integrating sphere measurements of tissue optical...

LUND UNIVERSITY

PO Box 117221 00 Lund+46 46-222 00 00

Integrating Sphere Measurements of Tissue Optical Properties For Accurate PdtDosimetry

Nilsson, A. M. K; Berg, R; Andersson-Engels, Stefan

Published in:Laser Interaction with Hard and Soft Tissue II, Proceedings of

DOI:10.1117/12.199183

Published: 1995-01-01

Link to publication

Citation for published version (APA):Nilsson, A. M. K., Berg, R., & Andersson-Engels, S. (1995). Integrating Sphere Measurements of Tissue OpticalProperties For Accurate Pdt Dosimetry. In H. J. Albrecht, G. P. Delacretaz, T. H. Meier, R. W. Steiner, L. O.Svaasand, M. J. C. vanGemert, & A. Katzir (Eds.), Laser Interaction with Hard and Soft Tissue II, Proceedings of(Vol. 2323, pp. 47-57). SPIE. DOI: 10.1117/12.199183

General rightsCopyright and moral rights for the publications made accessible in the public portal are retained by the authorsand/or other copyright owners and it is a condition of accessing publications that users recognise and abide by thelegal requirements associated with these rights.

• Users may download and print one copy of any publication from the public portal for the purpose of privatestudy or research. • You may not further distribute the material or use it for any profit-making activity or commercial gain • You may freely distribute the URL identifying the publication in the public portal

http://dx.doi.org/10.1117/12.199183

http://portal.research.lu.se/portal/en/publications/integrating-sphere-measurements-of-tissue-optical-properties-for-accurate-pdt-dosimetry(ab0efa58-dc6b-4164-aa47-6de6e43afda8).html

Integrating sphere measurements of tissue optical propertiesfor accurate PDT dosimetry

Annika M. K. Nilsson, Roger Berg andStefan Andersson-Engels

Department of Physics, Lund Institute of Technology,P.O. Box 118, S-221 00 Lund, Sweden

[email protected]

ABSTRACT

A set-up with an integrating sphere and a narrow-beam arrangement was used in order toderive the optical properties in vitro of 1 mm thick tissue slabs. The measuredmacroscopic quantities, the reflectance and the total transmittance, were correlated to thetissue optical properties by Monte Carlo simulations. Mie calculations were performed tobe able to calibrate the set-up with a solution of latex spheres and ink. Finally, the opticalproperties of rat liver samples were measured, before and after photodynamic therapy,showing approximately a 40 % increase of the absorption coefficient at 650nm due to thetreatment.

1. INTRODUCTION

The tissue optical properties, which characterise the interaction between light and tissue,can be described by the g-factor, scattering (i) and absorption (ia) coefficient. Whenignoring inelastic scattering and fluorescence, a photon interacting with the tissue has twooptions at every interaction site; either to be absorbed or to be further scattered. Theabsorption and scattering coefficient is defined as the probability of absorption andscattering, respectively, per unit infinitesimal path length. The attenuation coefficient,is given by the sum of the absorption and scattering coefficient. Finally, the anisotropyfactor, g, specifies the average cosine of the scattering angle, where the two opposite lightdistributions due to isotropic and totally forward scattering are characterised by g=O andg=1, respectively.

These tissue characteristics are important for all kinds of medical laser applications, tounderstand the interaction mechanisms between light and tissue. The knowledge of lighttransport in tissue is particularly essential for dosimetry in conjunction with photodynamictumour treatment, in order to follow and control the light interaction volume. It istherefore important to study what happens to the g-factor, scattering and absorption

O-8194-1656-8/95/$6.OO SPIE Vol. 2323 / 47Downloaded from SPIE Digital Library on 08 Jul 2011 to 130.235.188.104. Terms of Use: http://spiedl.org/terms

coefficient during the treatment irradiation. If the treatment causes changes to the opticalproperties, the treatment volume, and hence the effects, will alter.

Various optical integrating sphere techniques have been developed to quantify tissueoptical properties, and a number of studies have recently utilised these techniques13.

2. MEASUREMENTS OF TISSUE OPTICAL PROPERTIES -MATERIAL AND METHODS

The three optical quantities of tissue characteristic, g, .t5 and were determined in athree step procedure. A 1 mm thick tissue slab was put between two microscope objectglasses (1 mm thickness), with 1 mm thick glass spacers in-between. The total thickness, 3mm, of the resulting glass cuvette, was verified by a vernier calliper. The cuvette wassequentially placed at the two sample positions of the integrating sphere, measuring thereflectance and the total transmittance, respectively. Following those measurements, thecollimated transmittance was measured with a narrow-beam technique. Finally, thecorresponding g, and ta were derived by spline interpolation in a table of Monte Carlosimulated data.

