Photochemistry and Photobiology Vol. 56, No. 3 , pp. 379-384, 1992 Printed in Great Britain. All rights rcscrved

003 1-x655/92 $05 .oo+ o m Copyright @ 1992 Pcrgarnon Press Ltd

EFFECTS OF LIGHT BEAM SIZE ON FLUENCE DISTRIBUTION AND DEPTH OF NECROSIS IN SUPERFICIALLY APPLIED PHOTODYNAMIC THERAPY OF NORMAL RAT BRAIN QUN

CHEN"4*, BRIANc. WILSON.', MARY0 . DERESKI',MICHAEL s. PATTERSON', MICHAEL C H O P P ~and ~ ~ FREDw.H E T z E L ' ' ~

'Departments of Radiation Oncology and *Neurology, Henry Ford Hospital, Detroit, MI 48202, USA, .'Medical Physics, Hamilton Regional Cancer Center, and McMaster University, Hamilton, Ontario, Canada and 4Department of Physics, Oakland University, Rochester, MI 48309, USA (Received 22 October 1991; accepted 11 February 1992)

Abstract-The light fluence distributions of 632.8 nm light incident on the exposed surface of normal rat brain in vivo have been measured using an interstitial, stereotactically-mounted optical fiber detector with isotropic response. The dependence of the relative fluence rate on depth and the spatial distribution of fluence were compared for incident beam diameters of 3 and 5 mm. The fluence rate at depth of 1-6 mm along the optical axis within the brain tissue was approximately 70% greater for a 5 mm diameter beam than for a 3 mm beam, at the same incident fluence rate, although the plots of the relative fluence rate vs depth were parallel over the depth range 1-6 mm. The depths of necrosis resulting from photodynamic treatment of brain tissue using the photosensitizer Photofrin and irradiation by 632 nm light with 3 and 5 mm incident beams were also measured. The observed difference in necrosis depths was consistent with the measured difference in fluence. The importance of beam size in photodynamic treatment with small diameter incident light fields is discussed.

face irradiation, the fluence rate, @(d),at depth d below the surface on the central axis of the irradiating beam, can be described by

INTRODUCTION

Photodynamic therapy (PDT) has been used to treat various types of tumors (Dougherty, 1984). To apply this therapy, a photosensitizer (e.g. Photofrin, Quadralogic Technologies, Vancouver, BC) is first administered to the patient. After typically 24-72 h, which allows selective retention of the drug by tumor, light irradiation of a proper wavelength is applied to the tumor site to activate the drug. This elicits a series of photochemical reactions resulting in the production of cytotoxic agents, such as singlet oxygen, which may achieve local tumor cell killing. Detailed reviews of PDT can be found in Wilson and Jeeves (1987), Kaye et al. (1988) and Gomer et al. (1989). Optical dosimetry is clearly an important factor for successful PDT treatment. Light may be delivered either interstitially, using optical fibers, or superficially with an external incident beam. For the latter, which is the focus of this paper, the light "dose" involves the parameters of delivered fluence rate (mW cm-*), irradiation time (s) and irradiation field area (cm2). These allow the fluence (J cm-*) and total delivered energy (J) to be calculated. However, these parameters do not provide any information on how the energy is distributed and absorbed within the target tissue. Light propagation in tissue has been intensively studied in the last 10 years (see reviews by Wilson and Patterson, 1986; Patterson et al., 1991). In sur'To whom correspondence should be addressed. P#

Y114

where dcffis the effective penetration depth and @ 1 due to backscattered photons. Equation (1) applies only for d 3 1-2dcff,corresponding to the so-called diffusion region; close to the surface the simple exponential form breaks down. The dependence of d,, on tissue absorption and scattering has been described by diffusion theory (Star et al., 1988; Flock et al., 1989). The dependence of k, and hence of the absolute fluence rate deep in the tissue, on field size has been examined experimentally in vitro (Marijnissen and Star, 1987) and by Monte Carlo modelling (Flock et al., 1989). However, little work has been done on the effect of field size on fluence in vivo or the resulting biological consequences. In this paper, fluence-depth measurements made in normal rat brain in vivo are presented for two different incident beam sizes, over which significant changes in k would be expected, and compared with the measurement of PDT-induced depth of necrosis in the tissue. The objective of the study was to demonstrate the field-size effect in vivo and the

