Spatially selective photocoagulation of biological tissues: feasibility study utilizing cryogen spray cooling.

Successful laser treatment of selected dermatoses such as hemangiomas requires thermally induced damage to blood vessels while protecting the epidermis. We present and test a procedure in a rabbit liver tissue model that utilizes cryogen spray cooling during continuous Nd:YAG laser irradiation to induce deep photocoagulation necrosis while protecting superﬁcial tissues from thermal injury. Gross and histologic observations are consistent with calculated thicknesses of protected and photocoagulated tissues and demonstrate the feasibility of inducing spatially selective photocoagulation when cryogen spray cooling is used in conjunction with laser irradiation. This procedure may be useful in the thermal treatment of some pathological conditions for which it is desired that deep photocoagulation be induced while protecting superﬁcial tissues.


Introduction
Various thermally mediated therapeutic procedures utilizing microwave, infrared, or visible electromagnetic radiation have become of interest to many investigators in recent years. [1][2][3] In many of these therapeutic procedures the objective is to induce coagulation necrosis of certain tissue components while protecting superficial tissues from thermal injury. For example, successful laser treatment of dermatoses such as port-wine stain lesions, hemangiomas, and telangiectasias is based on the photocoagulation of blood vessels without inducing thermal injury to the overlaying epidermis and papillary dermis, which could result in skin-surface textural changes or scarring. [4][5][6][7][8] Cryogen spray cooling is a potentially effective method for protecting superficial tissues from thermal injury. 9 By the application of short-duration spurt of a cryogen 1of the order of milliseconds2, rapid and selective cooling of tissue is possible: superficial tissues are cooled while the temperature of deeper tissues remain unchanged. 9 Evaporation of the cryogen on the surface provides the mechanism for rapid removal of heat from the tissue. For example, when tetrafluoroethane 1boiling point < 226°C2 is used as a cryogen, surface-temperature drops of the order of 30-40°C have been obtained within 5-100 ms. [9][10][11] Theoretical results indicate that spray cooling the skin with tetrafluoroethane just prior to pulsed-laser irradiation can selectively cool the epidermis and yet permit photocoagulation of dilated port-wine stain blood vessels. 12,13 Successful blanching of port-wine strain lesions without either epidermal thermal injury or skin-surface textural changes has been reported when the skin is precooled with tetrafluoroethane immediately prior to flash-lamp-pumped pulsed-dyelaser irradiation with a relatively high light dosage 1e.g., 10 J@cm 2 2. 10,11,13 In this paper we present the results of an experiment using continuous Nd:YAG laser irradiation and cryogen spray cooling. The Nd:YAG laser 1l 5 1064 nm2 light is poorly absorbed by biological tissues and penetrates several millimeters into tissues to achieve deep photocoagulation, while the cryogen spray cooling protects against thermal damage to superficial tissues. A theoretical analysis of the tissue thermal response to continuous laser irradiation in the presence of cryogen spray cooling is presented, and clinical implications of this procedure for the treatment of hemangiomas are discussed.