48 / SP1E Vol. 2323

Figure 1. The total transmittance and the reflectance were measured with the tissue samplesituated at the integrating sphere position 1 and 2, respectively.

Set-up for Integrating sphere measurements

Optical fibre

Downloaded from SPIE Digital Library on 08 Jul 2011 to 130.235.188.104. Terms of Use: http://spiedl.org/terms

2.1 Integrating sphere and narrow-beam measurements

Figure 1 shows the set-up for integrating sphere measurements. As a light source a 75 Whigh pressure Xenon lamp was used in combination with a monochromator to select thewavelength. The light was guided to the integrating sphere in a quartz fibre with thediameter 200 .tm. A lens followed by an aperture stop formed a collimated light beamwith a diameter of 5 millimetres. The beam reached the tissue slab either at the entry portof the integrating sphere for the total transmittance measurements, or at the exit port forthe reflectance measurements. The integrating sphere by Oriel was 20.3 cm in diameterand its two beam ports, here called entry and exit port, was opposite to one another andhad a diameter of 2.54 cm. The inner surface was covered with a highly reflectingmedium, barium sulphate, which reflects the entering light so that the total light intensityspread in the sphere was measured by a photo diode. The diode was positioned in thesphere surface perpendicular to the entering light beam. The reflectance and the totaltransmittance were obtained by

R=RBsx(IR/If)T=RBsX(LJIf)

where RBS symbols the reflectance of barium sulphate and 'R the light intensity measuredby the photo diode with the tissue sample positioned at the exit port. 'T is the lightintensity measured with the sample at the entry port and a barium sulphate plug in the exitport. Finally, 'ref is the intensity measured without any sample, with the light beamentering the first port and the barium sulphate plug in the exit port.The specularly reflected light from the tissue sample, when placed at the exit port, wasblocked by a baffle mounted inside the sphere, so that this light could not interfere withthe measurements of the reflectance.A lock-in technique in combination with a light chopper was used to prevent thesurrounding light from disturbing the measurements.

Figure 2 shows the radial light profile of the transmitted and reflected light at the twoborders of the cuvette containing a tissue sample. This is the result of Monte Carlosimulations (see below) with g=0.909, j=9.77 mm1 and ia=O.2OS mm1, typical data ofliver tissue measured at 650 nm. In figure 2 it can be calculated that approximately 0.1-0.2% of the transmitted and reflected light are lost due to the limited port diameter. Thisensures that the photo diode of the integrating sphere did not measure too smallmagnitudes of the transmittance and reflectance. J. H. Tones et. al. pointed out in theirstudy4 that the absorption coefficient easily can be overestimated when the port-to-beam-size ratio is too low. A beam diameter of 5 mm and a port diameter of 2.54 cm, as in ourcase, did apparently not yield a too low port-beam ratio.

SPIE Vol. 2323 /49Downloaded from SPIE Digital Library on 08 Jul 2011 to 130.235.188.104. Terms of Use: http://spiedl.org/terms

Figure 2. Reflected and transmitted intensity at the surfaces of the tissue slab versus the radius. The thickerlines represent the outer surface of the sphere, which prevents the light from reaching the photo diode.

Figure 3 shows the set-up for the narrow-beam measurements, where the collimatedtransmittance of the tissue sample was obtained. The light source was the same as for theintegrating sphere set-up. By using a more sensitive detector, as a photo multiplier tube, itwas possible to spatially filter the light beam from the optical fibre hard and to keep thediameter of the aperture stop in front of the detector as small as 2 millimetres. These twoprocedures were performed to prevent the scattered light of the tissue sample to beregistered as collimated transmitted light. Another action was to keep the distance betweenthe sample and the detector as long as possible, in this case around 60 cm. The collimatedtransmittance could then be derived as

Tl=TNDxIl/IO

where TND symbols the transmittance of a neutral density filter used when measuring 1. Iis the light intensity measured by the photo multiplier tube with a reference glass cuvettefilled with water at the sample position. I is the intensity of the light transmitting thetissue sample without any interaction. The attenuation coefficient was obtained by

t=-1n(T01)/d

50 / SPIE Vol. 2323

0.60

0.50

Intensity, 0.40

reflected 0.30

(a.u.) 0.20

0.10

0

0.02 0.42 0.82 1.22 1.62 2.02 2.42 2.82

radius (cm)

2.50

Intensity, 2.00

transmitted 1.50

(a.u.) 1.00

0.50

0—I I III I I I I I

0.02 0.42 0.82 1.22 1.62 2.02 2.42 2.82

radius (cm)


with the thickness, d, of the tissue slab.

