Lasers in Surgery and Medicine 11:331340 (1991)
Optical Properties of Human Sclera, and Their Consequences for Transscleral Laser Applications Alfred Vogel,
PhD,
Christian Dlugos, BS, Roland Nuffer, BS, and Reginald Birngruber, PhD
H. Wacker Laboratory of Medical Laser Applications, University Eye Hospital Munich, 8000 Munich 2, Germany (A. V., C.D., R.N., R.6.); Wellman Laboratory of Photomedicine, Massachusetts-General Hospital, Boston, Massachusetts 021 14 (R.B.)
The spectral dependence of the optical properties of human sclera adjacent to the limbus was investigated and related to the potentials of transscleral photocoagulation. The total transmission, absorption, and reflection, as well as the angular distribution of the transmitted and reflected light were measured at five laser wavelengths (442 nm, 514 nm, 633 nm, 804 nm, and 1,064 nm), both for noncontact and contact applications. Absorption and scattering coefficients were determined using the KubelkaMunk model for light propagation through a scattering tissue. The scleral transmission is only 6%at 442 nm but increases to 35% at 804 nm and to 53%at 1,064 nm. The absorption is high at short wavelengths with 40%at 442 nm but it is only 6%at 804 nm and 1,064 nm. The reflection is generally higher than 40% and shows little wavelength dependence. The transmitted light is scattered diffusely at short wavelengths, but at 804 nm and 1,064 nm it exhibits a fairly narrow angular distribution in forward direction. Fiber contact leads to an increase of transmission, with a factor of 3.5 at 442 nm, of 2.0 at 804 nm, and 1.5 at 1,064 nm. Our results indicate that the diode laser (804 nm) and the NdYAG laser (1,064 nm) with contact delivery are best suited for transscleral photocoagulation. words: cyclodestruction, scleral absorption, scleral scattering, scleral transmission, transscleral photocoagulation
INTRODUCTION
Transscleral laser radiation has been applied for cyclodestructive procedures in glaucomatreatment [l-151, and for transscleral photocoagulation of the retina [4,6,18,191.Cyclodestructive procedures have been attempted with various wavelengths using ruby lasers [ll, argon lasers [4],Nd:YAG lasers [2,3,5-161, and diode lasers [17,18], whereas transscleral coagulation of the retina has been performed with argon lasers [41, Nd:YAG lasers [6], and diode lasers [18,191. Thereby, either noncontact delivery using a slit lamp or contact delivery through a n optical fiber with a flat or focusing fiber tip has been employed. Recently, special contact lenses have been developed to combine slit lamp delivery and con0 1991 Wiley-Liss, Inc.
tact application [20,211. In this study, the optical properties of the human sclera are investigated in order to provide criteria for selecting the optimal laser wavelength and delivery system for transscleral photocoagulation. The rationale of transscleral laser applications is to affect the target tissue while causing as little damage to the sclera as possible. This requires high light transmission through the sclera and low absorption within the sclera. The direc-
Accepted for publication April 17, 1991. Address reprint requests to Dr. Alfred Vogel, H. Wacker Laboratory of Medical Laser Applications, University Eye Hospital Munich, Mathildenstr. 8, 8000 Munich 2, Germany.
Vogel et al. tional distribution of the transmitted light is also of interest, because it determines the energy density at the target tissue. If the scattering 442 nm within the sclera is very strong, a focusing of the light incident at the sclera will hardly increase necd the energy concentration at the target tissue. Therefore, the total transmission, absorption, and reflection of the human sclera were measured at five laser wavelengths (442 nm, 514 nm, 633 nm, 804 nm, and 1,064 nm), and the angular distribution of the transmitted and reflected light was determined. All investigations were performed both for contact and for noncontact laser applications. 332
t
U
Voltmeter
n 442
nm
MATERIAL AND METHODS
All measurements were performed in vitro with sclera specimens obtained from fresh human autopsy eyes which had been used for keratoplasty. The sclera samples were taken from a site adjacent to the limbus and had a size of about 15 x 15 mm2. Conjunctiva, retina, and choroid were carefully removed. Thereby, the pigmented lamina fusca remained attached t o the sclera. The specimens were stored in physiological saline solution and, during the measurements, they were kept moist with saline solution. Fifteen specimens were used for noncontact measurements, and seven specimens for contact measurements. Five lasers with different wavelengths were used for the investigations: a helium-cadmium laser (RCA Electro Optics, model LD 2186) emitting at 442 nm, an argon ion laser (Spectra Physics, model 168) emitting at 514 nm, a heliumneon laser (Coherent Radiation, model 80) emitting at 633 nm, a diode laser (Zeiss, laboratory device) emitting at 804 nm, and a cw Neodymium:YAG laser (MBB Medilas 11) emitting at 1,064 nm. In the noncontact measurements, the sclera sample was illuminated with the unexpanded laser beam in all cases except that of the diode laser. The strongly divergent diode laser radiation was first coupled into an optical fiber, and the light emitted from the fiber tip was then collimated with a lens and focused onto the sample. The laser power was always limited t o 5 mW in order t o avoid changes of the optical properties of the sclera during the measurements. For the contact measurements, the laser beams were coupled into an optical fiber having a flat fiber tip with 600 p m diameter. In all contact measurements, the laser power at the distal fiber tip was 1.5 mW.
.,
He-
cd
\
\ ....I
I
b)
U
\
Photodiode
Voltmeter
Fig. 1. Experimental arrangement for the measurement of total transmission, absorption, and reflection of the sclera (a), and for the measurement of the angular distribution of the transmitted and reflected light (b).
Measurement of Transmission, Absorption, and Reflection
The total transmission, absorption, and reflection of the sclera samples were determined using a custom made integrating sphere (Fig. la). For absorption measurements, the sample has to be placed at the center of the sphere without impairing the development of a homogeneous light distribution within the integrating sphere. Therefore, the diameter of the sphere was 0.64 m, this size being large enough t o render direct absorption measurements possible. The inner surface of the sphere had a coating of barium sulfate (BaSO,). The reflectivity of the coating was determined for each laser wavelength by comparison with the known reflectivity of a BaS0,-standard provided by Zeiss. The reflection coefficients ranged from 90%at 442 nm t o 92.5%at 1064 nm. All light intensity measurements were performed with a silicium photodiode (TS Optoelectronic, model SD 444), the signal of which was amplified
Optical Properties of Human Sclera
(TS Optoelectronic, model SDC) and displayed with a digital voltmeter (Schlumberger, model A 203). For the transmission measurements, the sclera specimen was placed at position P I in the wall of the integrating sphere as shown in Figure la, and the voltage UT at the photodiode was measured. The specimen was then replaced by a standard target with very high reflectivity, and the reference voltage UR for the empty sphere was determined. The total transmission of the specimen is given by
333
sity observed in the direction given by the angle a.It is assumed in equations (2a) and (2b) that the spatial distribution of the scattered light is radially symmetric around the axis of the laser beam. To calculate from the measurement data, one must consider that the radiant intensity is only determined at discrete anlges ai in steps of 10". Therefore, equations (2a) and (2b) must be replaced by the approximation
+
9
Q~
=
271C I ( ~Sinai J ~
c t
(34
i=l T T
UT
T=-p UR
(1)
19
Q~
whereby p denotes the reflection coefficient of the standard target. The total absorption A and reflection R of the sclera specimens were measured in a corresponding manner, with the specimens being placed at positions Pa and P,, respectively. Goniometric Measurements
The angular distribution of the transmitted and reflected light in the far field was determined with the setup shown in Figure lb. The laser beam was incident normally onto the sclera sample. The photodiode could be moved around the sclera sample at a constant distance of 115 mm. The diameter of the sensitive area of the diode was 10 mm, resulting in an angular resolution of 5". Measurement values were recorded every 10". The data concerning the angular distribution of the radiant intensity were also used t o calculate the total radiant flux +T and +R of the transmitted and reflected light. Theoretically, the ratio +T/+R derived from the goniometric measurements equals the ratio TIR obtained by the measurements with the integrating sphere. A comparison of the respective values allows for a cross-check of the validity of both measurements. The total radiant flux of the transmitted and reflected light is related t o the angular light distribution by
=
2-rr
C ~ ( qSinai )
(3b)
i=ll
where al = 0", cta = = d18 rad = 10".
lo", . . . . aI9= 180",and ha
Determination of Absorption and Scattering Coefficients
The absorption and scattering coefficients of the sclera were calculated using the KubelkaMunk model for light propagation in turbid and absorbing media [22,231. The Kubelka-Munk theory describes the propagation of a uniform, diffuse irradiance through an isotropically scattering tissue layer with no reflections at the boundaries. These assumptions, especially that of diffuse illumination of the sample, are not entirely met in our experiments. Nevertheless, the Kubelka-Munk model is a much better approximation for laser light propagation through the sclera than Beer's law analysis, which ignores the effect of scattering. Its advantage as compared t o more complicated models [231 is that the scattering coefficient S and the absorption coefficient K can be directly expressed in terms of the measured reflection R, transmission T, and thickness d of the sample:
711 2 r
J
K
0
=
(x-1)S,
n
cDR
=
27-r
I
I(a)sin a da
(2b) where
7112
Here a = 0 denotes the forward direction of the incident laser beam, and I(a)is the radiant inten-
x =
2R
(4b)
Vogel et al.
