Skin model surface temperatures during single and multiple cryogen spurts used in laser dermatologic surgery

Background: Although cryogen spray cooling (CSC) is used to minimize the risk of epidermal damage during laser dermatologic surgery, concern has been expressed that CSC may induce cryo-injury. In order to address this concern, it is necessary to evaluate the effects of prolonged exposure of human skin phantoms (HSP) to CSC. Objective: To measure the minimum surface temperature ( T min


INTRODUCTION
The importance of cryogen spray cooling (CSC) in conjunction with laser therapy of various dermatoses such as port wine stain (PWS) [1,2], telangiectasias, hemangiomas [3], hair-removal, and rhytides [4] has been widely documented [5].CSC prevents excessive heating of the superficial skin layers, thus allowing the safe use of higher light doses while avoiding complications such as hypertrophic scarring or dyspigmentation [6].
When compared to other epidermal cooling methods, CSC is particularly useful for treatment of superficial targeted lesions, since it permits [7]: (1) accurate control of the cryogen application time (typically 5-100 milliseconds) and, consequently, cooling time; and (2) high heat transfer rates as cryogen is deposited onto the skin and evaporates very quickly therefrom.These two characteristics are instrumental in achieving efficient and spatially selective epidermal cooling.
Despite these advantages, concern has been expressed that CSC may induce cryo-injury [8].In order to address this concern, computational models [9], epoxy phantoms [7,10] and, more recently, human skin tissue culture models (RAFT) [11] have been used to evaluate the effects of prolonged exposure of human skin to CSC.
In this work, we investigate systematically the thermal response of human skin phantoms (HSP) by measuring the minimum surface temperature (T min ) and the time (t Tmin ) at which it occurs, as well as determine the time the sprayed surface remains below 08C (sub-zero time, Dt s ) and À268C (residence time, Dt r ) during the application of single (SCS) and multiple (MCS) cryogen spurts for two initial HSP substrate temperatures T i : 23 and 708C.

Numerical Simulation
We computed the radiant energy distribution resulting from a single laser pulse (beam radius ¼ 5 mm; fluence ¼ 4 J/ cm 2 ) using the well-known Monte Carlo code (MCML/ CONV) developed by Wang et al. [12,13].In order to estimate the total heat extracted from the HSP surface throughout a specific cryogen-laser pattern, we developed a numerical simulation using commercial finite element software (FEMLAB TM , COMSOL, Burlington, MA).The output of the Monte Carlo code is entered as the heat source required by the FEMLAB code [14] to simulate heat diffusion.For simplicity, each cryogen spurt was simulated by imposing a convective boundary condition on the HSP surface with constant h (heat transfer coefficient) and T c (cryogen temperature) values of 7,000 W/m 2 K and À408C, respectively.
We simulated a cryogen-laser pattern similar to those employed in commercially available devices, as represented in Figure 1.The thickness, thermal [9] and optical [15] properties of human skin layers used for calculations of the radiant energy distribution and total heat extracted, respectively, are shown in Table 1.

HSP
The HSP consisted of a thin (90 mm) rectangular (3.42Â3.50mm) silver foil placed on top of an epoxy resin and a type-K thermocouple of $50 mm bead diameter with a response time of 3 milliseconds (Omega, Stamford CN), positioned in between.Thermal paste was applied around the bead to ensure good thermal contact.The foil was sufficiently thin to ensure fast response and a spatially averaged measurement.The purpose of the HSP was to provide thermal properties on the same order of magnitude as those of human skin [9,10].In general, the CSC procedures are used clinically at room temperature, which is the reason for selecting T i ¼ 238C; however, it is believed that 708C is the threshold temperature for instantaneous skin-thermal injury [16].Therefore, and in order to simulate heat generated by the laser pulses, in another series of experiments a copper plate heated (by a pair of thermo-electric coolers: TEC) the HSP to T i of 708C prior to cryogen application.

Cryogen Delivery
The only FDA-approved cryogen compound currently used in laser dermatologic surgery is 1,1,1,2 tetrafluoroethane, also known as R134a, with boiling temperature T b %À26.28C at atmospheric pressure.R-134a is contained at saturation pressure (6.7 bar at 258C) and delivered through a standard high-pressure hose to a control valve.A commercial cryogen spray nozzle (with approximate inner diameter of 0.5 mm) used for laser treatment of vascular lesions and hair removal was employed to spray the cryogen on to the HSP.
The nozzle-to-sprayed surface distance, z, was 31 mm, (similar to that currently used in several commercially available CSC devices) for all experiments.The relative humidity was 39% and the room temperature %238C.A schematic of the experimental setup is shown in Figure 2.
In this study, we employed electronically controlled SCS and MCS spray patterns.T min , t Tmin , Dt s , and Dt r , were systematically measured for 17 spray patterns: (a) 1 SCS of Fig. 1.Schematic showing alternate cryogen spray, delay time, and laser pulses employed in our numerical simulation.