Materials and Methods
Light emitted from a Nd:YAG laser 1LaserSonics Inc., Santa Clara, Calif.2 was used as a heat source to induce deep tissue photocoagulation in freshly excised rabbit liver. Although liver is a relatively homogeneous medium 1unlike skin2, it nevertheless serves as a convenient tissue phantom in which photocoagulation necrosis can be clearly observed.
The experimental apparatus for laser irradiation in conjunction with cryogen spray cooling is shown in Fig. 1. Laser light delivered through a 600-µm core-diameter silica multimode optical fiber was directly incident onto the liver sample surface. Various irradiation parameter sets 1Table 12 were selected to cover a relatively wide range. Irradiation times of the order of seconds were used to achieve a large volume of photocoagulated tissue. Each parameter set was repeated at least three times with little variation in results.
Chlorodifluoromethane 1boiling point of < 240°C2 1Aldrich Chemical Company, Milwaukee, Wisc.2, a nonflammable refrigerant, was sprayed onto the liver through an electronically controlled standard automobile fuel-injection solenoid valve positioned 4 cm from the sample surface at a 30°angle from the tissue normal. The time duration of the cryogen spurt was fixed at 50 ms by a programmable digital delay generator 1Model DG535, Stanford Research Systems, Sunnyvale, Calif.2. The cooled site on the tissue surface was concentric with the laser-irradiated site and ,10 mm in diameter.
Radiometric measurement of the surface temperature at the center of the laser-irradiated site was used to trigger the delivery of cryogen spurts. When the radiometric surface temperature reached a prespecified threshold value 130°C2, a cryogen spurt of 50-ms duration was delivered onto the liver surface. In this way, repetitive, pulsed cryogen spray cooling during continuous laser irradiation was accomplished.
Infrared emission from the liver surface was detected by the use of a 1-mm 2 liquid-N 2 -cooled HgCdTe detector 1Model MDD-10E0-S1, Cincinnati Electronics, Mason, Oh.2, optically filtered at the cold stop by a 10.6-14-µm bandpass filter. Because the infrared absorption coefficient of water in this range is approximately 60 mm 21 1Ref. 142, we expect that contributions to the infrared signal originate predominantly from superficial tissues 1approximately the uppermost 15 µm2.
The detector was placed at the focal plane of a 25-mm-diameter f@1 Ge lens configured for unit magnification. For an improved signal-to-noise ratio, the pupil was stopped to a 5-mm diameter and the infrared radiation was amplitude modulated 13.5 kHz2 by the use of a chopper 1Model SR540, Stanford Research Systems2 and synchronously detected by a lock-in amplifier 1Model SR850, Stanford Research Systems2. The output signal of the lock-in amplifier was used as the threshold trigger for the digital delay generator. The infrared detection system was calibrated for temperature changes above and below the ambient value 123°C2. The lock-in amplifier output voltage was measured as a function of the surface temperature of 112 an aluminum block coated with highly emissive 1e < 0.972 black paint 1TC-303 black, GIE Corporation, Provo, Ut.2 and heated by a resistive element from 23°C to 75°C, and 122 a thermoelectric cooler 1ITI FerroTec, Chelmsford, Mass.2 that was coated with the same black paint and cooled from 23°C to 220°C. The surface temperature of the aluminum block and the thermoelectric cooler was measured by the use of a precision thermistor 1Model 8681, Keithley Instruments, Cleveland, Ohio2. The radiometric temperature was computed from calibration data by use of a polynomial fit.

Results and Discussion
In Fig. 2 we present a cross section of rabbit liver tissue dissected immediately after the procedure along the central axis of laser irradiation 1parameter set 2, t irrad 5 15 s2. Photocoagulated tissue 1<4 mm thick, 5-mm maximum width2 below a protected layer 1<400 µm thick2, between the arrows, is observed. Photocoagulation was accompanied by an observable loss of the natural red color and whitening of the tissue, indicative of optical-property changes, in part caused by thermal denaturation of hemoglobin molecules. In the absence of cryogen spray cooling there was no protected layer, and the region of photocoagulation extended from the surface deep into tissue. Histologic sections of the nonirradiated and irradiated 1parameter set 2, t irrad 5 10 s2-cooled rabbit liver tissue are shown in Figs. 3. The nonirradiated section 3Fig. 31a24 shows the red blood cells appearing as small dark regions within blood vessels and capillaries, which appear as white. A confined zone of photocoagulation is observed in the irradiated and cryogen-spray-cooled section 3Fig. 31b24. Red blood cells are still present in superficial tissue 1indicated by the arrow2, but they are completely absent in the deeper regions because of photocoagulation of the blood vessels and capillaries. The photocoagulated tissue is more compact when compared with the protected superficial and the nonirradiated tissues, an indication that thermal destruction of blood vessels and capillaries has taken place.
Photocoagulation necrosis without damage to the superficial tissues was also achieved with parameter sets 1 and 4. Irradiation times in parameter set 1 190 and 120 s2, however, may be considered too long for clinical use. Protection of the superficial tissues was not achieved with parameter set 3. When the laser beam was focused to a small spot 13-mm diameter2, the resulting irradiance was so large 12830 kW@m 2 2 that the amount of heat generated was greater than that which could be removed by cooling. An example of the recorded radiometric surface temperature in response to laser irradiation 1parameter set 2, t irrad 5 15 s2 and repetitive cryogen spray cooling is shown in Fig. 4. The radiometric surface temperature shows cycles of laser-induced temperature increases to 30°C, the prespecified threshold temperature for the application of cooling, followed by rapid temperature reductions to approximately 25°C after cryogen is sprayed onto the tissue surface. The frequency of cryogen spurts increases toward the end of the irradiation time, indicating that more heat is diffusing toward the surface as the internal temperature of the liver sample increases.
Temperature distributions within the tissue during continuous laser irradiation and repetitive pulsed cooling can be calculated by the solution of the

