A comparative study of human skin thermal response to sapphire contact and cryogen spray cooling

Surface cooling, in conjunction with various thermally mediated therapeutic procedures, can provide a means to protect superficial tissues from injury while achieving destruction of deeper targeted structures. We have investigated the thermal response of in-vivo human skin to: (1) contact cooling with a sapphire window (6-12/spl deg/C); and (2) spray cooling with a freon substitute cryogen [tetrafluoroethane; boiling point /spl ap/-26/spl deg/C at 1 atmospheric pressure (atm)]. Measurements utilizing infrared radiometry show surface temperature reductions from 30/spl deg/C to 14-19/spl deg/C are obtained within approximately is in response to sapphire contact cooling. Surface temperature reductions to values between 5/spl deg/C and -9/spl deg/C are obtained in response to 20-100-ms cryogen spurts. Computational results, based on fitting the measured radiometric surface temperature to estimate heat transfer parameters, show: (1) temperature reductions remain localized to approximately 200 /spl mu/m of superficial tissue; and (2) values of heat flux and total energy removed per unit skin surface area at least doubled when using cryogen spray cooling.


I. INTRODUCTION
T HERMALLY-MEDIATED therapeutic procedures utiliz- ing microwave, infrared, or visible electromagnetic radiation are of considerable interest to many investigators and clinicians.In many of these procedures, the objective is to induce thermal injury to subsurface targeted structures while preserving superficial tissue.For example, successful laser treatment of selected hypervascular cutaneous malformations such as port wine stain (PWS) birthmarks, hemangiomas (benign vascular tumors of infancy), and telangiectasias requires photocoagulation of subsurface targeted blood vessels while protecting the overlying epidermis from thermal injury.
Surface cooling can provide a means to protect superficial skin structures from undesired thermal injury [1]- [5].Heat may be removed from skin by generating a thermal gradient at the surface.We analyze cases where a thermal gradient is generated by: 1) placing a cold sapphire window in contact with skin; and 2) spraying a short (ms) cryogen spurt onto the skin surface; both sapphire contact and cryogen spray cooling (CSC) are currently utilized by clinicians to induce selective photocoagulation of PWS blood vessels without damaging the overlying epidermis [6], [7].Theories incorporating thermal and radiometric considerations to compute skin temperature distributions, and experiments that use infrared radiometry to measure the in-vivo response of human skin to these cooling modalities are presented.

A. Heat Transfer
We assume skin is a semi-infinite medium cooled by sudden contact with a: 1) semi-infinite sapphire window; or 2) thin film ( 10-40 m) consisting of liquid cryogen and ice (formed as a result of water vapor condensation) at the surface (Fig. 1).We assume the lateral dimensions of the cooled surface are much larger than skin thickness ( 1 mm) so the temperature distribution may be computed by solving the one-dimensional heat conduction equation (1) where ( C) is skin temperature in response to cooling (beginning at time zero), (m) is distance into the skin (with origin at the skin surface), (s) is time, and is skin thermal diffusivity (1.1 10 m s [8].Inasmuch as skin is a multilayered structure, an accurate analysis of heat conduction should take into account the thermal properties of various components (i.e., epidermis and dermis).However for simplicity, we assume a homogeneous medium with spatially uniform thermal properties.A Robin boundary condition at the skin surface is implemented (2) where is thermal conductivity (0.45 W m K for skin [8]), ( C) is temperature of the sapphire window or cryogen-ice film, and (W m K ) is the average heat transfer coefficient over the surface due to either: 1) thermal resistance at the sapphire-skin interface or 2) convective/evaporative effects when cryogen is sprayed onto the skin surface.
When cooling is induced by a sapphire window in contact with skin, solution to (1) with boundary condition (2) is [9] (3) where (4a) In the above expressions, is skin initial temperature prior to cooling (assumed 30 C), is the complementary error function , is defined as the thermal effusivity with corresponding to the sapphire window 46 W m K , 15.1 10 m s [10]) or skin, and is the average heat transfer coefficient that accounts for thermal resistance associated with sapphire contact cooling (SCC) at the sapphire-skin interface.