334 RefIection 60
50
-
40
-
30 -
20
-
10
I
I
400
500
I
I
600
700
800
900
lOOO[nm]llOO
Wavelength
!
1.0 400
500
600
700
800
900
1OOO[nrn]llOO
Wavelength
Fig. 2. Total reflection, absorption, and transmission of human sclera, plotted as a function of the laser wavelength. The error bars represent the standard deviation of the measurement values.
Fig. 3 . Ratio of the scleral transmission values with and without fiber contact, plotted as a function of laser wavelength. The ratio was determined for gentle and firm pressure of the fiber tip against the scleral surface.
R ESULTS Transmission, Absorption, and Reflection
cannot be enhanced by more than a factor of 1.5. This is because the transmission value surpasses 50% even without fiber contact (Fig. 2).
Figure 2 shows the results of the noncontact measurements with the integrating sphere. The values for total transmission, absorption, and reflection of the sclera are plotted as a function of the laser wavelength. In the visible range of the spectrum, the largest part of the incident light is reflected. The small wavelength dependence of the reflection corresponds t o the white appearance of the sclera. The transmission increases continuously with increasing wavelength, until above 1,000 nm it becomes larger than the reflection. At 1,064 nm, the transmission is about 53%. At wavelengths below 600 nm, more light is absorbed within the sclera than transmitted through the sclera. The absorption shows, however, a continuous decrease with increasing wavelength; at 804 nm and 1,064 nm, it amounts to only 6%. The scleral transmission increases at all wavelengths, when the laser light is delivered through an optical fiber, with the fiber tip in contact with the tissue. Figure 3 shows the ratio of the transmission values with fiber contact and noncontact delivery as a function of laser wavelength. The ratio was determined both for “gentle” and for “firm” pressure of the fiber tip against the sclera. As already observed by Rol et al. [241, the scleral transmission increases with the pressure applied on the sclera. The improvement of transmission is wavelength dependent and largest for short wavelengths. At 1,064 nm, even with firm pressure, the total transmission
Angular Light Distribution
The results of the goniometric measurements are exemplified in the polar diagrams shown in Figures 4 and 5. The examples are chosen such that the corresponding calculated ratio +T/+R of the transmitted and reflected light flux resembles as closely as possible the mean value determined for the respective wavelength. The size of the diagrams is normalized with respect t o the radiant flux incident onto the sclera. This means that in each diagram an equal distance from the point of incidence represents the same angular transmission or reflection. The noncontact measurements (Fig. 4) demonstrate that, in the visible range of the spectrum, both the transmitted and reflected light are scattered diffusely. With increasing wavelength, however, the scatter profile of the transmitted light exhibits a more and more elongated form. Thus, near infrared radiation is not only better transmitted through the sclera than visible radiation (Fig. 21, but the transmitted light is also less scattered in sideward direction. The scatter profiles for fiber contact with gentle pressure are presented in Figure 5. A comparison with Figure 4 shows that the scleral transmission is considerably higher with fiber contact than without, especially for short wavelengths. This result confirms the observations made with the integrating sphere (Fig. 3). As in
Optical Properties of Human Sclera
335 442 nrn
442 nm
514 nm 514 nm
n 633 nm
*
633 nrn
Fig. 4. Typical examples of polar diagrams for the angular distribution of radiant intensity at different laser wavelengths in the noncontact mode. The size of the diagrams is normalized with respect to the incident radiant flux.
Fig. 5 . Typical angular distribution of radiant intensity at different laser wavelengths in the contact mode with gentle pressure. The size of the diagrams is normalized with respect to the incident radiant flux. The scale is as in Figure 4.
the noncontact mode, the scatter profiles become more elongated for longer wavelengths. The scatter profiles at 804 nm and at 1,064 nm appear to be a little broader with fiber contact than in the noncontact case, when the laser beam is directed perpendicular onto the sclera sample. The broadening of the forward scattering may partly be attributable to the divergence of the light emitted from the flat fiber tip. The tissue effect of the laser light depends not only on the total transmission of the sclera, but, also on the directional distribution of the transmitted light. Only that part of the laser energy scattered into the central solid angle will contribute t o the desired effect at the application site. The directional distribution of the laser energy or laser light flux, respectively, cannot directly be read from Figures 4 and 5 containing polar diagrams of the radiant intensity in one plane only. The solid angle in which a certain radiant intensity is detected can be taken into account using equation (3). Taking the diagram shown in Figure 4 for a wavelength of 1,064 nm as an example, it is found that only 40% of the transmitted light is scattered into the central solid angle of 25" around the laser beam axis. This percentage is even smaller at shorter wavelengths,
exhibiting a less elongated scatter diagram: 35% at 804 nm, and 21% at 514 nm. Contact application leads to an increase of the part of the transmitted light scattered into the cone angle of 25". At 1,064 nm, the percentage was found to change from 40% t o 49%, when fiber contact with gentle pressure was applied. In Figure 6, the ratio +T/+R of total transmitted and reflected radiant flux calculated from the goniometric measurements is compared with the ratio T/R obtained with the integrating sphere. Theoretically, both values should be the same for each laser wavelength. The agreement is indeed very good; the differences observed are smaller than the standard deviations for both measurement series. Absorption and Scattering Coefficients
The absorption and scattering coefficients of the sclera were calculated with equations (4a)(4c), assuming a thickness of the sclera adjacent to the limbus of 0.8 mm [251. Figure 7 shows the mean values of the absorption coefficient obtained from the noncontact measurements, as well as the scattering coefficients determined in the noncontact mode and with fiber contact. The values of the absorption coefficient with fiber contact are
Vogel et al.
336 1.5
80
-
I
Pol 70 -
.-
60
$\ Reflection
-
a:
I- 1.0 -
50 -
1
40 -
ek 0.5 --
30 -
?a
20 10 -
I
400
I
I
I
I
I
500
600
700
800
900
i
I
1000[nm]l100
Wavelength
-
\ \
\. \
\ 30
\ \
Y
20
--.
'. .-.
. '. .
- - _ .- -.~.. - - _ _- ---.._ . .......... a - _ _
\
.........8 . - ..........................................
500
600
500
600
700
I
800
900
, 1000[nm]1100
Fig. 8. Comparison of the results of the present study with present study, human sclera the results of other studies. -, adjacent to the limbus, n = 15; ---,Smith and Slein 1261, human sclera adjacent to the limbus, n = 16;- -, Manson et al. 1271, origin and number of the samples not identified. -, Rol et al. 1241, human sclera with a thickness of 0.71 m m , n = 1.
DISCUSSION Interpretation of the Measurement Results
\
J.__
400
I
s
\
v)
10
I
1,064 nm. Scattering is reduced by fiber contact, especially when firm pressure is applied t o the sclera.
- \
40
__.
I
Wavelength
Fig. 6. Comparison of the measurement results obtained with the goniometer and the integrating sphere. is the ratio of the total transmitted and reflected radiant flux calculated from the data of the goniometric measurements using equation (3). T/R is the ratio of the total transmission and reflection determined with the integrating sphere. Theoretically, both values should be equal for each laser wavelength. The error bars represent the standard deviation of the measurement values. [cm-11
:ion
700
800
900
1000 [nm] 1100
Wavelength Fig. 7. Absorption coefficient K and scattering coefficient S of human sclera. K and S were calculated from the measured values of R and T using equations (4a)-(4c). K is given for the noncontact measurements, and S is plotted for the noncontact mode as well as for fiber contact with gentle and firm prescontact with gentle pressure; sure: --, noncontact mode; ---, . . . , contact with firm pressure.