379

380

QUN

CHENet al.

direct correspondence of this with a quantifiable biological end-point. MATERIALS AND METHODS

male Fisher rats (220-330 8) were used' surgical procedures were performed under general anesthesia using i+,, injection of Ketamine (44 mg/kg) and Xylazine (13 mglkg). For anesthetic maintenance, half doses of the anesthetic drugs were administered every half hour or as needed. The animal body temperature was maintained at 36°C 2 0.5"C using a heating pad. mearuremenL A total of rats were Opticalfluence studied. No photosensitizer was administered to these animais. After anesthesia the head was immobilized and a midline incision made to expose the skull. Craniotomy was performed to expose the dural surface in 2 regions: (a) a 5 m m diameter area centered on the right hemisphere 3 mm to the coronal and off the for light irradiation; (b) a 1.5 by 8 mm area extending from midline to jaw, 3 mm posterior to the coronal suture, for optical probe insertion [ ~ i I ~( ~.) ] ne . dura was left intact, For convenience, the irradiation beam was provided by a 10 mW He-Ne laser. It was assumed that the optical properties at the wavelength of the He-Ne laser (632.8 nm) are not significantly different from those at the wavelength of the dye laser used in the PDT treatment (632 2 2 nm). The output light was coupled into a 400 micron optical fiber containing a distal output microlens to produce uniform irradiation. The irradiating beam had a divergence of about 30". The irradiating fiber was mounted vertically above the 5 m m craniotomy [Fig. l(b)]. Circular apertures were used to produce 3 or 5 mm diameter irradiated areas on the exposed dural surface, The incident light fluence rate (420 mW cm-2) delivered to the brain surface remained constant during each experiment, NO heating of the tissue was produced at this incident fluence rate, The light detecting probe consisted of an isotropic optical fiber (Laserguide, CA) within an 18 gauge biopsy needle. The 800 micron diameter isotropic fiber tip protruded slightly from the needle end. This probe was mounted on a micromanipulator, and the fiber tip was inserted into the brain tissue through the 1 by 8 mm slot on the skull, with a positional resolution of 0.1 mm in all 3 directions of the micromanipulator. The accuracy of the fiber tip position with respect to the irradiated brain sur-

face was S 1 mm. A total of 4-6 co-planar probe insertions (tracks) were made in each animal. Measurements started along the track at the greatest depth from the irradiated surface, so that damage caused by insertion would minimally influence the measurement made on the subsequent tracks [Fig. l(b)]. The light collected by the detecting fiber was measured as the mean photon rate, using a photomultiplier tube (Hamamatsu Japan) connected to a loo MHz sing1e photon counter (Photochemical Research Associates, Ontario, Canada). The were made as the probe was advanced in steps of 0.5 mm. The vertical separation of the tracks was 1 mm. The ambient light level, other than the laser was PhorodYamic rherapy lesion deprh measurement. The Photodynamic sensitizer was prepared according to the manufacturer's instructions to a concentration of 1 mg/mL. The drug was administered i.p. to 5 animals (12.5 mg/kg). After 48 h, each animal was anesthetized. Bilateral craniOtomY (5 mm diameter), centered 3 mm posterior to the coronal Suture and 3 mm from the sagittal suture, was performed to expose the dural surface for light irradiation. The treatment light was Provided by a dye laser (model CR599) Pumped by an argon-ion h e r (model INNOVA 70, both lasers from Coherent Laser h c , CA). The dye laser output light, with wavelength tuned to 632 2 nm, was coupled through a microlens-tiPPed optical fiber (Laserguide, CA). The irradiated area at the brain surface was adjusted, using optical apertures, to 3 or 5 mm diameter at each craniotomy site. The treatment irradiation geometry was, thus, identical to that for the fluence mapping above. The hemisphere of the brain receiving 3 or 5 m m diameter beam irradiation and the order of the irradiation were randomly selected. The in& dent flUenCe rate and exposure time were 100 m W Cm-' and 350 S. Both craniotomies were exposed to the air for the same length Of time. After treatment, craniotomies were covered with absorbable hemostat (Johnson &Johnson Inc, CA) and the scalp was sutured closed. The animal was kept warm and allowed to recover from anesthesia. The rats were sacrificed at 24 h post light irradiation by perfusion through the left cardiac ventricle of 250 mL 4% neutral buffered formaldehyde following vascular washout with 250 mL heparinized saline. The brains were subsequently removed and sliced into 3 m m thick coronal sections. R9559

*

Figure 1. Experimental setup for the optical measurement. (a) Craniotomy sites shown as the shaded tracks along which area, (b) optical probe within the brain, showing the first (---) and last (-) measurements were made.