As
, thermal resistance at sapphire-skin interface diminishes and approaches a perfect contact between the two media; conversely, as , resistance becomes infinite.When cooling is induced by spraying cryogen onto the skin surface, a solution that has a similar form to (3) is obtained [9] (5) here, where is the average heat transfer coefficient that accounts for convective/evaporative effects due to CSC.
The instantaneous heat flux, (W m ), and total energy removed per unit area of skin surface, (J m ) are, respectively

B. Infrared Radiometry
Radiometric measurements are dependent on the temperature distribution below the surface and the infrared properties of skin in the detected spectral bandwidth (3-5 m in this study).The change in radiometric temperature ( C), due to cooling is computed by integrating the infrared emission over all depths [5] (8) Here, we have assumed that the presence of an infrared attenuating cryogen-ice film (thickness, ) makes where .

A. Sapphire Contact Cooling
A contact cooling device, consisting of two circular sapphire windows (each 3 mm in thickness and 20 mm in diameter) separated by an evacuated chamber, was constructed.The evacuated chamber was used to prevent formation of an infrared absorbing film of mist (resulting from water vapor condensation) on the contact sapphire window.A recirculating chiller (T-250, ThermoTek Inc.) delivered water (4 C; lowest attainable temperature) to a copper coil wrapped around the contact sapphire window, which was cooled to temperatures between 6 C and 12 C (Fig. 2).Skin of human volunteers was cooled by bringing fingers or arms in sudden contact with the cold sapphire window within the field of the infrared radiometer.

C. Temperature Measurements
Emission in the 3-5 m spectral range was collected with an InSb infrared focal plane array (IR-FPA) camera (Amber Engineering Inc., Goleta CA) configured for unit magnification.The IR-FPA camera system acquired 30-400 infrared emission images per second from skin, and was externally triggered by the digital delay generator.The infrared signal collected by each detector element in the IR-FPA was digitized with a 12-b (0-4095) A/D converter, and stored in a computer.The infrared detection system was calibrated for temperature changes above and below the ambient value (23 C).Average pixel values over a circular region ( 1-mm diameter; 800 pixels) were measured as a function of the surface temperature of: 1) an aluminum block coated with highly emissive black paint (TC-303 black, GIE Corp., Provo, UT) and heated by a resistive element from 23 C to 75 C; and 2) a thermoelectric cooler (ITI FerroTec, Chelmsford, MA) coated with the same black paint, and cooled to 20 C. Simultaneously, a surface mount precision thermistor (8681, Keithley Instruments, Cleveland, OH) measured the temperatures of the aluminum block and thermoelectric cooler.Calibrations were performed with and without room temperature SCC device placed against the aperture.Responsivity (ratio of pixel value to temperature) of the infrared detection system was higher with the SCC device present (Fig. 3).We attribute this effect to background emission specularly reflected from the sapphire window surface, resulting in an additional signal component detected by the IR-FPA.IV. RESULTS

A. Measured and Theoretical Radiometric Temperatures
Examples of measured and predicted radiometric surface temperature profiles in response to SCC are presented in Fig. 4. Measurements represent average temperatures in an approximately 2-mm-diameter circular skin surface region.Theoretical values shown are for 2000 W m K , and 12 C. Considering that water is the major constituent of skin, we used 10-20 mm , characteristic of liquid water in the 3-5 m spectral region [15].
We attribute the discrepancy between measured and theoretical values at early times to: 1) a thin infrared absorbing mist layer that sometimes formed on the contact surface of the cold sapphire window; and 2) difficulty in determining the exact time of contact between the sapphire window and skin surface.The cold (relative to skin surface temperature) mist provides an additional signal component detected by the IR-FPA at early times; as the mist and skin surface reach an equilibrium temperature, good agreement between measured and theoretical values is obtained.
Inasmuch as the sapphire window temperature, and thermal contact at the sapphire-skin interface were variable, steady state surface temperature in response to SCC ranged between 14-19 C. Reasonable agreement between measured and theoretical values was obtained by proper selection of , , and .In general, ranged between 2000-10 000 W m K , indicating variability in thermal resistance at the sapphire-skin interface.Value of ranged between 6-12 C. In contrast to SCC, rapid temperature reductions to values between 5 C and 9 C (average temperatures in approximately a 2-mm-diameter circular skin surface region) were obtained in response to CSC ( 20-100 ms) [Fig.5(a)].After the cryogen spurt was terminated, the radiometric temperature returned to ambient values.Increased spurt duration resulted in a longer time to return to baseline temperature.
The physical model of the cryogen-ice film on skin surface gives good agreement with infrared thermal measurements