not given, since they are similar to those without contact. At all wavelengths, scattering within the sclera is much stronger than absorption. Both scattering and absorption decrease considerably with increasing wavelengths; the value of the absorption coefficient drops from 8.6 cm-' at 442 nm t o about 0.6 cm-' at 804 nm, and 0.7 cm-I at
Figure 8 compares the measurement results of this study with the data for the spectral dependence of total scleral reflection and transmission reported by other groups [24,26,27]. The shape of the curves is similar in all cases, and the differences of the absolute values can be partly explained by differences in the preparation of the sclera specimen and by variations of the scleral thickness. It should be noted, however, that in the earlier studies only reflection and transmission have been measured. In two of the studies [24,261, the sum of the values for transmission and ref lection sometimes exceeds 100%.This indicates measurement inaccuracies, since the absorption also has t o be considered. In the present study, the measurement accuracy was increased by using a large integrating sphere which allowed direct absorption measurements. The deviation of the sum (R + T + A) of the mean values from 100% was smaller than 1.0% for all wavelengths. For interpreting the absorption and scattering properties of the sclera, it is helpful t o recall its composition and structure. The sclera consists mainly of water (65%), collagen (75% of dry
Optical Properties of Human Sclera 337 weight), noncollageneous proteins, and muco- portional to K4.After single scattering events polysaccharides [251. Up t o 1% are lipids (de- with d = A, the distribution of the scattered light pending on age) [251, and about 0.5% inorganic is predominantly forward-directed [391. When substances [28]. It is remarkable that the scleral multiple scattering occurs, the spatial distribuabsorption reaches up t o 40% in the visible range tion of the radiation as it passes through the samof the spectrum (Fig. 21, although the absorption ple quickly becomes isotropic (i.e., diffuse), reof collagen and other proteins is almost negligible gardless of the spatial distribution obtained for a above wavelengths of 400 nm [29]. Water absorp- single scattering. With this background, the scattion is also negligible in the visible part of the ter diagrams in Figures 4 and 5 can be inspectrum and plays a role only in the near-in- terpreted. At short wavelengths, a diffuse disfrared region with a first weak absorption peak at tribution of the transmitted light is observed 975 nm [301. One possible explanation for the indicating multiple scattering. With increasing measurement results is indicated by the fact that wavelength, the scattering becomes weaker as exthe wavelength dependence of the scleral absorp- pected theoretically, and the scatter diagrams extion (Figs. 2, 7) resembles that of melanin [311. hibit a more elongated form. The decrease of scleral transmission at This may be due t o the pigmented cells found within the sclera, particularly on the inner sur- shorter wavelengths (Fig. 2) is partly due t o the face (lamina fusca) [32,33]. Another chromophore increase of absorption but is mainly a conseabsorbing in the blue-green part of the spectrum quence of the increased scattering. Again a comis p-carotene [341. It is highly lipophilic and may parison with the optical properties of porcine be bound to some extent to the lipids contained sclera adds to the understanding of the data within the sclera. This explanation is indirectly obtained for humans. The light transmission supported by comparing the absorption character- through porcine slcera is considerably lower than istics of human sclera with that of porcine sclera. through human sclera (only 31% at 1,064 nm as Unlike humans, most other mammals retain rel- compared t o 53%in humans), and the distribution atively little p-carotene and have white fat as op- of the transmitted light is more diffuse [36]. This posed t o the yellow fat found in humans [351. can only be due t o increased scattering in the porMoreover, domestic pigs are hardly pigmented, cine sclera, since its absorption is smaller than and their sclera probably contains very little mel- that of human sclera. The assumption that the anin. The lack of both chormophores corresponds reduced transmission is caused by increased scatt o the small absorption of 6% or less observed in tering is consistent with the greater thickness of porcine sclera over the whole range of the visible the porcine sclera (1.1 mm in the limbal region spectrum [361. It should be noted that although [40] as compared with 0.7-0.8 mm in humans the scleral absorption in human sclera reaches up [25,401) and with the stronger irregularity in the t o 40% at short wavelengths, its absorption coef- order of stromal fibrils and lamellae [401. The light reflection of the sclera is basically ficient (Fig. 7) amounts to only about 1/300 of the diffuse (Fig. 4); i.e., it relies mainly on the scatvalue for the retinal pigment epithelium. The light scattering within the sclera is tering within the tissue. The regular reflectance mainly due t o the refractive index difference be- of the scleral surface amounts to only 2-4% for tween the inhomogeneously distributed collagen normal or nearly normal light incidence, and fibrils (n = 1.47) and the ground substance sur- therefore plays a minor role. Neglecting the rounding the fibrils (n = 1.36) [331. The diameter scleral absorption, the reflection should therefore of the fibrils varies between 30 nm and 300 nm, be the higher, the stronger the scattering within with an average of 60 nm [25,33]. This gives rise the sclera; i.e., it should increase with decreasing to variations of the refractive index of the sclera wavelength. This tendency is indeed observed beon a scale comparable t o the wavelength of light, tween 1,064 nm and 633 nm (Fig. 21, but in the hence t o scattering [37,381. Since the scattering short wavelength range it is covered by the instructures have a size d in the order of the wave- crease of absorption within the sclera. In porcine length of light, they cause neither pure Rayleigh sclera with its little absorption over the whole scattering (d > A) [39]. In the intermediate domain, scat- tinuously with decreasing wavelength [36]. tering varies inversely with wavelength, as demThe increase of transmission due t o fiber onstrated in Figure 7, but not as strongly as contact (Figs. 3-5) may be attributed to the disRayleigh scattering, where the variation is pro- placement of ground substance caused by the
338
Vogel et al.
pressure of the fiber tip. This causes a thinning of the sclera and a reduction of the distance between the collagen fibrils leading to modifications in the interference of the light scattered from adjacent fibrils [37]. If more water than proteins and mucopolysaccharides are displaced, the concentration of the remaining ground substance increases and its refractive index becomes closer t o that of the collagen fibrils [331. As a consequence of all effects described, the scattering is reduced (Fig. 7) so that the transmission increases and the angular distribution of the transmitted light becomes narrower. This effect is stronger when higher pressure is applied to the sclera. Clinical Consequences
nm and 77% at 1064 nm. It can be concluded from these data that the Nd:YAG laser should require two times more energy for transscleral photocoagulation in the noncontact mode than the diode laser. In the contact mode with strong pressure onto the sclera, the required energy should be 2.8 times higher for the Nd:YAG laser. When a fiber optic contact delivery system is used, the total transmission of the sclera is increased by about 50% at 1,064 nm and by 100% at 804 nm, (Fig. 3). An even higher increase of transmission is observed in the visible range of the spectrum, but this is not really significant for clinical applications because of the high scleral absorption at these wavelengths. Moreover, fiber contact leads to an increase of the percentage of light transmitted in the forward direction. Thus, contact application causes a reduction of the energy required for transscleral photocoagulation. A considerable reduction can, however, only be achieved if firm pressure is applied t o the sclera (Fig. 3) [241. Using a contact glass with a large contact area, i.e., with small pressure onto the sclera, the same energy was found t o be necessary for cyclophotocoagulation as in the noncontact mode [201. A t long wavelengths, when the transmitted light is mainly scattered in forward direction (Figs. 4, 5), the energy requirements may be further reduced by focusing the laser beam [45]. However, the focusing power of fiber tips is limited by the small difference in refractive index between glass or sapphire tips and sclera [46,47]. Regarding the transmission and absorption qualities of the sclera, the results of our study support the use of a contact delivery system for transscleral laser applications. The wavelengths best suited are in the near-infrared region, like those of the diode laser (804 nm) or of the Nd:YAG laser (1,064 nm). The use of the diode laser is thereby still restricted by limitations of the maximal laser power available. This will change, however, in the near future.
