Analysis of nonablative skin resurfacing

Nonablative skin resurfacing is a dermatologic procedure utilizing pulsed laser irradiation and dynamic cooling to induce selectively a wound healing response in the papillary and upper reticular dermis. Using temperature measurements of human skin provided by pulsed photothermal radiometry immediately following laser irradiation (lambda equals 1.32 micrometer), spatial distribution of thermal damage is predicted in response to various potential therapeutic laser- cryogen doses. Results of our analysis suggest that appropriate application of pulsed laser irradiation and cryogen spray cooling may be used to protect the epidermis and selectively confine thermal injury to the papillary and upper reticular dermis. Development of nonablative skin resurfacing will require understanding the relationship between the degree of dermal photocoagulation and the cutaneous wound healing response following laser irradiation.


INTRODUCTION
Laser-assisted skin resurfacing recently has emerged as a dermatologic procedure for reducing facial rhytides (i.e., wrinkles). In practice, a superficial layer of skin material is removed through an ablative mechanism frequently utilizing infrared radiation emitted from a CO2 laser (10.6 jtm). Recent investigations into the origin of the underlying mechanisms of laser ablative resurfacing procedures, however, indicate removal of skin material may not be the only process required for successful therapeutic outcome. In a histological study, Cotten et. al. noted reduction of human facial rhytides following CO2 laser skin resurfacing may correspond to the formation of a new band of connective tissue formed in the papillary dermis.' The new tissue was composed of dense, compact collagen bundles oriented parallel to the skin surface. In addition, thickness of the new band of connective tissue strongly correlated to a corresponding measure of thermal denatured collagen. A light microscopic polarization sensitive histological study by Thomsen et. al., of rodent skin irradiated with laser parameters corresponding to a CO2 resurfacing procedure in humans indicated thermal alteration of collagen fibrils in the dermis may occur at depths up to 250 m.
Inasmuch as the CO2 resurfacing technique partially relies on induction of a wound healing response following thermal injury in the superficial skin layers, a novel laser procedure utilizing "dynamic cooling' has been devised to protect the epidermis from damage while achieving thermal denaturation of proteins in the papillary and upper reticular dermis. In this procedure, the epidermis is rapidly cooled prior to pulsed laser irradiation by a short spurt of cryogen sprayed onto the skin surface.2 Subsequent absorption of pulsed laser energy by water in skin results in a temperature increase in the papillary dermis sufficient to cause spatially selective protein thermal denaturation. Because skin temperature at superficial depths is decreased by dynamic cooling, thermal injury in the epidermis may be substantially reduced or eliminated.
The purpose of this study is to use temperature measurements provided by pulsed photothermal radiometry of human skin immediately following laser irradiation to predict the spatial distribution of thermal damage in response to various potential therapeutic laser-cryogen doses. Application of a non-negatively constrained conjugate gradient inversion algorithm3 and model of dynamic cooling are used to predict the space-time temperature response of skin. Spatial distribution of thermal injury is computed at potential therapeutic laser-cryogen doses by application of a first-order rate SPIE Vol. 2970 • 0277-786X197/$10.00 367 368 damage model.4 Comparison of the predicted spatial distribution of thermal injury in skin with that observed in animal model studies is discussed.

METHODS AND MATERIALS
To evaluate the thermal response of human skin to laser irradiation without dynamic cooling, pulsed photothermal radiometry measurements (PPTR) are recorded5 at seventy-one in-vivo test sites on volar forearm of human volunteers. The radiometric temperature change (S(t)) of an irradiated site is measured by collecting and focusing infrared emission from the skin onto a 1 mm2 liquid nitrogen cooled HgCdTe detector sensitive in the 10-14 tm spectral region. An f/i, 25 mm diameter germanium lens configured for unit magnification is positioned 50 mm from the detector to give a 1 mm2 square field at the skin surface. A silica multimode optical fiber (400 jim core diameter) is used to deliver pulsed laser radiation (x=l.32 jtm) emitted from the open tip onto the skin surface. Value of laser dosage (5-30 Jcm2) is determined for each irradiated skin site from independent measurements ofthe spot size (17-28 mm2) and pulse energy (1-5.3 1).
For each recorded PPTR signal (AS(t)), the one-dimensional spatial distribution of the initial temperature increase in skin (KFL(z,t=O)) immediately following pulsed laser irradiation is computed by application of a non-negatively constrained conjugate gradient inversion algorithm;6 S(t) and TL(z,t=O) are related by a Fredhoim integral equation ofthe first kind (Eq. 1A).
Here, P-a 5 the infrared absorption coefficient of skin at the detected wavelength(s), and Cd is a proportionality constant determined by the detection system. Additionally, erfcx(u)=exp(u2)erfc(u) is the exponential complementary error function and u,1 are functions (Eq. 1B) of space (z), time (t), thermal diffusivity of skin (D), and heat transfer coefficient (h) of the air-skin boundary.
Spatial temperature distribution immediately following cryogen spray cooling (zT(z,t=O), Eq. 2A) has been extensively investigated experimentally and theoretically by Anvari et. a!.,2 Here, -t, < t 0, where t is the spray cooling time, T is the temperature ofthe cryogen-ice mixture on the skin (e.g., T -10 °C), and T0 is the initial skin temperature. Immediately following pulsed laser irradiation, we assume a heat loss coefficient h at the air-skin boundary so that the spatial distribution of the initial temperature increase (zT(z,tO)) is a Knowledge of the thermal diffusivity (D) of human skin7 together with the spatial distribution of the initial temperature increase (zT(z,t=O)) allows computation of the space-time evolution of the thermal profile (Eq. 4, T(z,t)), Figure 1: Steps to compute the fraction of denatured protein (1(z)) from a measured PPTR signal (zS(t)).