Optical beam size in PDT

381

from a limited number of measured points. The relative light fluence rates along the central irradiation axis were calculated, normalizing the measurements to a reference point closest to the irradiated surface. Figures 3(a) and 3(b) demonstrate the relative optical fluence rate as functions of depth for 3 and 5 mm diameter irradiation beams, respectively. Linear regression was performed to equation (1) for all points between 1 and 6 mm below the irradiated surface. The mean values of d,, were 1.33 k 0.07 and 1.24 2 0.05 (SD) mm for 3 and 5 mm diameter surface irradiations, respectively (n = 5, r2 > 0.95). These were not significantly different. The mean relative photon count rates along the central axis were computed for the 3 and 5 mm diameter beams as shown in Fig. 4. As expected, these were higher at all depths (1-6 mm) for the larger beam. This was also found for all 5 animals individually at all depths. The differences in the relative light fluence rate distributions for the two cases are also demonstrated in Fig. 2. For example, the two contour lines labeled zero represent the

The brain slices corresponding to the area directly beneath the irradiated surface (typically 2 or 3 slices) were processed for histologic evaluation. Each was placed in paraffin and a 6 pm section was cut and collected every 0.5 m m from each brain slice. The paraffin sections were then stained with hematoxylin and eosin for light microscopic examination. Measurements of the lesion depth were taken perpendicular to the pial brain surface with an eyepiece micrometer on the section exhibiting maximal depth to the point where no cellular alterations were evident. Coronal section size was normalized to fresh tissue measurements, eliminating any tissue preparation size artifacts. This procedure was carried out for the PDT lesions in both hemispheres for each animal. RESULTS

Optical jluence rate measurement

Figure 2 shows representative logarithmic contour plots of light fluence rate distribution in brains irradiated by 5 and 3 mm diameter beams. The light fluence rate was attenuated by l/e across the shaded zones. The contour lines near the edges of the plotted region may not be accurate since they were generated

1 -

2S-

4-

i

:?.-

S-

6-

3 mrn BEAM 0

111111111 I

I

I

I

1

I

I

SURFACE I

I

I

I

I

1mm

U

Figure 2. Representative contour plots of the natural logarithm of light fluence rate distributions for 3 and 5 mm diameter beams. *represents the measured points. The contour lines were generated by computer using commercial software (Surfer, Golden Software Inc, Golden, CO). Arrows show the incident light.

382

QUN

3 mm

I I! 8 b a W U W

-2.50

.

2

-3.50 .

b

2

\

AVERAQED ATTENUATION DEPTH:

0

.

EY

0 -

(a) -1.50

K

CHENet al.

0

1 -

-2

2

3

4

0 50

5

6

7

8

5 mm

-4

1.24tO.05 mm

3mm beam: 1.33t0.07 mm

-

-3 1

5mm b..m:

POT INDUCED LESION DEPTH:

-

- 5 '

0

Smm b*.m'

3.22t0 24 mm

I 3mm beam: 2.53t0.51 mm '

'

'

a

.

'

'

'

'

'

'

'

-050

5Y

- 1.50

b

-2.50

Lc

3 5

-3.50

DEPTH (mm)

-4.50

DEPTH (mm)

Figure 3. Representative plots of the natural logarithm of the relative optical fluence rate along the centerline of the irradiation beam as functions of depth for (a) 3 mm and (b) 5 mm diameter irradiation beams. Each symbol represents a single animal: both beam sizes measured in each animal. The data were normalized to that at 1 mm depth in each animal.

same measured light fluence rate. The shapes of the two contour lines are clearly different, that for the 5 mm beam enclosing a much larger area (and hence volume), and reaching a greater depth.