The aim in transscleral photocoagulation is t o achieve focal cyclodestruction or retinal coagulation without damaging the sclera. This requires the use of a laser wavelength exhibiting little scleral absorption, high transmission through the sclera, and a narrow angular distribution of the transmitted light. In addition, the absorption in the target tissue should be as high as possible. Our results show that the light of Nd: YAG lasers (1,064 nm) and of diode lasers (804 nm) is well transmitted through the sclera and hardly absorbed in scleral tissue. This corresponds to the observation made in various histological studies that after Nd:YAG and diode laser cyclocoagulation at moderate energy levels, the tissue effects are confined t o the ciliary body [2,8,10,12,17-19,41,42]. Only at high-energy levels (5 J in a rabbit eye) far above the coagulation threshold and at long exposure times (300 ms t o 2 s), was scleral damage found, obviously having been caused by heat conduction from the ciliary body into the sclera [8,411. Comparison of the laser effects in pigmented rabbits and albino rabbits has shown that cyclocoagulation relies on the light absorption in the ciliary pigment epithelium [1,43,441. The most important absorbing structure for retinal photocoagulation is the retinal pigment epithelium [311. In both cases, the absorbing chromophore is ACKNOWLEDGMENTS melanin. Since the melanin absorption is about We appreciate stimulating discussions with three times higher at 804 nm than at 1,064 nm F. Fankhauser and P. Rol. [311,the light of the diode laser is better absorbed within the ciliary body and the retina than Nd: YAG laser light. The total transmission of the REFERENCES at 804 nm (35%) than at is 1, Beckman H, Kinoshita A, Rota AN, Sugar HS, Transnm (53%), but this difference diminishes with fiscleral ruby laser irradiation of the ciliary body in the ber contact, if strong pressure is applied to the treatment of intractable glaucoma. Trans Am Acad Ophthalmol Otolaryngol 1972; 46:423-436. sclera. In this case, the transmission is 71% at 804
Optical Properties of Human Sclera 2. Wilenski ZT, Welch D, Mirolovich M. Transscleral cyclocoagulation using a Neodymium:YAG laser. Ophthalm Surg 1985; 16:95-98. 3. Fankhauser F, van der Zypen E, Kwasniewska S, Rol P, England C. Transscleral cyclophotocoagulation using a Neodymium:YAG laser. Ophthalm Surg 1986; 17:94100. 4. Federman IL, Ando F, Schubert HD, Eagle RC. Contact laser for transscleral photocoagulation. Ophthalm Surg 1987; 18:183-184. 5 . Devenyi RG, Trope GE, Hunter WH, Badeeb 0. Neodymium:YAG transscleral cyclocoagulation in human eyes. Ophthalmology 1987; 94:1519-1522. 6. Kwasniewska S, Fankhauser F, van der Zypen E, Rol P, Henchoz PD, England C. Acute effects following transscleral contact irradiation of the ciliary body and the retinal choroid with the cw Nd:YAG laser. Lasers Light Ophthalmol 1988; 225-34. 7. England C, van der Zypen E, Fankhauser F, Kwasniewska S. A comparison of optical methods used for transscleral cyclophotocoagulation in rabbit eyes produced with the Nd:YAG laser: A morphological, physical and clinical analysis. Lasers Light Ophthalmol 1988; 2: 87-102. 8. Heidenkummer HP, Birngruber R, Lorenz B, Gabel VP, Gao L, Tang SH. Morphologische Befunde des Kaninchenziliarkorpers nach transskleraler cw Nd:YAG Laser Koagulation. In: Wollensack J, ed. “Laser in der Ophthalmologie.” Stuttgart: Enke, 1988; pp 184-191. 9. Mehdorn E, Lucke K, Steinmetz M, Messmer E. Transsklerale Ziliarkorperkoagulation mit dem Nd:YAG cw Laser bei Sekundarglaukom. Fortschr Ophthalmol 1988; 86:102-106. 10. Hampton C, Shields MB. Transscleral neodymium:YAG cyclophotocoagulation. A histological study of human autopsy eyes. Arch Ophthalmol 1988; 106:1121-1123. 11. Klapper RM, Wandel T, Donnefeld E, Perry HD. Transscleral neodymium:YAG thermal cyclophotocoagulation in refractory glaucoma; a preliminary report. Ophthalmology 1988; 95:719-722. 12. Latina MA, Pate1 S, de Kater AW, Goode S, Nishioka NS, Puliafito CA. Transscleral cyclophotocoagulation using a contact laser probe: A histologic and clinical study in rabbits. Laser Surg Med 1989; 9:465-470. 13. Brancato R, Leoni G, Trabucchi G, Pietroni C. Contact transscleral cyclophotocoagulation with cw Nd:YAG laser in uncontrolled glaucoma. Ophthalm Surg 1989; 20: 547-551. 14. Trope GE, Ma S. Mid-term effects of Nd:YAG transscleral cyclophotocoagulation in glaucoma. Ophthalmology 1990; 97:73-75. 15. Schuman JS, Puliafito CA, Allingham RR, Belcher CD, Bellows AR, Latina MA, Shingleton BJ. Contact transscleral continuous wave Neodymium:YAG laser cyclophotocoagulation. Ophthalmology 1990; 97571-580. 16. Schuman JS, Puliafito CA. Laser cyclophotocoagulation. Int Ophthalmol Clin 1990; 3O:lll-119. 17. Schuman JS, Jacobson JJ, Puliafito CA, Noecker R J , Reidy WT. Experimental use of semiconductor diode laser in contact transscleral cyclophotocoagulation in rabbits. Arch Ophthalmol 1990; 108:1152-1157. 18. Okamoto S, Takahashi H, Fukado Y, Ozawa T. Laser diode application for transscleral photocoagulation. Laser Light Ophthalmol 1990; 3:29-37.
339
19. Peyman GA, Naguib KS, Gaasterland D. Transscleral application of a semiconductor diode laser. Laser Surg Med 1990; 10:569-575. 20. Simmons RB, Blasini M, Shields M, Erichson PJ. Comparison of transscleral Neodymium:YAG cyclophotocoagulation with and without a contact lens in human antopsy eyes. Am J Ophthalmol 1990; 109:174-179. 21. Durr V, Henchoz PD, Fankhauser F, Kwasniewska S, van der Zypen E. Results and methods of transscleral laser cyclodestruction: A new contact lens for use with non-contact systems. Lasers Light Ophthalmol 1990; 3: 123-131. 22. Kubelka P. New contributions t o the optics of intensely light-scattering materials. Part I. J Opt SOCAm 1948; 381448-457. 23. Cheong WF, Prahl SA, Welch AJ. A review on the optical properties of biological tissues. IEEE J Quantum Electron 1990; QE 26:2166-2185. 24. Rol P, Niederer P, Durr U, Henchoz PD, Fankhauser F. Experimental investigations on the light scattering properties of the human sclera. Lasers Light Ophthalmol 1990; 3:201-221. 25. Torcynski E. Sclera. In: Duane TD, Jaeger EA, eds. “Biomedical Foundations of Ophthalmology,” Vol. I. Philadelphia: J.B. Lippincott Company, 1988, pp 1-23. 26. Smith RS, Stein MN. Ocular hazards of transscleral laser radiation. I. Spectral reflection and transmission of the sclera, choroid and retina. Am J Ophthalmol 1968; 66: 21-31. 27. Manson N, Mc Carthy J, Henry L, Terry C. Transscleral tunable dye laser lesions. In: Marshall J, ed: “Laser Technology in Ophthalmology.” Amsterdam: Kugler & Ghedeni, 1988, pp 163-172. 28. Duke-Elder S. The cornea and sclera. In: “System of Ophthalmology,” Vol. IV: “The Physiology of the Eye and of Vision.” London: Henry Kimpton, 1968, pp 337-363. 29. Hillenkamp F. Interaction between laser radiation and biological systems. In: Hillenkamp F, Pratesi R, Sacchi CA, eds. “Lasers in Biology and Medicine.” New York: Plenum Press, 1980, pp 37-68. 30. Tyler JE. Optical properties of water. In: Driscoll WG, Vaughan W, eds. “Handbook of Optics.” New York: McGraw-Hill, pp 15-28. 31. Gabel VP, Birngruber R, Hillenkamp F. Visible and near infrared light absorption in pigment epithelium and choroid. In: Shimizu K, ed. “International Congress Series No. 450, Twenty-third Concilium Ophthalmologicum, Kyoto. Amsterdam: Excerpta MedicaiElsevier, 1978, pp 650-662. 32. Fine BS, Yanoff M. “Ocular Histology.” New York: Harper & Row, 1979, pp 186-193. 33. Maurice DM. The cornea and sclera. In: Davson H, ed. “The Eye,” Vol. 1. New York: Academic Press, 1962, pp 289-368. 34. Prince MR, Deutsch TF, Methews-Roth MM, Margolis R, Parrish JA, Oseroff AR. Preferential absorption in atheromas in vitro; implications for laser angioplasty. J Clin Invest 1986; 78295-302. 35. Anderson RR, Parrish JA. Optical properties of human skin. In: Regan JD, Parrish JA, eds. “The Science of Photomedicine.” New York: Plenum Press, 1982, pp 147-194. 36. Nuffer R. Transsklerale Streuungs- und Absorptionsmessungen. Diploma thesis, Polytechnical of Munich, Munich, 1988.
340
Vogel et al.
37. Benedek GB. Theory of transparency of the eye. Appl Opt 1971; 10:459-465. 38. Miller D, Benedek GB. “Intraocular Light Scattering.” Springfield, Illinois: Charles C. Thomas, 1973. 39. Kerker M. “Scattering of Light and Other Electromagnetic Radiation.” New York: Academic Press, 1969. 40. KO11 R. Experimentelle Untersuchungen zur Sklerostomie mittels Nd:YAG Laser. Dissertation thesis, Ludwig-Nlaximilians-University, Munich, 1990. 41. Schubert HD. Effects of exposure time in cw Nd:YAG contact transscleral photocoagulation and photofiltration. Lasers Light Ophthalmol 1990; 3:53-59. 42. Smith RS, Stein MN. Ocular hazards of transscleral laser radiation. 11. Intraocular injury produced by ruby and neodymium lasers. Am J Ophthalmol 1969; 67:lOO-110. 43. Schubert HD, Federman JL. A comparison of cw Nd:YAG
44.
45.
46.
47.
contact transscleral cyclophotocoagulation with cyclocryopexy. Invest Ophthalmol Vis Sci 1989; 305364542. Cantor LB, Nichols DA, Katz LJ, Moster MR, Poryzees E, Shields JA, Spaeth GL. Neodymium YAG transscleral cyclophotocoagulation. The role of pigmentation. Invest Ophthalmol Vis Sci 1989; 30:1834-1837. Stolzenburg S, Kress S, Muller-Stolzenburg N. Thermal side reactions during in vitro contact cyclophotocoagulation with the continuous wave Nd:YAG laser. Ophthalm Surg 1990; 21:356-358. Rol P, Fankhauser F, Kwasniewska S, Niederer P. A comparison of ophthalmic fiber optic microfocusing systems. Lasers Light Ophthalmol 1988, 2:115-124. Royston D, Waynant R, Banks A, Ramee S, White CJ. Optical properties of fiber optic surgical tips. Appl Opt 1989; 28:799-803.
