazemi J., Nguyen, F. T., Wax, A., Zhang, Q.

G., Dasari, R. R., Feld, M. S., (2003). Instrumentation

for multimodal spectroscopic diagnosis of epithelial

dysplasia. Technol Cancer Res Treat 2:505-514

Volynskaya, Z., Haka, A. S., Bechtel, K. L., Fitzmaurice,

M., Shenk, R., Wang, R., Nazemi, J., Dasari, R. R.,

Feld, M. S., (2008). Diagnosing breast cancer using

diffuse reflectance spectroscopy and intrinsic

fluorescence spectroscopy. J Biomed Opt 13:024012

Yu, C., Lau, C., O’Donoghue, G., Mirkovic, J., McGee,

S., Galindo, L., Elackattu, A., Stier, E., Grillone, G.,

Badizadegan, K., Dasari, R. R., Feld, M. S., (2008).

Quantitative spectroscopic imaging for noninvasive

early cancer detection. Opt Express 16:16227-16239

Zonios, G., Perelman, L. T., Backman, V. M., Manoharan

R., Fitzmaurice, M., Van Dam, J., Feld, M. S., (1999).

Diffuse reflectance spectroscopy of human

adenomatous colon polyps in vivo. Appl Opt 38:6628-

6637

DEVELOPMENT OF AN EX VIVO QUANTITATIVE SPECTROSCOPIC SCANNER

279