RESULTS AND DISCUSSION
Although the dynamics of a MCS spray pattern of CSC and laser-induced heating is different from that of a SCS of same Dt total followed by one laser pulse, both spray patterns can be compared in terms of the total heat extracted.According to the numerical simulation described above with the optical and thermal parameters described therein, the total heat extracted through the pattern shown in Figure 1 is $10.9 kJ/m 2 , which is similar to that ($9.2 kJ/ m 2 ) from the pre-heated HSP (at 708C) exposed to a SCS of 40 milliseconds.For this reason, the Dt c for all MCS patterns was selected to be 40 milliseconds.
Figure 4 shows a typical HSP surface temperature measurement in response to a MCS pattern as well as the definitions employed to describe the temperature profile as a function of time.As spray droplets impinge on the surface, very fast heat extraction occurs at the cryogen-HSP interface.A rapid decrease in surface temperature (quantitatively similar to that expected to occur on human skin) is noted and continues for some time after spurt termination.A T min is reached at a certain time (t Tmin ), which depends on the Dt total employed in each cryogen spray pattern.Often, the interface can reach and maintain a local constant temperature near (T b ) for several milliseconds which may be attributed to the presence of a thin residual cryogen layer, which evaporates during and after spurt termination.Aguilar et al. [7] defined the Dt r as the period that the surface temperature remains below (the cryogen boiling point) T b % À26.2.The increase in temperature from T b up to the freezing temperature of water (T m ) is gradual and relatively linear.Often, a plateau is noted at T m , which may be attributed to condensation, freezing, and subsequent melting of ambient water on the surface.The period the surface temperature remains below 08C is defined as the Dt s .Water condensation and freezing on the sprayed surface do not occur during the spurt, despite the sub-zero temperature that may have been reached [17,18].This is because tetrafluorethane is a hydrophobic compound, which impedes the mixing of cryogen with water and, therefore, does not allow frost to form on the sprayed surface until the cryogen has evaporated completely [7].
Figures 5 and 6 show, respectively, T min and t Tmin as a function of Dt total for all spray patterns described in Figure 3 and for the two initial HSP temperatures, T i (23 and 708C).For T min , the differences between SCS and MCS with the same Dt total are small.Note, however, that T min has a strong dependence on T i (between 12 and 188C).For T i ¼ 238C, T min varies between À40 and À468C ($15%) and, for T i ¼ 708C, T min varies between À20 and À328C ($60%) for the range of Dt total under study.From Figure 5, it is possible to distinguish between two different cooling regimes.For Dt total 110 milliseconds, T min shows a linear dependence with Dt total .In the interval 110 milliseconds < Dt total 280 milliseconds, T min reaches an asymptotic value (À468C for T i ¼ 238C and À328C for T i ¼ 708C), for both SCS and MCS. Figure 6 also shows the linear dependence of t Tmin with Dt total , although the same dependence holds throughout the entire range of Dt total .The slope of this linear behavior is $1, which means that for both SCS and MCS, T min is always reached at spurt termination for all parameters under study and there is no dependence on T i .Figure 7 shows Dt s as a function of Dt total for SCS and MCS at the two T i .Similar to T min , it is possible to distinguish two different cooling regimes.In the interval Dt total 110 milliseconds, the differences in Dt s between SCS and MCS for each T i are negligible.Dt s increases linearly with Dt total and the slope is greater by a factor of two for T i ¼ 238C as compared to T i ¼ 708C.In the interval 110 milliseconds < Dt total 280 milliseconds, Dt s is notably different between SCS and MCS for both T i , and Dt s becomes greater for SCS than MCS by a factor of almost 2.
Figure 8 shows Dt r as a function of Dt total for SCS and MCS at the two T i .Similar qualitative behavior is seen for Dt r compared to Dt s (Fig. 7), except that the magnitude of Dt r for experiments at T i ¼ 708C compared to those at 238C in the interval 110 milliseconds < Dt total 280 milliseconds, can be up to four times greater.The responses to the    different spray patterns suggest that the increase in Dt s depends on the time the surface temperature remains almost constant near T b , as well as on the surface temperature rise from T b to T m (both of which may be attributed to the presence and linear evaporation of a residual cryogen layer and to the temperature gradient generated within the HSP during cryogen deposition).Longer spurts induce greater heat extraction, leading to lower HSP surface temperatures but, also, deeper cooling.The latter situation leads to lower temperature gradients close to the surface upon spurt termination and, therefore, the rate of heat transfer from the HSP to the cryogen layer is lower for longer spurts, prolonging the evaporation time, that is, Dt s , which is clearly shown in Figure 9.Note that longer spurts lead to lower average slopes of the temperature curves in the region from T b to T m , as illustrated by the dotted lines.It should be noted that the temperature in this range increases $100% faster for the shortest Dt total (40 milliseconds) than for the longest (110 milliseconds).The superposition of these two linear processes leads to the linear dependency of Dt s on Dt total .
Tuqan et al. [11] recently performed histologic evaluations of in vitro model human skin exposed to four (marked with ''*'' in Fig. 3) of our first five MCS spray patterns.Their work demonstrated that for spurts with a Dt total of 110 milliseconds or less, delivering the cryogen in MCS patterns increases the risk of cryo-injury as compared to one SCS with the same Dt c .As such, when the Dt total is less than 110 milliseconds, the risk of injury appears to depend on Dt total and not Dt c .In this study, we have shown that Dt s continues to increase for SCS patterns while it remains essentially constant for MCS patterns within the interval 110 milliseconds < Dt total 280 milliseconds, implying that for 110 milliseconds < Dt total it may be beneficial to use MCS to minimize the risk of cryo-injury to human skin during laser dermatologic surgery.
It is important to recognize that this HSP do not fully represent the thermal and mechanical behavior of human skin (e.g., skin indentation during CSC [19]), which might influence the magnitude of all the parameters measured in this study and, also, that our experimental procedures did not consider the combined application of cryogen and laser irradiation.Evidently, the heat generated by laser light absorption and scattering within the superficial layers of human skin would reduce both Dt s and Dt r .However, the present results describe the relative variation in the dynamics of the surface temperature of HSP between SCS and MCS, which is valuable information that will aid in the design of more appropriate procedures for optimal CSC on human skin.