Optical Properties of Human Sclera, and Their Consequences for Transscleral Laser Applications Alfred Vogel,
PhD,
Christian Dlugos, BS, Roland Nuffer, BS, and Reginald Birngruber, PhD
H. Wacker Laboratory of Medical Laser Applications, University Eye Hospital Munich, 8000 Munich 2, Germany (A. V., C.D., R.N., R.6.); Wellman Laboratory of Photomedicine, Massachusetts-General Hospital, Boston, Massachusetts 021 14 (R.B.)
The spectral dependence of the optical properties of human sclera adjacent to the limbus was investigated and related to the potentials of transscleral photocoagulation. The total transmission, absorption, and reflection, as well as the angular distribution of the transmitted and reflected light were measured at five laser wavelengths (442 nm, 514 nm, 633 nm, 804 nm, and 1,064 nm), both for noncontact and contact applications. Absorption and scattering coefficients were determined using the KubelkaMunk model for light propagation through a scattering tissue. The scleral transmission is only 6%at 442 nm but increases to 35% at 804 nm and to 53%at 1,064 nm. The absorption is high at short wavelengths with 40%at 442 nm but it is only 6%at 804 nm and 1,064 nm. The reflection is generally higher than 40% and shows little wavelength dependence. The transmitted light is scattered diffusely at short wavelengths, but at 804 nm and 1,064 nm it exhibits a fairly narrow angular distribution in forward direction. Fiber contact leads to an increase of transmission, with a factor of 3.5 at 442 nm, of 2.0 at 804 nm, and 1.5 at 1,064 nm. Our results indicate that the diode laser (804 nm) and the NdYAG laser (1,064 nm) with contact delivery are best suited for transscleral photocoagulation. words: cyclodestruction, scleral absorption, scleral scattering, scleral transmission, transscleral photocoagulation
INTRODUCTION
Transscleral laser radiation has been applied for cyclodestructive procedures in glaucomatreatment [l-151, and for transscleral photocoagulation of the retina [4,6,18,191.Cyclodestructive procedures have been attempted with various wavelengths using ruby lasers [ll, argon lasers [4],Nd:YAG lasers [2,3,5-161, and diode lasers [17,18], whereas transscleral coagulation of the retina has been performed with argon lasers [41, Nd:YAG lasers [6], and diode lasers [18,191. Thereby, either noncontact delivery using a slit lamp or contact delivery through a n optical fiber with a flat or focusing fiber tip has been employed. Recently, special contact lenses have been developed to combine slit lamp delivery and con0 1991 Wiley-Liss, Inc.
tact application [20,211. In this study, the optical properties of the human sclera are investigated in order to provide criteria for selecting the optimal laser wavelength and delivery system for transscleral photocoagulation. The rationale of transscleral laser applications is to affect the target tissue while causing as little damage to the sclera as possible. This requires high light transmission through the sclera and low absorption within the sclera. The direc-
Accepted for publication April 17, 1991. Address reprint requests to Dr. Alfred Vogel, H. Wacker Laboratory of Medical Laser Applications, University Eye Hospital Munich, Mathildenstr. 8, 8000 Munich 2, Germany.
Vogel et al. tional distribution of the transmitted light is also of interest, because it determines the energy density at the target tissue. If the scattering 442 nm within the sclera is very strong, a focusing of the light incident at the sclera will hardly increase necd the energy concentration at the target tissue. Therefore, the total transmission, absorption, and reflection of the human sclera were measured at five laser wavelengths (442 nm, 514 nm, 633 nm, 804 nm, and 1,064 nm), and the angular distribution of the transmitted and reflected light was determined. All investigations were performed both for contact and for noncontact laser applications. 332
t
U
Voltmeter
n 442
nm
MATERIAL AND METHODS
All measurements were performed in vitro with sclera specimens obtained from fresh human autopsy eyes which had been used for keratoplasty. The sclera samples were taken from a site adjacent to the limbus and had a size of about 15 x 15 mm2. Conjunctiva, retina, and choroid were carefully removed. Thereby, the pigmented lamina fusca remained attached t o the sclera. The specimens were stored in physiological saline solution and, during the measurements, they were kept moist with saline solution. Fifteen specimens were used for noncontact measurements, and seven specimens for contact measurements. Five lasers with different wavelengths were used for the investigations: a helium-cadmium laser (RCA Electro Optics, model LD 2186) emitting at 442 nm, an argon ion laser (Spectra Physics, model 168) emitting at 514 nm, a heliumneon laser (Coherent Radiation, model 80) emitting at 633 nm, a diode laser (Zeiss, laboratory device) emitting at 804 nm, and a cw Neodymium:YAG laser (MBB Medilas 11) emitting at 1,064 nm. In the noncontact measurements, the sclera sample was illuminated with the unexpanded laser beam in all cases except that of the diode laser. The strongly divergent diode laser radiation was first coupled into an optical fiber, and the light emitted from the fiber tip was then collimated with a lens and focused onto the sample. The laser power was always limited t o 5 mW in order t o avoid changes of the optical properties of the sclera during the measurements. For the contact measurements, the laser beams were coupled into an optical fiber having a flat fiber tip with 600 p m diameter. In all contact measurements, the laser power at the distal fiber tip was 1.5 mW.
.,
He-
cd
\
\ ....I
I
b)
U
\
Photodiode
Voltmeter
Fig. 1. Experimental arrangement for the measurement of total transmission, absorption, and reflection of the sclera (a), and for the measurement of the angular distribution of the transmitted and reflected light (b).
Measurement of Transmission, Absorption, and Reflection
The total transmission, absorption, and reflection of the sclera samples were determined using a custom made integrating sphere (Fig. la). For absorption measurements, the sample has to be placed at the center of the sphere without impairing the development of a homogeneous light distribution within the integrating sphere. Therefore, the diameter of the sphere was 0.64 m, this size being large enough t o render direct absorption measurements possible. The inner surface of the sphere had a coating of barium sulfate (BaSO,). The reflectivity of the coating was determined for each laser wavelength by comparison with the known reflectivity of a BaS0,-standard provided by Zeiss. The reflection coefficients ranged from 90%at 442 nm t o 92.5%at 1064 nm. All light intensity measurements were performed with a silicium photodiode (TS Optoelectronic, model SD 444), the signal of which was amplified
Optical Properties of Human Sclera
(TS Optoelectronic, model SDC) and displayed with a digital voltmeter (Schlumberger, model A 203). For the transmission measurements, the sclera specimen was placed at position P I in the wall of the integrating sphere as shown in Figure la, and the voltage UT at the photodiode was measured. The specimen was then replaced by a standard target with very high reflectivity, and the reference voltage UR for the empty sphere was determined. The total transmission of the specimen is given by
333
sity observed in the direction given by the angle a.It is assumed in equations (2a) and (2b) that the spatial distribution of the scattered light is radially symmetric around the axis of the laser beam. To calculate from the measurement data, one must consider that the radiant intensity is only determined at discrete anlges ai in steps of 10". Therefore, equations (2a) and (2b) must be replaced by the approximation
+
9
Q~
=
271C I ( ~Sinai J ~
c t
(34
i=l T T
UT
T=-p UR
(1)
19
Q~
whereby p denotes the reflection coefficient of the standard target. The total absorption A and reflection R of the sclera specimens were measured in a corresponding manner, with the specimens being placed at positions Pa and P,, respectively. Goniometric Measurements
The angular distribution of the transmitted and reflected light in the far field was determined with the setup shown in Figure lb. The laser beam was incident normally onto the sclera sample. The photodiode could be moved around the sclera sample at a constant distance of 115 mm. The diameter of the sensitive area of the diode was 10 mm, resulting in an angular resolution of 5". Measurement values were recorded every 10". The data concerning the angular distribution of the radiant intensity were also used t o calculate the total radiant flux +T and +R of the transmitted and reflected light. Theoretically, the ratio +T/+R derived from the goniometric measurements equals the ratio TIR obtained by the measurements with the integrating sphere. A comparison of the respective values allows for a cross-check of the validity of both measurements. The total radiant flux of the transmitted and reflected light is related t o the angular light distribution by
=
2-rr
C ~ ( qSinai )
(3b)
i=ll
where al = 0", cta = = d18 rad = 10".
lo", . . . . aI9= 180",and ha
Determination of Absorption and Scattering Coefficients
The absorption and scattering coefficients of the sclera were calculated using the KubelkaMunk model for light propagation in turbid and absorbing media [22,231. The Kubelka-Munk theory describes the propagation of a uniform, diffuse irradiance through an isotropically scattering tissue layer with no reflections at the boundaries. These assumptions, especially that of diffuse illumination of the sample, are not entirely met in our experiments. Nevertheless, the Kubelka-Munk model is a much better approximation for laser light propagation through the sclera than Beer's law analysis, which ignores the effect of scattering. Its advantage as compared t o more complicated models [231 is that the scattering coefficient S and the absorption coefficient K can be directly expressed in terms of the measured reflection R, transmission T, and thickness d of the sample:
711 2 r
J
K
0
=
(x-1)S,
n
cDR
=
27-r
I
I(a)sin a da
(2b) where
7112
Here a = 0 denotes the forward direction of the incident laser beam, and I(a)is the radiant inten-
x =
2R
(4b)
Vogel et al.