2.2 Monte Carlo simulations

Light interaction with tissue can be modelled by Monte Carlo simulations, a random walkfor photon packets. The g-factors, scattering and absorption coefficients are defined aswell as the thickness of the material layers. A photon packet, given a photon weight, issent into the tissue and the step size between each interaction position is randomised witha mean step size of (s1a)1. At the interaction sites an attenuation of the photon weightis done in proportion to the magnitude of the absorption coefficient, until it falls below athreshold level and the photon packet is terminated. If it is not, the photon packet with itsnew photon weight is further scattered. The deflection angle, 9, is randomised with theHenyey-Greenstein probability distribution5:

p(cos 9)=( 1 -g2)/2( 1 +g2-2gcos 9)213

This is done until the photon packet reaches either the first or the second boundary and isadded to the other reflected or transmitted photons, respectively, yielding the reflectanceand transmittance. Monte Carlo simulations link the g-factors, scattering and absorptioncoefficients to the measured reflectance and transmittance.

SPIE Vol. 2323 /51

Set-up for Narrow-beam Measurements

Photomultiplier tube

Figure 3. The direct transmittance was measured with a narrow-beam experiment.


10 (T)-Colour Mapped-2.92635 to -2.68748-2.68748 to -2.44861-2.44861 to -2.20975-2.20975 to -1.97088-1.97088 to -1.73201-1.73201 to -1.49314-1.49314 to -1.25427125427to 101541 0154 to 07765350 776535 to 0 537667

-0.537667 to -0.298798-0.298798+

g=O.990

Iog(R)

Figure 4. Results of Monte Carlo Simulations presented as log (R) and log (T) versus and afor g=O.99. Log (T) is colour mapped.

We used a Monte Carlo simulation program written by L. Wang and S. L. Jacques6 on aDECpc ocXP 150-computer. The tissue geometry could be defined as a multilayeredmedium, in our case a 1 mm thick tissue slab with the refractive index 1 .4 in-between two1 mm thick glass slides with refractive index 1.52. Simulations were performed with 100000 photon packets for all combinations of 17 g-factors in the range of 0.6 to 0.99, 10 p-coefficients in the range of 5 to 95mm1 and for 10 J.ta-coefficients in the range of 0.02 to5 mmd. The corresponding reflectance and transmittance were arranged in 17 tables, onefor each g-factor. An example of these tables, the one for g=0.990, is graphically shown infigure 4.

2.3 Interpolation

To find the corresponding g-factor, j- and i.a coefficients to the measured totaltransmittance and reflectance, a two dimensional spline interpolation of each g-table was

52 ISPIE Vol. 2323

0


performed. The interpolation resulted in 17 pairs of possible and ta coefficients - onepair for each g-factor table. The result is shown in figure 5.

Figure 5. The result after a two dimensional spline interpolation for T=O.480 and R=O.155among the Monte Carlo simulated data.

The -L coefficient measured at the narrow-beam experiment, was then used to do a onedimensional spline interpolation among these 17 g-factors and their - ta pairs. Finally acomplete set of optical properties was obtained.

2.4 Calibration

To ensure that the measured optical properties were correct, calibration measurementswere performed. The operating procedure was therefore accomplished for a suspension ofuniform latex microspheres and ink, yielding measured optical properties between 500 and800 nm, which were compared with reference values.0.5 ml latex suspension with 10 % solid plastic spheres and 90 % water was mixed with2.5 ml ink solution (a mixture of 0.25 ml ink and 100 ml water) and 22 ml water. Thelatex spheres from Duke Scientific Corporation had the diameter of 0.806 iim. Sphericalparticles of this size interact with light as Mie scatterers and by Mie calculations the g-factor and the scattering coefficient can be derived7. We used a Pascal program by J. R.Zijp and J. J. ten Bosch8 to do the Mie computations. The input parameters were the sizeparameter

and the relative refractive index

x=2lcnmedr/?Lvac

m=np/nmed

SPIE Vol. 2323 / 53

T-measured: 0.480141

R-measured: 0.154815

g0.6000.6270.6500.6750.7000.7250.7500.7750.8000.8250.8500.8750.9000.9250.9500.975

Ms-3.657182757-2.589222086-1.7 14502963-1.600116122

-0.7041435-0.223778726

0.6124958881.1739408672.1445826822.911493081

4.2437808 15.6814104497.4023474659.57862903713.2412603777 t)1771'7'U

pa0.1115584160. 1139704110.1148329850.1187431870.1186318630.1220130420.1251802480.127384569