(5)
PPTR signals recorded in response to pulsed laser irradiation indicate a rapid temperature increase followed subsequently by a slower monotonic decrease (Figure 2A). Application ofthe non-negatively constrained conjugate gradient inversion algorithm to the measured PPTR signal (AS(t)) indicates a subsurface peak of the initial temperature increase with a monotonic decrease at deeper positions ( Figure 2B). Magnitude of the temperature decay with increasing skin depth given by the inversion algorithm is greater than that predicted from simple diffusion theory; we believe the observed difference may be due to the use of one-dimensional models that do not correctly account for lateral propagation of photon or thermal flux.

369
Spatial distribution of thermal injury (1(z)) is computed at two potential therapeutic laser doses (25Jcm2 and 30 Jcm2) by application of a first-order rate damage model (Eq. 5), here, 1(z) represents the fraction (0 1(z) 1) of denatured protein at depth z. Values for temperature dependent parameters [i.e., A(T(z,t)) and EA(T(z,t))] in the damage integral (Eq. 5) were determined by Weaver and Stoll. 8 We summarize ( Figure 1) the processing steps to compute the fraction of denatured protein (1(z)) from a measured PPTR signal (S(t)). The peak temperature increase (iSpeaic) from recorded PPTR signals is plotted (Figure 3) versus the corresponding laser dose for each in-vivo skin test site. The recorded data suggest Spe varies linearly with increasing dose. Linear least squares analysis ofthe recorded data gives a slope of 1.2 (°CJ1cm2). The initial temperature increase (tT(z,t=O)) due to pulsed laser irradiation and cryogen spray precooling (t=4O ms, T=-10 °C, and h=4O kWm2) is determined by superposing ATL and Tc ( Figure 4). The combined effect of pulsed laser irradiation and cryogen spray cooling selectively reduces temperature of the epidermis while creating a peak subsurface increase at z=O.2 mm.

PPTR
Determination ofthermal damage distribution (1(z)) following pulsed laser irradiation and cryogen spray cooling requires computation of T(z,t). Computation of T(z,t) at three time points ( Figure 5 Spatial distribution of thermal injury (1(z)) was computed using the damage integral analysis discussed above (Eq. 5).
Mean fraction of denatured protein ( Figure 6) was determined by averaging each 1(z) for for each irradiated skin site at potential therapeutic doses, D=25 Jcm2 and 30 Jcm2.

DISCUSSION
Results of our analysis suggest that appropriate application pulsed laser irradiation and cryogen spray cooling may be used to protect the epidermis and selectively confme the spatial distribution of thermal injury to the papillary and upper reticular dermis. Although a number of assumptions (e.g., one-dimensional model) in our analysis may compromise the numerical accuracy of the results, the ability to limit preferentially spatial distribution of thermal injury to the papillary dermis is supported by histological studies in a porcine skin model.9 Furthermore, although the spatial distribution of thermal injury in skin may be confmed in the papillary and upper reticular dermis, therapeutic efficacy of nonablative skin resurfacing is detennined primarily by the cutaneous wound healing response.
Despite the advances that have been achieved in understanding the important cellular and biochemical processes involved in the cutaneous wound healing response,1° many important questions remain. For example, changes in the wound healing response for various levels of dermal photocoagulation and accompanying damage to the skin microvasculature require additional study. Specifically, in nonablative skin resurfacing, dermal photocoagulation just above the thermal damage threshold will result in slightly increased levels of new collagen and proteoglycan synthesis. As the degree of dermal photocoagulation becomes greater, subsequent protein synthesis will also increase in order to replace greater quantities of thermally damaged tissue. One may not expect, however, that the degree of dermal photocoagulation may be arbitrarily increased in nonablative resurfacing procedures without increasing the risk of complications such as blistering, induration, or possibly skin sloughing. For example, increased levels of dermal photocoagulation may irreversibly damage and temporarily occlude the microvasculature in the papillary dermis resulting in a thick (-300 jim) region ofnecrotic tissue which can not be regenerated without sloughing.
Inasmuch as the primary advantage of nonablative skin resurfacing is reduced patient discomfort and simpler postoperative wound care, retaining the viability of at least a portion ofthe dermal microvasculature is important. Additional quantitative studies are required to ascertain accurate values of the scattering coefficient of the papillary and reticular dermis and the dependence of subsurface fluence on the numerical aperture of incident light. Inasmuch as the optical properties of skin vary between different sites and subjects, a subtherapeutic diagnostic temperature measurement in response to pulsed laser irradiation may be required to determine the appropriate cryogen-laser dosimetry. Most