334 RefIection 60
50
-
40
-
30 -
20
-
10
I
I
400
500
I
I
600
700
800
900
lOOO[nm]llOO
Wavelength
!
1.0 400
500
600
700
800
900
1OOO[nrn]llOO
Wavelength
Fig. 2. Total reflection, absorption, and transmission of human sclera, plotted as a function of the laser wavelength. The error bars represent the standard deviation of the measurement values.
Fig. 3 . Ratio of the scleral transmission values with and without fiber contact, plotted as a function of laser wavelength. The ratio was determined for gentle and firm pressure of the fiber tip against the scleral surface.
R ESULTS Transmission, Absorption, and Reflection
cannot be enhanced by more than a factor of 1.5. This is because the transmission value surpasses 50% even without fiber contact (Fig. 2).
Figure 2 shows the results of the noncontact measurements with the integrating sphere. The values for total transmission, absorption, and reflection of the sclera are plotted as a function of the laser wavelength. In the visible range of the spectrum, the largest part of the incident light is reflected. The small wavelength dependence of the reflection corresponds t o the white appearance of the sclera. The transmission increases continuously with increasing wavelength, until above 1,000 nm it becomes larger than the reflection. At 1,064 nm, the transmission is about 53%. At wavelengths below 600 nm, more light is absorbed within the sclera than transmitted through the sclera. The absorption shows, however, a continuous decrease with increasing wavelength; at 804 nm and 1,064 nm, it amounts to only 6%. The scleral transmission increases at all wavelengths, when the laser light is delivered through an optical fiber, with the fiber tip in contact with the tissue. Figure 3 shows the ratio of the transmission values with fiber contact and noncontact delivery as a function of laser wavelength. The ratio was determined both for “gentle” and for “firm” pressure of the fiber tip against the sclera. As already observed by Rol et al. [241, the scleral transmission increases with the pressure applied on the sclera. The improvement of transmission is wavelength dependent and largest for short wavelengths. At 1,064 nm, even with firm pressure, the total transmission
Angular Light Distribution
The results of the goniometric measurements are exemplified in the polar diagrams shown in Figures 4 and 5. The examples are chosen such that the corresponding calculated ratio +T/+R of the transmitted and reflected light flux resembles as closely as possible the mean value determined for the respective wavelength. The size of the diagrams is normalized with respect t o the radiant flux incident onto the sclera. This means that in each diagram an equal distance from the point of incidence represents the same angular transmission or reflection. The noncontact measurements (Fig. 4) demonstrate that, in the visible range of the spectrum, both the transmitted and reflected light are scattered diffusely. With increasing wavelength, however, the scatter profile of the transmitted light exhibits a more and more elongated form. Thus, near infrared radiation is not only better transmitted through the sclera than visible radiation (Fig. 21, but the transmitted light is also less scattered in sideward direction. The scatter profiles for fiber contact with gentle pressure are presented in Figure 5. A comparison with Figure 4 shows that the scleral transmission is considerably higher with fiber contact than without, especially for short wavelengths. This result confirms the observations made with the integrating sphere (Fig. 3). As in
Optical Properties of Human Sclera
335 442 nrn
442 nm
514 nm 514 nm
n 633 nm
*
633 nrn
Fig. 4. Typical examples of polar diagrams for the angular distribution of radiant intensity at different laser wavelengths in the noncontact mode. The size of the diagrams is normalized with respect to the incident radiant flux.
Fig. 5 . Typical angular distribution of radiant intensity at different laser wavelengths in the contact mode with gentle pressure. The size of the diagrams is normalized with respect to the incident radiant flux. The scale is as in Figure 4.
the noncontact mode, the scatter profiles become more elongated for longer wavelengths. The scatter profiles at 804 nm and at 1,064 nm appear to be a little broader with fiber contact than in the noncontact case, when the laser beam is directed perpendicular onto the sclera sample. The broadening of the forward scattering may partly be attributable to the divergence of the light emitted from the flat fiber tip. The tissue effect of the laser light depends not only on the total transmission of the sclera, but, also on the directional distribution of the transmitted light. Only that part of the laser energy scattered into the central solid angle will contribute t o the desired effect at the application site. The directional distribution of the laser energy or laser light flux, respectively, cannot directly be read from Figures 4 and 5 containing polar diagrams of the radiant intensity in one plane only. The solid angle in which a certain radiant intensity is detected can be taken into account using equation (3). Taking the diagram shown in Figure 4 for a wavelength of 1,064 nm as an example, it is found that only 40% of the transmitted light is scattered into the central solid angle of 25" around the laser beam axis. This percentage is even smaller at shorter wavelengths,
exhibiting a less elongated scatter diagram: 35% at 804 nm, and 21% at 514 nm. Contact application leads to an increase of the part of the transmitted light scattered into the cone angle of 25". At 1,064 nm, the percentage was found to change from 40% t o 49%, when fiber contact with gentle pressure was applied. In Figure 6, the ratio +T/+R of total transmitted and reflected radiant flux calculated from the goniometric measurements is compared with the ratio T/R obtained with the integrating sphere. Theoretically, both values should be the same for each laser wavelength. The agreement is indeed very good; the differences observed are smaller than the standard deviations for both measurement series. Absorption and Scattering Coefficients
The absorption and scattering coefficients of the sclera were calculated with equations (4a)(4c), assuming a thickness of the sclera adjacent to the limbus of 0.8 mm [251. Figure 7 shows the mean values of the absorption coefficient obtained from the noncontact measurements, as well as the scattering coefficients determined in the noncontact mode and with fiber contact. The values of the absorption coefficient with fiber contact are
Vogel et al.
336 1.5
80
-
I
Pol 70 -
.-
60
$\ Reflection
-
a:
I- 1.0 -
50 -
1
40 -
ek 0.5 --
30 -
?a
20 10 -
I
400
I
I
I
I
I
500
600
700
800
900
i
I
1000[nm]l100
Wavelength
-
\ \
\. \
\ 30
\ \
Y
20
--.
'. .-.
. '. .
- - _ .- -.~.. - - _ _- ---.._ . .......... a - _ _
\
.........8 . - ..........................................
500
600
500
600
700
I
800
900
, 1000[nm]1100
Fig. 8. Comparison of the results of the present study with present study, human sclera the results of other studies. -, adjacent to the limbus, n = 15; ---,Smith and Slein 1261, human sclera adjacent to the limbus, n = 16;- -, Manson et al. 1271, origin and number of the samples not identified. -, Rol et al. 1241, human sclera with a thickness of 0.71 m m , n = 1.
DISCUSSION Interpretation of the Measurement Results
\
J.__
400
I
s
\
v)
10
I
1,064 nm. Scattering is reduced by fiber contact, especially when firm pressure is applied t o the sclera.
- \
40
__.
I
Wavelength
Fig. 6. Comparison of the measurement results obtained with the goniometer and the integrating sphere. is the ratio of the total transmitted and reflected radiant flux calculated from the data of the goniometric measurements using equation (3). T/R is the ratio of the total transmission and reflection determined with the integrating sphere. Theoretically, both values should be equal for each laser wavelength. The error bars represent the standard deviation of the measurement values. [cm-11
:ion
700
800
900
1000 [nm] 1100
Wavelength Fig. 7. Absorption coefficient K and scattering coefficient S of human sclera. K and S were calculated from the measured values of R and T using equations (4a)-(4c). K is given for the noncontact measurements, and S is plotted for the noncontact mode as well as for fiber contact with gentle and firm prescontact with gentle pressure; sure: --, noncontact mode; ---, . . . , contact with firm pressure.