Figure 4. Comparison of the mean photon count rates along the irradiation beam centerline for the 2 beam sizes, relative to that at 1 mm depth under the 5 mm beam. The two square symbols with their error bars represent the averaged depth of necrosis in each case. The 5 mm point has been placed on the corresponding fluence-depth line, so that, for perfect agreement of the optical and necrosis measurement, the 3 mm point should also lie on the 3 mm line.

Photodynamic therapy lesion measurement

Photodynamic therapy treated brains showed gross damage at 24 h, similar to that reported in detail previously (Dereski et al., 1989). Figure 5 shows a brain slice 24 h after PDT light treatment. Gross lesions were found in both hemispheres treated with either 5 or 3 mm diameter beams. A brain micrograph at 24 h after the light treatment (Fig. 6 ) shows that the lesion and intact regions were distinguished by sharply demarcated boundaries. Very few viable neurons were observed within

Figure 5 . A representative brain slice at 24 h after PDT light treatment.

Optical beam size in PDT

383

(Fig. 7). The lesion depth was statistically greater for regions treated with the larger diameter beam (paired t-test, P < 0.05). DISCUSSION

Figure 6. A representative brain micrograph at 24 h after the light treatment using a 5 mm beam, showin8 the lesion (upper portion) and intact tissue (lower portion) distinguished by a sharply demarcated boundary in both hemispheres. The arrows indicate the incident light.

the lesion volume, while most neurons appeared normal in the intact region. The lesion sizes at the irradiated brain surfaces were approximately equal to or greater than the incident beam size at the brain surface(s). The demarcation boundaries had shapes similar to those of the optical fluence rate contour lines deter,mined from the optical measurements (Fig. 2). The maximum distances from the irradiated brain surface to the lesion boundary were 2.53 f 0.56 and 3.22 2 0.27 mm for the 3 and 5 mm, respectively

c

= =

5I

5 mm :3.22t0.27mm

3 mm :2.53t0.~6mm

4

1

2

3

p c 0.05

4

5

ANIMAL NUMBER

Figure 7. Comparison of the maximum lesion depths in 5 animals, each treated with 5 mm and 3 mm diameter beams, evaluated by histological analysis.

This work has investigated the optical fluence rate distribution in normal rat brain and the corresponding size of PDT-generated lesions in superficial light irradiation utilizing different beam sizes. The data from the optical measurements indicate that for an identical incident fluence rate, the larger diameter irradiation beam results in a higher optical fluence rate in the diffusion range (beyond 1 mm in depth). It is, of course, well known in radiation therapy that a larger beam produces a higher central-axis radiation dose, due to the enhanced photon fluence from radiation scattered onto the axis from the larger irradiated tissue volume. It can be seen in Fig. 2, however, that there is also considerable spreading of the fluence contours beyond the geometric edge of the incident beam, due to the very high albedo of brain tissue at 633 nm (Flock et al., 1987; Cheong et al., 1990). It has been shown (Patterson et al., 1990) that the PDT threshold (corresponding to the number of photons absorbed by the photosensitizer per unit volume of tissue required to produce necrosis) should not depend on the irradiation geometry, being a measure of the intrinsic photodynamic sensitivity of the tissue. Hence, since the fluence is greater at all depths for the larger beam, the depth of necrosis should also be larger, as has been demonstrated here. For fixed fluence, the maximum lesion depths for the two beam sizes is indicated on Fig. 4, and is consistent with the difference in fluence. The effect of PDT on normal brain has been investigated by several groups (Cheng et al., 1984; Kaye and Morstyn, 1987; Dereski et al., 1989), and significant differences exist among the reported data. In particular, we previously reported a PDT treatment of normal brain using a 5 mm diameter beam and 100 J cm-*, which produced lesion 3 mm in depth (Dereski et al., 1989), whereas in a similar experiment using a 3 mm diameter beam, Cheng et al. (1984) reported minimal effect. The data reported here may partially explain this inconsistency. This work demonstrates that the irradiation beam size can be an important parameter in PDT treatment planning, at least for small beams (Star et al., 1992). This has been given little attention in most PDT reports. Usually the beam size is chosen simply to correspond to the tumor size, and, irrespective of the size, an identical incident fluence (J cm-*) is delivered to the area. Our data clearly demonstrate that this results in different light penetration patterns and consequent maximum lesion depths. When a larger beam is used, the effect of beam size