CONCLUSIONS
An HSP was constructed and used to measure the dynamic temperature produced by MCS with the same total cryogen delivery time (Dt c ¼ 40 milliseconds) as SCS but with different Dt total .The temporal distribution of temperature for either SCS or MCS is strongly dependent on T i .Our results show that it is possible to distinguish between two different cooling regimes.For Dt total 110 milliseconds, the differences between SCS and MCS with the same Dt total are negligible for all variables under study (T min , t Tmin , Dt s , Dt r ), which show a linear dependence on Dt total .The longer Dt total , the longer t Tmin , Dt s , Dt r , and the lower T min .This response occurs for the two T i studied herein.In the interval 110 milliseconds < Dt total 280 milliseconds, however, only Dt s shows notable differences between SCS and MCS.For the same Dt total , T min is similar for SCS and MCS and Dt s become larger for SCS than MCS by a factor of almost 2.
These results suggest that similar epidermal protection may be attained with SCS and MCS in the interval Dt total 110 milliseconds.For 110 milliseconds < Dt total ¼ 280 milliseconds, MCS help to maintain Dt s similar to that of SCS at Dt total ¼ 100 milliseconds, which may be beneficial to prevent cryo-injury.
Dt c ¼ 40 milliseconds; (b) 8 MCS patterns with identical Dt c of 40 milliseconds, but with a constant time interval between consecutive spurts, which resulted in a variation of the total cooling time (Dt total ) from 50 to 280 milliseconds; and (c) 8 SCS patterns that matched the Dt total of the MCS patterns described above.A schematic of the experimental spray patterns is shown in Figure 3.

Fig. 3 .
Fig. 3. SCS and MCS spray patterns under study.The asterisk denotes the spray patterns studied by Tuqan et al. [11].Fig. 4. Schematic of a typical temperature curve versus time for a MCS at a spray distance (z) ¼ 31 mm.Characteristic features are minimum temperature (T min ); time at which T min is reached (t Tmin ); sub-zero time (Dt s ); liquid cryogen residence time (Dt r ); boiling point of R134a (T b ), and freezing point of water (T m ).

Fig. 5 .
Fig. 5. Minimum surface temperature (T min ) versus total cooling time (Dt total ) for SCS and MCS at two different T i HSP temperatures.Two different cooling regimes are distinguished: For Dt total 110, T min shows a linear decrease with the Dt total , and for 110 milliseconds < Dt total 280 milliseconds, T min is similar for SCS and MCS.

Fig. 6 .
Fig. 6.Time of minimum temperature (t Tmin ) versus Dt total , for SCS and MCS patterns at two different T i temperatures (23 and 708C).t Tmin keeps the same linear increase for all cryogen spray patterns and for both T i .

Fig. 7 .
Fig. 7. Dt s as a function of the Dt total .Dt s shows two different cooling regimes, in the interval Dt total 110 milliseconds, the differences between SCS and MCS for Dt s are negligible.In the interval 110 milliseconds < Dt total 280 milliseconds, Dt s becomes greater for SCS than for MCS by a factor of almost 2.

Fig. 8 .
Fig. 8. Dt r versus total cooling time (Dt total ).Note the obvious similarity between SCS and MCS, and the pronounced linear increase of Dt r at T i ¼ 238C.

Fig. 9 .
Fig. 9. Dynamic surface temperature versus time for three of the SCS studied herein with different Dt total .

TABLE 1 .
Thickness, Thermal and Optical Properties of Human Skin Layers Used With the Finite Element and MCML/CONV Models for l ¼ 1,450 nm Fig. 2. Experimental system employed to measure the sprayed HSP surface temperatures during the application of SCS and MCS, for two different T i HSP temperatures (23 and 708C).