not given, since they are similar to those without contact. At all wavelengths, scattering within the sclera is much stronger than absorption. Both scattering and absorption decrease considerably with increasing wavelengths; the value of the absorption coefficient drops from 8.6 cm-' at 442 nm t o about 0.6 cm-' at 804 nm, and 0.7 cm-I at
Figure 8 compares the measurement results of this study with the data for the spectral dependence of total scleral reflection and transmission reported by other groups [24,26,27]. The shape of the curves is similar in all cases, and the differences of the absolute values can be partly explained by differences in the preparation of the sclera specimen and by variations of the scleral thickness. It should be noted, however, that in the earlier studies only reflection and transmission have been measured. In two of the studies [24,261, the sum of the values for transmission and ref lection sometimes exceeds 100%.This indicates measurement inaccuracies, since the absorption also has t o be considered. In the present study, the measurement accuracy was increased by using a large integrating sphere which allowed direct absorption measurements. The deviation of the sum (R + T + A) of the mean values from 100% was smaller than 1.0% for all wavelengths. For interpreting the absorption and scattering properties of the sclera, it is helpful t o recall its composition and structure. The sclera consists mainly of water (65%), collagen (75% of dry
Optical Properties of Human Sclera 337 weight), noncollageneous proteins, and muco- portional to K4.After single scattering events polysaccharides [251. Up t o 1% are lipids (de- with d = A, the distribution of the scattered light pending on age) [251, and about 0.5% inorganic is predominantly forward-directed [391. When substances [28]. It is remarkable that the scleral multiple scattering occurs, the spatial distribuabsorption reaches up t o 40% in the visible range tion of the radiation as it passes through the samof the spectrum (Fig. 21, although the absorption ple quickly becomes isotropic (i.e., diffuse), reof collagen and other proteins is almost negligible gardless of the spatial distribution obtained for a above wavelengths of 400 nm [29]. Water absorp- single scattering. With this background, the scattion is also negligible in the visible part of the ter diagrams in Figures 4 and 5 can be inspectrum and plays a role only in the near-in- terpreted. At short wavelengths, a diffuse disfrared region with a first weak absorption peak at tribution of the transmitted light is observed 975 nm [301. One possible explanation for the indicating multiple scattering. With increasing measurement results is indicated by the fact that wavelength, the scattering becomes weaker as exthe wavelength dependence of the scleral absorp- pected theoretically, and the scatter diagrams extion (Figs. 2, 7) resembles that of melanin [311. hibit a more elongated form. The decrease of scleral transmission at This may be due t o the pigmented cells found within the sclera, particularly on the inner sur- shorter wavelengths (Fig. 2) is partly due t o the face (lamina fusca) [32,33]. Another chromophore increase of absorption but is mainly a conseabsorbing in the blue-green part of the spectrum quence of the increased scattering. Again a comis p-carotene [341. It is highly lipophilic and may parison with the optical properties of porcine be bound to some extent to the lipids contained sclera adds to the understanding of the data within the sclera. This explanation is indirectly obtained for humans. The light transmission supported by comparing the absorption character- through porcine slcera is considerably lower than istics of human sclera with that of porcine sclera. through human sclera (only 31% at 1,064 nm as Unlike humans, most other mammals retain rel- compared t o 53%in humans), and the distribution atively little p-carotene and have white fat as op- of the transmitted light is more diffuse [36]. This posed t o the yellow fat found in humans [351. can only be due t o increased scattering in the porMoreover, domestic pigs are hardly pigmented, cine sclera, since its absorption is smaller than and their sclera probably contains very little mel- that of human sclera. The assumption that the anin. The lack of both chormophores corresponds reduced transmission is caused by increased scatt o the small absorption of 6% or less observed in tering is consistent with the greater thickness of porcine sclera over the whole range of the visible the porcine sclera (1.1 mm in the limbal region spectrum [361. It should be noted that although [40] as compared with 0.7-0.8 mm in humans the scleral absorption in human sclera reaches up [25,401) and with the stronger irregularity in the t o 40% at short wavelengths, its absorption coef- order of stromal fibrils and lamellae [401. The light reflection of the sclera is basically ficient (Fig. 7) amounts to only about 1/300 of the diffuse (Fig. 4); i.e., it relies mainly on the scatvalue for the retinal pigment epithelium. The light scattering within the sclera is tering within the tissue. The regular reflectance mainly due t o the refractive index difference be- of the scleral surface amounts to only 2-4% for tween the inhomogeneously distributed collagen normal or nearly normal light incidence, and fibrils (n = 1.47) and the ground substance sur- therefore plays a minor role. Neglecting the rounding the fibrils (n = 1.36) [331. The diameter scleral absorption, the reflection should therefore of the fibrils varies between 30 nm and 300 nm, be the higher, the stronger the scattering within with an average of 60 nm [25,33]. This gives rise the sclera; i.e., it should increase with decreasing to variations of the refractive index of the sclera wavelength. This tendency is indeed observed beon a scale comparable t o the wavelength of light, tween 1,064 nm and 633 nm (Fig. 21, but in the hence t o scattering [37,381. Since the scattering short wavelength range it is covered by the instructures have a size d in the order of the wave- crease of absorption within the sclera. In porcine length of light, they cause neither pure Rayleigh sclera with its little absorption over the whole scattering (d > A) [39]. In the intermediate domain, scat- tinuously with decreasing wavelength [36]. tering varies inversely with wavelength, as demThe increase of transmission due t o fiber onstrated in Figure 7, but not as strongly as contact (Figs. 3-5) may be attributed to the disRayleigh scattering, where the variation is pro- placement of ground substance caused by the
338
Vogel et al.
pressure of the fiber tip. This causes a thinning of the sclera and a reduction of the distance between the collagen fibrils leading to modifications in the interference of the light scattered from adjacent fibrils [37]. If more water than proteins and mucopolysaccharides are displaced, the concentration of the remaining ground substance increases and its refractive index becomes closer t o that of the collagen fibrils [331. As a consequence of all effects described, the scattering is reduced (Fig. 7) so that the transmission increases and the angular distribution of the transmitted light becomes narrower. This effect is stronger when higher pressure is applied to the sclera. Clinical Consequences
nm and 77% at 1064 nm. It can be concluded from these data that the Nd:YAG laser should require two times more energy for transscleral photocoagulation in the noncontact mode than the diode laser. In the contact mode with strong pressure onto the sclera, the required energy should be 2.8 times higher for the Nd:YAG laser. When a fiber optic contact delivery system is used, the total transmission of the sclera is increased by about 50% at 1,064 nm and by 100% at 804 nm, (Fig. 3). An even higher increase of transmission is observed in the visible range of the spectrum, but this is not really significant for clinical applications because of the high scleral absorption at these wavelengths. Moreover, fiber contact leads to an increase of the percentage of light transmitted in the forward direction. Thus, contact application causes a reduction of the energy required for transscleral photocoagulation. A considerable reduction can, however, only be achieved if firm pressure is applied t o the sclera (Fig. 3) [241. Using a contact glass with a large contact area, i.e., with small pressure onto the sclera, the same energy was found t o be necessary for cyclophotocoagulation as in the noncontact mode [201. A t long wavelengths, when the transmitted light is mainly scattered in forward direction (Figs. 4, 5), the energy requirements may be further reduced by focusing the laser beam [45]. However, the focusing power of fiber tips is limited by the small difference in refractive index between glass or sapphire tips and sclera [46,47]. Regarding the transmission and absorption qualities of the sclera, the results of our study support the use of a contact delivery system for transscleral laser applications. The wavelengths best suited are in the near-infrared region, like those of the diode laser (804 nm) or of the Nd:YAG laser (1,064 nm). The use of the diode laser is thereby still restricted by limitations of the maximal laser power available. This will change, however, in the near future.