3316
APPLIED OPTICS @ Vol. 35, No. 19 @ 1 July 1996 heat-conduction equation. In one dimension, where T 1°C2 is the temperature, z 1m2 is the distance into the tissue 1with the origin at the tissue surface2, t1s2 is time, a is the thermal diffusivity 12 3 10 27 m 2 @s for liver tissue 15 2, k is the tissue thermal conductivity 10.59 W@mK2, 15 and Q L 1W@m 3 2 is the volumetric heat generation caused by laser irradiation. We assume that Q L is distributed as where µ a is the absorption coefficient of liver at 1064 nm 10.88 mm 21 2, 16 and f 0 , in watts per meter squared, is the irradiance. Our simple assumption of the exponential decay of Q L with depth does not take into account the scattering of laser light within tissue. [17][18][19][20] Backscattering of laser light is expected to cause the maximum value of Q L to occur just below the surface of the sample. 21 Deep in the tissue, Q L is overestimated by the neglect of scattering. 22 However, as a first approximation, the exponential model provides insight into the thermal response of tissue during irradiation in conjunction with cryogen spray cooling. The effect of blood perfusion on the resulting temperature distributions is not taken into account because this is an in vitro study. However, for in vivo studies and in highly vascularized tissues, contributions of blood perfusion to heat transfer can be significant for heating times, of the order of minutes. 23 We assume a periodic temperature variation 1Fig. 42 and write a thermal boundary condition of the form T1z 5 0, t2 5 T thresh 2 3 T thresh 2 T min 2 11 2 cos 2pft2 4 .