B. Sensitivity Analysis
To investigate the influence of various parameters in fitting the experimental data, sensitivity studies were performed.Increasing from 4000 to 40 000 W m K resulted in lower radiometric temperatures for early times, whereas a value as high as 80 000 W m K had a negligible effect [Fig. 6 The rate of cryogen-ice film build up had an appreciable effect on the radiometric temperature change; considerably lower temperatures were achieved as was increased from 0.35 to 1.0 and 1.75 m ms , respectively [Fig.6(d)].

C. Computed Thermal Responses
Computed values of heat flux, and total energy removed per unit skin surface area in response to SCC and CSC, are presented in Fig. 7.We have assumed a "best" case for SCC by using 10 000 W m K corresponding to relatively minimal thermal resistance at the sapphire-skin interface, and 4 C.For CSC, we have assumed 40 000 W m K , and 7 C.When cryogen is sprayed onto the skin surface, the immediate heat flux is approximately four times greater than conductive cooling using a cold sapphire window [Fig.7(a)].As skin surface temperature becomes closer to that of the sapphire window or the cryogen-ice film, the thermal gradient and, subsequently, heat flux diminish over time.However, total energy removed per unit skin surface area continues to increase.For example, due to CSC is approximately four and two times greater than by SCC immediately and 100 ms after cooling, respectively [Fig.7(b)].
The relationship between depth of skin cooling sapphire contact time or cryogen spurt duration is demonstrated with isothermal contour plots (Fig. 8).When a cold sapphire window 4 C) is placed on the skin surface, reducing the temperature at 100 m to 10 C requires approximately 2 s of contact time; deeper tissue structures (up to 500 m) are cooled to temperatures between 10 C and 20 C [Fig.8(a)].
When cryogen is sprayed onto the skin surface, temperature reductions to approximately between 0 C and 20 C are obtained for 10 100 m when the spurt duration is 40 ms [Fig.8(b)].Temperatures of deeper tissue structures (e.g., located in 200 m) are not reduced substantially even when the spurt duration is longer (e.g., 100 ms).In conclusion, with CSC, cooling remains localized in the epidermis while leaving the temperature of deeper structures unaffected.