The aim in transscleral photocoagulation is t o achieve focal cyclodestruction or retinal coagulation without damaging the sclera. This requires the use of a laser wavelength exhibiting little scleral absorption, high transmission through the sclera, and a narrow angular distribution of the transmitted light. In addition, the absorption in the target tissue should be as high as possible. Our results show that the light of Nd: YAG lasers (1,064 nm) and of diode lasers (804 nm) is well transmitted through the sclera and hardly absorbed in scleral tissue. This corresponds to the observation made in various histological studies that after Nd:YAG and diode laser cyclocoagulation at moderate energy levels, the tissue effects are confined t o the ciliary body [2,8,10,12,17-19,41,42]. Only at high-energy levels (5 J in a rabbit eye) far above the coagulation threshold and at long exposure times (300 ms t o 2 s), was scleral damage found, obviously having been caused by heat conduction from the ciliary body into the sclera [8,411. Comparison of the laser effects in pigmented rabbits and albino rabbits has shown that cyclocoagulation relies on the light absorption in the ciliary pigment epithelium [1,43,441. The most important absorbing structure for retinal photocoagulation is the retinal pigment epithelium [311. In both cases, the absorbing chromophore is ACKNOWLEDGMENTS melanin. Since the melanin absorption is about We appreciate stimulating discussions with three times higher at 804 nm than at 1,064 nm F. Fankhauser and P. Rol. [311,the light of the diode laser is better absorbed within the ciliary body and the retina than Nd: YAG laser light. The total transmission of the REFERENCES at 804 nm (35%) than at is 1, Beckman H, Kinoshita A, Rota AN, Sugar HS, Transnm (53%), but this difference diminishes with fiscleral ruby laser irradiation of the ciliary body in the ber contact, if strong pressure is applied to the treatment of intractable glaucoma. Trans Am Acad Ophthalmol Otolaryngol 1972; 46:423-436. sclera. In this case, the transmission is 71% at 804
Optical Properties of Human Sclera 2. Wilenski ZT, Welch D, Mirolovich M. Transscleral cyclocoagulation using a Neodymium:YAG laser. Ophthalm Surg 1985; 16:95-98. 3. Fankhauser F, van der Zypen E, Kwasniewska S, Rol P, England C. Transscleral cyclophotocoagulation using a Neodymium:YAG laser. Ophthalm Surg 1986; 17:94100. 4. Federman IL, Ando F, Schubert HD, Eagle RC. Contact laser for transscleral photocoagulation. Ophthalm Surg 1987; 18:183-184. 5 . Devenyi RG, Trope GE, Hunter WH, Badeeb 0. Neodymium:YAG transscleral cyclocoagulation in human eyes. Ophthalmology 1987; 94:1519-1522. 6. Kwasniewska S, Fankhauser F, van der Zypen E, Rol P, Henchoz PD, England C. Acute effects following transscleral contact irradiation of the ciliary body and the retinal choroid with the cw Nd:YAG laser. Lasers Light Ophthalmol 1988; 225-34. 7. England C, van der Zypen E, Fankhauser F, Kwasniewska S. A comparison of optical methods used for transscleral cyclophotocoagulation in rabbit eyes produced with the Nd:YAG laser: A morphological, physical and clinical analysis. Lasers Light Ophthalmol 1988; 2: 87-102. 8. Heidenkummer HP, Birngruber R, Lorenz B, Gabel VP, Gao L, Tang SH. Morphologische Befunde des Kaninchenziliarkorpers nach transskleraler cw Nd:YAG Laser Koagulation. In: Wollensack J, ed. “Laser in der Ophthalmologie.” Stuttgart: Enke, 1988; pp 184-191. 9. Mehdorn E, Lucke K, Steinmetz M, Messmer E. Transsklerale Ziliarkorperkoagulation mit dem Nd:YAG cw Laser bei Sekundarglaukom. Fortschr Ophthalmol 1988; 86:102-106. 10. Hampton C, Shields MB. Transscleral neodymium:YAG cyclophotocoagulation. A histological study of human autopsy eyes. Arch Ophthalmol 1988; 106:1121-1123. 11. Klapper RM, Wandel T, Donnefeld E, Perry HD. Transscleral neodymium:YAG thermal cyclophotocoagulation in refractory glaucoma; a preliminary report. Ophthalmology 1988; 95:719-722. 12. Latina MA, Pate1 S, de Kater AW, Goode S, Nishioka NS, Puliafito CA. Transscleral cyclophotocoagulation using a contact laser probe: A histologic and clinical study in rabbits. Laser Surg Med 1989; 9:465-470. 13. Brancato R, Leoni G, Trabucchi G, Pietroni C. Contact transscleral cyclophotocoagulation with cw Nd:YAG laser in uncontrolled glaucoma. Ophthalm Surg 1989; 20: 547-551. 14. Trope GE, Ma S. Mid-term effects of Nd:YAG transscleral cyclophotocoagulation in glaucoma. Ophthalmology 1990; 97:73-75. 15. Schuman JS, Puliafito CA, Allingham RR, Belcher CD, Bellows AR, Latina MA, Shingleton BJ. Contact transscleral continuous wave Neodymium:YAG laser cyclophotocoagulation. Ophthalmology 1990; 97571-580. 16. Schuman JS, Puliafito CA. Laser cyclophotocoagulation. Int Ophthalmol Clin 1990; 3O:lll-119. 17. Schuman JS, Jacobson JJ, Puliafito CA, Noecker R J , Reidy WT. Experimental use of semiconductor diode laser in contact transscleral cyclophotocoagulation in rabbits. Arch Ophthalmol 1990; 108:1152-1157. 18. Okamoto S, Takahashi H, Fukado Y, Ozawa T. Laser diode application for transscleral photocoagulation. Laser Light Ophthalmol 1990; 3:29-37.
339
19. Peyman GA, Naguib KS, Gaasterland D. Transscleral application of a semiconductor diode laser. Laser Surg Med 1990; 10:569-575. 20. Simmons RB, Blasini M, Shields M, Erichson PJ. Comparison of transscleral Neodymium:YAG cyclophotocoagulation with and without a contact lens in human antopsy eyes. Am J Ophthalmol 1990; 109:174-179. 21. Durr V, Henchoz PD, Fankhauser F, Kwasniewska S, van der Zypen E. Results and methods of transscleral laser cyclodestruction: A new contact lens for use with non-contact systems. Lasers Light Ophthalmol 1990; 3: 123-131. 22. Kubelka P. New contributions t o the optics of intensely light-scattering materials. Part I. J Opt SOCAm 1948; 381448-457. 23. Cheong WF, Prahl SA, Welch AJ. A review on the optical properties of biological tissues. IEEE J Quantum Electron 1990; QE 26:2166-2185. 24. Rol P, Niederer P, Durr U, Henchoz PD, Fankhauser F. Experimental investigations on the light scattering properties of the human sclera. Lasers Light Ophthalmol 1990; 3:201-221. 25. Torcynski E. Sclera. In: Duane TD, Jaeger EA, eds. “Biomedical Foundations of Ophthalmology,” Vol. I. Philadelphia: J.B. Lippincott Company, 1988, pp 1-23. 26. Smith RS, Stein MN. Ocular hazards of transscleral laser radiation. I. Spectral reflection and transmission of the sclera, choroid and retina. Am J Ophthalmol 1968; 66: 21-31. 27. Manson N, Mc Carthy J, Henry L, Terry C. Transscleral tunable dye laser lesions. In: Marshall J, ed: “Laser Technology in Ophthalmology.” Amsterdam: Kugler & Ghedeni, 1988, pp 163-172. 28. Duke-Elder S. The cornea and sclera. In: “System of Ophthalmology,” Vol. IV: “The Physiology of the Eye and of Vision.” London: Henry Kimpton, 1968, pp 337-363. 29. Hillenkamp F. Interaction between laser radiation and biological systems. In: Hillenkamp F, Pratesi R, Sacchi CA, eds. “Lasers in Biology and Medicine.” New York: Plenum Press, 1980, pp 37-68. 30. Tyler JE. Optical properties of water. In: Driscoll WG, Vaughan W, eds. “Handbook of Optics.” New York: McGraw-Hill, pp 15-28. 31. Gabel VP, Birngruber R, Hillenkamp F. Visible and near infrared light absorption in pigment epithelium and choroid. In: Shimizu K, ed. “International Congress Series No. 450, Twenty-third Concilium Ophthalmologicum, Kyoto. Amsterdam: Excerpta MedicaiElsevier, 1978, pp 650-662. 32. Fine BS, Yanoff M. “Ocular Histology.” New York: Harper & Row, 1979, pp 186-193. 33. Maurice DM. The cornea and sclera. In: Davson H, ed. “The Eye,” Vol. 1. New York: Academic Press, 1962, pp 289-368. 34. Prince MR, Deutsch TF, Methews-Roth MM, Margolis R, Parrish JA, Oseroff AR. Preferential absorption in atheromas in vitro; implications for laser angioplasty. J Clin Invest 1986; 78295-302. 35. Anderson RR, Parrish JA. Optical properties of human skin. In: Regan JD, Parrish JA, eds. “The Science of Photomedicine.” New York: Plenum Press, 1982, pp 147-194. 36. Nuffer R. Transsklerale Streuungs- und Absorptionsmessungen. Diploma thesis, Polytechnical of Munich, Munich, 1988.
340
Vogel et al.
37. Benedek GB. Theory of transparency of the eye. Appl Opt 1971; 10:459-465. 38. Miller D, Benedek GB. “Intraocular Light Scattering.” Springfield, Illinois: Charles C. Thomas, 1973. 39. Kerker M. “Scattering of Light and Other Electromagnetic Radiation.” New York: Academic Press, 1969. 40. KO11 R. Experimentelle Untersuchungen zur Sklerostomie mittels Nd:YAG Laser. Dissertation thesis, Ludwig-Nlaximilians-University, Munich, 1990. 41. Schubert HD. Effects of exposure time in cw Nd:YAG contact transscleral photocoagulation and photofiltration. Lasers Light Ophthalmol 1990; 3:53-59. 42. Smith RS, Stein MN. Ocular hazards of transscleral laser radiation. 11. Intraocular injury produced by ruby and neodymium lasers. Am J Ophthalmol 1969; 67:lOO-110. 43. Schubert HD, Federman JL. A comparison of cw Nd:YAG
44.
45.
46.
47.
contact transscleral cyclophotocoagulation with cyclocryopexy. Invest Ophthalmol Vis Sci 1989; 305364542. Cantor LB, Nichols DA, Katz LJ, Moster MR, Poryzees E, Shields JA, Spaeth GL. Neodymium YAG transscleral cyclophotocoagulation. The role of pigmentation. Invest Ophthalmol Vis Sci 1989; 30:1834-1837. Stolzenburg S, Kress S, Muller-Stolzenburg N. Thermal side reactions during in vitro contact cyclophotocoagulation with the continuous wave Nd:YAG laser. Ophthalm Surg 1990; 21:356-358. Rol P, Fankhauser F, Kwasniewska S, Niederer P. A comparison of ophthalmic fiber optic microfocusing systems. Lasers Light Ophthalmol 1988, 2:115-124. Royston D, Waynant R, Banks A, Ramee S, White CJ. Optical properties of fiber optic surgical tips. Appl Opt 1989; 28:799-803.