132
Here, T thresh is the prespecified threshold temperature 130°C2, T min is the minimum surface temperature achieved as a result of cooling 125°C2, and f 1Hz2 is the spurt repetition rate.
The solution of Eq. 112 with boundary-condition equation 132 can be expressed as a superposition of the thermal responses caused by laser-induced heating and cryogen spray cooling: where 24  Fig. 4, it can be seen that there is a cryogen spurt released approximately every 1.6 s2 and µ a 5 0.88 mm 21 . From boundary-condition equation 132, it is seen that the surface temperature reaches a minimum when t 5 n@f 6 1@2f 1with n a positive integer2 and a maximum when t 5 n@f. The solid and dashed curves shown in Fig. 5 correspond to the calculated temperature distributions when surface temperatures reach a minimum and a maximum, respectively. For example, when n 5 5, the minimum and maximum surface temperatures occur at 7.2 and 8.0 s, respectively 3Fig. 51a24. With the assumption that the threshold temperature for photocoagulation necrosis is approximately 55°C when this temperature is maintained for seconds 25 3previous measurements of tissue temperature during continuous Nd:YAG laser irradiation have indicated that observable tissue whitening occurs at approximately 55°C 1Ref. 2624, superficial tissue 1<400 µm thick2 remains below the threshold temperature required for photocoagulation 155°C2, whereas deeper tissue is denatured. When n 5 10, the minimum and maximum surface temperatures occur at 15.2 and 16.0 s, respectively 3Fig. 51b24. The thickness of photocoagulated tissue has increased, while the temperature of the superficial tissue still remains below 55°C.
The heat flux q 1W@m 2 2, removed at the surface is obtained by the superposition of fluxes arising from cooling and laser irradiation: The calculated values of the heat flux removed at the tissue surface in response to laser irradiation 15 W, 7-mm irradiated spot diameter2, µ a 5 0.88 mm 21 , and cryogen spray cooling 1 f 5 0.6252 are shown in Fig. 6. The oscillatory behavior results from the nature of boundary-condition equation 132. The minimum and maximum values increase with irradiation time and approach steady-state values.
The calculated thicknesses of the photocoagulated and protected tissues as functions of the irradiation time for three different laser powers 15, 10, and 20 W2 but constant delivered laser energies 180 J2 and irradiated spot diameters 17 mm2 are shown in Fig. 7. Calculations of the thickness of photocoagulated tissue were based on the assumption of 55°C as the threshold temperature for tissue whitening. The physically unrealistic ripples in the curves result from the oscillatory boundary condition. The delay time before tissue photocoagulation begins increases as the laser power is lowered 3Fig. 71a24. The time required to photocoagulate a given tissue thickness is inversely related to the laser power.
The thickness of the protected superficial tissue decreases with the irradiation time 3Fig. 71b24. With a laser power of 5 W delivered for 16 s, an approximately 300-µm thickness of tissue is protected, whereas an almost 4.2-mm thickness of the deeper tissue is photocoagulated. Although we have not accounted for laser-light scattering and changes in optical properties resulting from photocoagulation, the calculated thicknesses of the protected and photocoagulated tissues are consistent with the experimental results and indicate the feasibility of inducing deep photocoagulation while protecting the superficial tissue. Nevertheless, the thickness of the photocoagulated tissue is somewhat overestimated because we have neglected light scattering and the radial diffusion of heat, which would reduce the temperature of deeper tissues. Furthermore, this simple theoretical model cannot successfully be applied to experiments involving high values of f 0 because unrealistically high temperatures and much smaller protected layers than those found experimentally are predicted. A more realistic model should account for the utilization of laser energy in bringing about phase changes associated with thermal dena-

3318
APPLIED OPTICS @ Vol. 35, No. 19 @ 1 July 1996 turation 27 and requires a more accurate value for the tissue absorption coefficient. 28 As illustrated with a liver tissue phantom and by our theoretical analysis, the feasibility of inducing spatially selective coagulation with cryogen spray cooling may have broad clinical implications. For example, successful laser treatment of large or bulky hemangiomas is based on photocoagulating blood vessels while avoiding thermal injury to the epidermis and papillary dermis. Because of the relatively deep tissue penetration at 1064 nm, the Nd:YAG laser has been used to coagulate bulky hemangiomas 29,30 in which the dermis is almost completely replaced by blood vessels. 31 Nevertheless thermal injury to the epidermis and papillary dermis remains a major concern. 5 The repetitive application of cryogen spray cooling during continuous Nd:YAG laser irradiation may provide a method with which the epidermis and papillary dermis can be protected.
Cooling tissue with water or air during microwave heating and continuous laser irradiation has been shown to be effective for protecting superficial tissues. 19,32,33 Cryogen spray cooling can offer compa-rable or greater temperature reductions. This advantage is evident in Fig. 4, from which it can be seen that the time-averaged surface-temperature reduction is ,16°C, which is probably somewhat larger than the surface-temperature drop when water or air cooling is used. 19,32 We are currently investigating the feasibility of inducing spatially selective coagulation using cryogen spray cooling in conjunction with laser irradiation for the treatment of hemangiomas and other selected dermatoses. The results of these studies will be reported in future papers.

Conclusions
We have experimentally and theoretically demonstrated the feasibility of inducing spatiall selective tissue coagulation during thermally mediated procedures by using cryogen spray cooling. Various clinical procedures, such as laser treatment of hemangiomas, could potentially benefit from cryogen-spray cooling in conjunction with heating of tissues in order to confine tissue necrosis to subsurface regions.