0.129129220.1351848710.1392580170.1405970940.1415280510.1426429750.1409150850.137497004

SumSq2.38058E-097.24277E- 101 .96358E-098.62029E- 102.42759E-092.52929E-092.20827E-094.79625E-094. 15582E-101.35416E-099.93728E-103.86109E- 109.08169E-1 13.37822E-1 15.60468E- 101.85387E-09

T-Interp.0.480141002

0.4801410.4801410.480141

0.4801410020.4801410020.4801410020.480141004

0.4801410.4801409990.480140999

0.4801410.4801410.480141

0.4801410010.480141

R-Interp.0.1548149990.1548150010.1548150020.1548150010.1548149990.1548149990.1548150010.154815002

0.1548150.1548150.1548150.1548150.1548150.1548150.154815

0.154814998


where med refractive index of the medium surrounding the particlesapart refractive index of the particler - radius of the particles2vac the wavelength of the interacting light in vacuum.

The refractive index, n, of polystyrene (latex) was found in reference 9

n=1.5683+1O.O87 x1011/ X2 vac (7vac in cm)

and the density p=l.O57 g/cm3 in reference 10.

The Mie calculations resulted in the g-factor and scattering coefficient of a suspension oflatex spheres without any absorber present and the absorption coefficient of the pure inksolution was obtained by absorbance measurements. It was assumed that the ink solutionwas a pure absorber as well as the latex sphere solution a pure scatterer. Therefore thereference values g, and ta for the mixture of the latex spheres and ink solution wasaccepted to be the Mie calculated scattering coefficient and g-factor of the latex spheresand the absorption coefficient of the pure ink solution.

0.9 a 7 0.3I0.25

0.8 6.5 S measured0.2 value

0.7 6 0.15 • referencevalue

0.10.6 5.5

0.05

0.5 1 5 1 0 I

g Ps(mm1) a(mm1)

Figure 6. Measured optical properties of the latex sphere - ink solution compared with reference values.

The measured optical properties of the latex sphere and ink solution were compared withthe reference values from the Mie calculations and absorbance measurements. The resultshowed good agreement, see figure 6, particularly for the g-factor and the scatteringcoefficient, where it was within 1-3%. The absorption coefficients had rather highstandard deviation, which other authors11 have commented as a result of the difference inmagnitude of the absorption and scattering coefficient. The a0 icients derived fromabsorbance measurements of the ink only, were within the limits of the standarddeviations.

54 / SPIE Vol. 2323


3. MEASUREMENTS OF TISSUE OPTICAL PROPERTIESIN CONJUNCTION WITH PHOTODYNAMIC THERAPY

Measurements of the optical properties of rat liver tissue, treated with photodynamictherapy, have been performed using the integrating sphere technique described above. Theoptical properties of the treated and the untreated area were measured and compared.

3.1 Operating procedure

For the experiments nine Spraque-Dawley rats were injected i. v. with 30 mg/kg bodyweight -amino-levulinic acid (ALA) 2.5 h prior to the treatment. The abdominal wall ofthe rat was cut open to expose the liver lobe. A circular region with a diameter of 1.5 cmwas irradiated with a light energy density of 60 J/cm2 at 635 nm. During the irradiationthe power density was kept well below 100 mW/cm2 to avoid any hyperthermic effect onthe tissue. As a light source a dye laser pumped with an intracavity frequency-doubledNd:YAG laser, Q-switched at a pulse repetition rate of 4 kHz, was used. Immediatelyfollowing each irradiation the treated tissue was resected and a 1 mm thick slice of thesuperficial tissue was cut and placed between two microscope object glasses with 1.0 mmspacers in-between. Two to four minutes after PDT, the optical properties of the treatedarea followed by an untreated area were derived with the technique described above.

1 12 0.3 .U _____________• 0.25

U before PDT0.8 0.2

I • after PDT10 0.15 ___________

0.6 0.1

0.05

0.41 8 0 1

g 5(mnr1) Jla(mm1)

Figure 7. Measured g-factor, scattering and absorption coefficient for a rat liver lobe at 650 nmbefore and after photodynamic therapy.

3.2 Results

Figure 7 shows typical values of the optical properties at 650 nm, measured before andafter PDT, for one of the nine liver lobes. The change of the g-factor in percent, Ag, iscalculated as:

SPIE Vol. 2323 / 55Downloaded from SPIE Digital Library on 08 Jul 2011 to 130.235.188.104. Terms of Use: http://spiedl.org/terms

and corresponding for At and Ata. The mean values and standard deviations of thechanges for the nine liver samples are shown in figure 8, indicating no change of the g-factor and scattering coefficient. However, the absorption coefficient changedapproximately 40 % during the treatment. The standard deviation of the Asia 5 ratherlarge, which partly can be explained by the small magnitudes of the absorptioncoefficients.