V. DISCUSSION
Inducing spatially selective photocoagulation in multilayered composite biological tissues is the objective in many thermally mediated therapeutic procedures that utilize microwave, infrared, or visible electromagnetic radiation.For example, successful laser treatment of cutaneous hypervascular malformations such as PWS is based on selective photothermal destruction of targeted blood vessels without thermal damage to the normal overlying epidermis.Selective cooling of superficial tissue is a technique to protect the epidermis from laser induced thermal injury.In this study, we have investigated the effectiveness of a cooled sapphire window, and a short cryogen spurt for selective cooling.
According to our infrared thermal measurements, the contact at the sapphire-skin interface is a determining factor in the resulting temperature profile.For relatively good thermal contact (i.e., minimal thermal resistance at the sapphire-skin interface), temperature reductions to approximately 14-19 C are achieved at the surface.Poor thermal contacts (i.e., large thermal resistance at sapphire-skin interface) results in higher temperatures (16 C) at the surface.
Apart from the difficulty of placing the sapphire window in contact with the skin surface for short times (less than 1 s), the resulting temperature reductions are relatively minimal even for good thermal contacts.Inasmuch as laser induced temperature increases of 30-70 C can be achieved during treatment of PWS [17], SCC may not reduce the temperature sufficiently within the epidermis immediately prior to pulsed irradiation.More importantly, however, sustained cooling does not confine the temperature reduction to the epidermis.Consequently laser energy is used first to "rewarm" the cooled targeted blood vessels, and then induce a sufficient temperature increase required for photocoagulation.
Infrared thermal measurements show surface temperature reductions to between 5 C and 9 C are obtained within 20-100 ms by CSC (Fig. 5).Liquid cryogen droplets (b.p. 26 C) strike the "hot" (30 C) skin surface and undergo evaporation.Temperature reductions are obtained as a result of skin supplying the latent heat of vaporization to cryogen droplets.Since relatively large temperature reductions are obtained within milliseconds, cooling remains localized to the epidermis.
Inasmuch as water vapor present in the surrounding air undergoes condensation upon release of the cryogen into the atmosphere, variable quantities of ice can be formed at skin surface based on the relative humidity level.Results of preliminary experiments performed in an atmospheric water vapor control chamber indicate that lowering the relative humidity level to 5% results in elimination of ice at skin surface [18].
Whether ice elimination is clinically desirable, however, requires further investigation.Nevertheless, reducing or eliminating ice minimizes the scattering of laser light by the cryogen-ice film which is formed over the skin surface.A reduction of 10%-15% in light transmission ( 585 nm; a commonly used wavelength for laser treatment of PWS) through the cryogen-ice film has been reported for tetrafluoroethane spurts ranging between 5-100 ms [19].
Although laser treatment of PWS with irradiation parameters 585 nm, pulse duration 0.5 ms, and dosage 5-10 J cm has yielded successful results in some cases, improved therapeutic outcome in a greater proportion of patients with large blood vessel diameters ( 100 m) may be expected by using alternative wavelengths, longer pulse durations (1-10 ms), or higher light dosages [20]- [25].Unfortunately, laser irradiation utilizing alternative parameters can result in more heat generated within the epidermis.In such a situation, a single cryogen spurt immediately prior to pulsed laser irradiation may not be sufficient to protect the epidermis from thermal injury.Additional cryogen spurts during and possibly after laser irradiation may be required to remove a sufficient amount of heat to protect the epidermis from subsequent hypertrophic scarring, atrophy, and changes in normal skin pigmentation.Further studies aimed at improving treatment of PWS and other cutaneous hypervascular malformations utilizing CSC in conjunction with laser irradiation are currently underway.

VI. CONCLUSION
Measurements utilizing infrared radiometry show contact cooling with a sapphire window reduces the skin surface temperature from 30 to approximately 14-19 C, depending on temperature of the sapphire window, and the thermal at sapphire-skin interface.CSC with tetrafluoroethane results in rapid (within milliseconds), and large (on the order of 30 C) surface temperature reductions.Computational results indicate when a cryogen spurt is sprayed onto the skin surface: 1) the amount of heat flux and total energy removed per unit skin surface area are at least doubled; and 2) spatially selective epidermal temperature reduction is achieved.

Fig. 1 .
Fig. 1.Model geometry consisting of skin in contact at z = 0 with a: (a) semi-infinite sapphire window or (b) cold and infrared-attenuating cryogen-ice film of varying thickness l(t).Various thermal and infrared properties are represented by the quantities within the parentheses.Thermal resistance between sapphire-skin interface is assumed.

Fig. 2 .
Fig. 2. Schematic of instrumentation for measuring the radiometric surface temperature of skin in response to: (a) sapphire contact or (b) cryogen spray cooling.

Fig. 3
Fig. 3 Calibration curves for the infrared detection system: in air ( ), and with intervening SCC device present ( ).
(a)].For , measured values differed considerably from those predicted indicating that the mechanical contact between the sapphire window and skin surface was a key determining factor in the resulting temperature profile.Increasing from 40 to 60 mm influenced the radiometric temperature during the 10-60-ms interval [Fig.6(b)].Changing from 10 to 12 mm had minimal effect on the radiometric temperature [Fig.6(c)], but when increased to 20 mm , lower values resulted during the first 60 ms.