384

QUN

CHENet al.

gradually becomes less significant. In conclusion, data from the present in vivo study and other theoretical and in vitro studies strongly suggest that PDT treatments using small irradiation beam sizes should be carefully calibrated for the effect of beam size, and the delivered optical dose should be adjusted according to the actual beam size and the desired treatment area and depth. This will depend on the tissue optical properties at the treatment wavelength. Acknowledgements-This work was supported by NIH grant No. Pol-CA43892 and the National Cancer Institute of Canada. The authors also wish to thank Dr. Thomas Farrell of the Hamilton Regional Cancer Center for helpful discussions and assistance and Ms. Patricia Ruffin for the preparation of the manuscript. REFERENCES

Cheng, M., D. McKean, D. Boisvert, J. Tulip and B. W. Mielke (1984) Effects of photodynamic therapy on normal rat brain. Neurosurgery 15, 804-810. Cheong, W. F., S. A. Prahl and A. J. Welch (1990) A review of the optical properties of biological tissues. IEEE J.Quant. Electr. 26, 2166-2185. Dereski, M. O., M. Chopp, Q. Chen and F. W. Hetzel (1989) Normal brain tissue response to photodynamic therapy: histology, vascular permeability and specific gravity. Photochem. Photobiol. 50, 653457. Dougherty, T. J. (1984) Photoradiation therapy of malignant tumors.CRC Crit. Rev. 2, 83-116. Flock, S. T., M. S. Patterson, B. C. Wilson and D. R. Wyman (1989) Monte Carlo modeling of light propagation in highly scattering tissues, I: model predictions and comparison with different theory. IEEE Trans. Biomed. Eng. 36, 1162-1168. Flock, S. T., B. C. Wilson and M. S. Patterson (1987) Total attenuation coefficients and scattering phase func-

tions of tissues and phantom materials at 633 nm. Med. Phys. 14, 835-841. Gomer, C. J., N. Rucker, A. Feffario and S. Wong (1989) Properties and applications of photodynamic therapy. Radiat. Res. 120, 1-18. Kaye, A. H. and G. Morstyn (1987) Photoradiation therapy causing selective tumor kill in a rat glioma model. Neurosurgery 20, 408- 413, Kaye, A. H., G. Morstyn and M. L. J. Apuzzo (1988) Photoradiation therapy and its potential in the management of neurological tumors. J . Neurosurg. 69, 1-14. Marijnissen, J. P. A. and W. M. Star (1987) Quantitative light dosimetry in vitro and in vivo. Lasers Med. Sci. 2, 235-242. Patterson, M. S., B. C. Wilson and R. Graff (1990) I n vivo tests of the concept of photodynamic threshold dose in normal rat liver photosensitized by aluminum chlorosulphonated phthalocyanine. Photochern. Photobiol. 51, 343-350. Patterson, M. S.,B. C. Wilson and D. R. Wyman (1991) The propagation of optical radiation in tissue I: models of radiation transport and their application. Lasers Med. Sci. 6, 155-168. Star, W. M., J. P. A. Marijnissen and M. J. C. van Gemert (1978) Light dosimetry: status and prospects. J. Photochern. Photobiol. 1, 149-168. Star, W. M., J. P. A. Marijnissen and M. J. C. van Gemert (1988) Light dosimetry in optical phantoms and in tissues: I. Multiple flux and transport theory. Phys. Med. Biol. 33, 437-454. Star, W. M., B. C. Wilson and M. S. Patterson (1992) Light delivery and optical dosimetry in photodynamic therapy of solid tumors. In Photodynamic Therapy. (Edited by T. J. Dougherty and B. Henderson), pp. 333-367. Marcel Dekker, New York. Wilson, B. C. and W. P. Jeeves (1987) Photodynamic therapy of cancer. In Photomedicine (Edited by E. BenHur and I. Rosenthal), Vol. 11, Chapter 6, pp. 127-177. CRC Press, Florida. Wilson, B. C. and M. S. Patterson (1986) The physics of photodynamic therapy. Phys. Med. 31, 327-360.