807060

Change, % 50403020l0ft — I I I

g i.ts

Figure 8. The change of the optical properties at 650urn during PDT.

4. DISCUSSION

An increasing a coefficient during PDT yields a decreased light penetration, which willaffect the treatment. This increase is important to know in order to be able to calculate thelight dose in conjunction with the treatment. Further studies are needed to investigate whatcauses the increased absorption coefficient. One suggestion is that it is due to blood stasisin the treatment region as a result of the treatment.A. Castellani eta!. describe in their article12 the damage of the tissue micro circulationduring PDT, which starts with microagglutination of the blood cells followed by bloodstasis. The blood vessels turn dilated and full of red blood cells, which most likely leads tohigher light absorption. This is in accordance with the increasing absorption coefficient,seen in our experiment.

5. ACKNOWLEDGEMENTS

The authors wish to thank Lexin Liu for obtaining the tissue samples and Anders Perssonfor helping out with the spline interpolation routine. This work was financially supportedby the Swedish research council for engineering sciences. The support from professorSune Svanberg is gratefully acknowledged.

56 / SPIE Vol. 2323Downloaded from SPIE Digital Library on 08 Jul 2011 to 130.235.188.104. Terms of Use: http://spiedl.org/terms

zc 1 ZEZ IA JIdS

€96T 'OT66 dd '98 J0A '•2wH P1d f 'UOi21jflOJTOOJiW pooq uo uuqdiodowq jo iojp oTartuApo1oq,, 'joiouo pui °d d o 'iujs y 66T '61 dd 'j TOA 'p44 .Jns

sisir-j 'uumos souqu UpJO1Oq :wu 1'90T pui IIIU 9 ? JOWfl2 Xfl JA!J 11J JO swdoJ poudO,, 'PH t1 •f PU sJpI1V •Jk iULT)jDi M I iJqSJjji U1A )J J J

Lc6T 'LT dd 'j 10A 'S WJ f 'IIH M U1 qfl T1 L 0T 6c61 '8O dd 'pj jOA '!°S iI°J •1 'ijqs •J .u pu •f •3 'U1flUP[ J f 6

661 'V'1 'c69T -169T dd ' TOA 'UUUiU 2d0 'SUOi2jflOO !F\I uiioj.id o wioid !oSM,, 'qoso u ç ç pui dir

Z66T id '9LT -OL1 dd •jOA '2d0 iddV 'SUOiUjflO !V1 WOJJ uoiun.uixoiddt jdtuis

1 :sjotpd Ioiqds JO saIn2x!UJ JOJ sp.idoid ponp',, 'juipoj H •Ik DU1 AJ9 •f 'jflj,S4 p •FTtI •tI '200!S .V •Fs1 d 'di!z I •1 'SPflOU.flV •D •f 'JJJO I L

Z66T 'J2U jou uosjpui ci ii 'sixj JO A2ISJA!Ufl 'D ppuis ui sgnssi paIic1I-!2Inw U! i.iodsuiJ1 2qj JO U!jOW OIJlD 2uOFs1,, 'snboiç .'-i •s pu UAt .rj 9

Ti'6T '8-OL dd

'6 10A '•f sqdoJ2s 'Axj u uopu SflJJTU,, 'U!2SUaI9 'TE I '/AU o i i'66T 'Lcz-6z dd 'N 1OA in sisj 'snbTuqo ou2uJo2oqdoJ2ods qm IUT,OJOO uogthosqi JO UOUJflSJAO :suuainsw

Ai.idoid pido nssi,, 'ipunoj, js pu zisp3 J 'qop r y 'sauoi F1 i 066T 't7T-LTJ dd 'ç J0A '101H

PN sAq 'puju u u snss siiq uinnnq pssp pu JIUJJOU

Jo sii.idoid poud,, 'IU1 TE 0 UOSJUd S V 'UUcM U ''d 9 A Z66T id 'T9TZ9dd '6 i0

oos idO 1 'jdams U qM saiqds uu1Jiu! '2Jm9 U1A j f pim jqd y g 'ioquJ 'j j 1 F 'so f j 'uLI)pid

066T °U 'c81z-991z dd '9 IoA "uoJ2oJ !.urnuO i aii 'snss poiojoiq Jo s.idoid jl?oudo Jo M!AaI V,, '°i i v pu jqd 'uoqj j-

S3DNDIkIEDI 9


integrating sphere measurements of tissue optical...

Documents