Comparisons of the in vivo and ex vivo models. A, Irregularity of ablation, ablation shrinkage, and inflammation. B, Remaining thickness and thickness of the zone of necrosis.
Control skin specimen representative of the 10 patients sampled, demonstrating marked dermal elastosis and mild to moderate melanocyte atypia and hypertrophy (original magnification ×400).
Skin treated with a carbon dioxide laser in the in vivo model (A) and the ex vivo model (B) (original magnification ×90). Irregularity of ablation was vastly higher in the ex vivo specimens for all laser treatments.
Skin treated with an erbium:YAG laser in the in vivo model (A) and the ex vivo model (B) (original magnification ×90).
Skin treated with a carbon dioxide laser followed by an erbium:YAG laser in the in vivo model (A) and the ex vivo model (B) (original magnification ×90).
Greene D, Egbert BM, Utley DS, Koch RJ. The Validity of Ex Vivo Laser Skin Treatment for Histological AnalysisA Prospective Controlled Study. Arch Facial Plast Surg. 1999;1(3):159-164. doi:
From Facial Plastic and Reconstructive Surgery, Division of Otolaryngology–Head and Neck Surgery (Drs Greene and Koch), and the Department of Pathology/Dermatopathology (Dr Egbert), Stanford University Medical Center, Palo Alto, Calif; Palo Alto Veterans Affairs Health Care System (Drs Greene, Egbert, Utley, and Koch); and the Department of Otolaryngology, Cleveland Clinic Florida, Ft Lauderdale, Fla (Dr Greene).
Background Laser treatment of skin following removal from human subjects has been the staple of laser research. However, no study has been done to assess the efficacy of ex vivo skin for predicting the behavior of laser treatments in living human tissue.
Objective To assess the validity of the ex vivo model by comparing histological characteristics of skin treated with laser prior to and following its removal in rhytidectomy.
Study Design Nonrandomized controlled intervention study in which each patient served as both experimental subject and control for different skin sites.
Patients Ten patients with actinic skin changes.
Interventions Patients underwent laser treatment to 4 left preauricular sites 1 hour prior to rhytidectomy as follows: carbon dioxide laser treatment alone, carbon dioxide laser treatment followed by erbium:YAG laser treatment, erbium:YAG laser treatment alone, and erbium:YAG laser treatment followed by carbon dioxide laser treatment. The skin was examined by a dermatopathologist blinded to the identity of each specimen. Untreated skin was also removed and immediately subjected to laser treatment identical to that employed in the in vivo skin. This skin was examined histologically.
Main Outcome Measures Regularity of ablation, depth of the necrotic zone, amount of skin removed, degree of collagen injury, and degree of inflammation.
Results There were significant differences between the ex vivo and in vivo groups. The ex vivo specimens demonstrated more than 10 times the irregularity of ablation of the in vivo specimens (irregularity index of 3.0 for the ex vivo group vs 0.25 for the in vivo specimens; P<.05). The incidence of collagen injury was slightly lower for the ex vivo group (1.0 vs 1.3), as was the degree of inflammation (1.4 vs 1.5). The greatest differences were the significantly smaller necrotic zone in the ex vivo specimens (51 vs 71 µm) and the smaller amount of skin removed (118 vs 234 µm). These findings were consistent for all 4 laser treatment regimens studied.
Conclusions Significant differences were found between the in vivo and ex vivo models. Irregularity of ablation in the ex vivo specimens was 10 times that in the living specimens, limiting histological accuracy in the ex vivo model. The ex vivo skin model underestimated the amount of tissue ablation. This suggests that an in vivo model should be adopted as the standard for laser research.
RAPID ADVANCES in laser technology have required histological studies to evaluate the effects of lasers on tissue, such as depth of penetration, necrotic zone, uniformity of ablation, and other characteristics.1-5 The majority of studies have used skin treated ex vivo following resection of the skin during rhytidectomy or blepharoplasty, making ex vivo skin studies the criterion standard for studying laser-induced histological changes. More recently, reports using human and animal in vivo models have appeared that provide data on the effects of laser in physiologically normal living skin.2, 6-8 To our knowledge, the similarities and differences of these 2 models have not been examined.
Living skin has several characteristics that differ greatly from those of resected skin samples. Skin in vivo hasactive circulation by way of the subdermal plexus and superficial capillaries, lymphatic drainage, the production of sweat and sebum, and a high resting turgor. When skin is removed from the patient, as in rhytidectomy, these characteristics are also removed. The skin quickly loses turgor and shrinks, since it is no longer under tension. How these physiological differences affect reaction to laser treatment is unknown.
The carbon dioxide laser has an established record of safety and efficacy for the application of facial skin resurfacing.1, 9-10 The advantages of the carbon dioxide laser are reported to be thermal contraction of collagen,11 widespread familiarity of physicians with carbon dioxide laser technology, improved technology with the addition of the computer pattern generator handpiece,12 lower equipment costs than alternative lasers, and adequate delineation by previous authors of the histological effects on tissue and the clinical outcomes expected from this laser.1-12
Potential disadvantages of the carbon dioxide laser include the creation of a thermal injury zone deep to the ablation zone as well as diminishing ablation depth and increasing thermal necrosis zone with each successive pass.1-2,13-16 Excessive thermal injury has been implicated as a possible cause of the prolonged erythema, hyperpigmentation, and delayed wound healing sometimes observed after treatment with the carbon dioxide laser.1-2,16-17
Considerably less information is available for the erbium:YAG laser. The erbium:YAG laser has a coefficient of absorption in water that is 10 times greater (7700/cm) than that of carbon dioxide and reduced optical penetration in tissue (1 vs 20-30 µm).1-2 The erbium:YAG laser is reported to cause less thermal necrosis and less tissue ablation per pass, and therefore less erythema and fewer wound healing complications.4, 18-20 The use of the erbium:YAG laser in conjunction with the carbon dioxide laser to reduce the thermal necrosis zone while maintaining some of the collagen-tightening advantages of the carbon dioxide laser is promising.13, 21 In a previous study using an in vivo preauricular skin model,13 we found that treatments that combined carbon dioxide and erbium:YAG lasers were superior in terms of healing and beneficial histological changes.
Criteria critical to the assessment of differences between different laser regimens include the amount of tissue removed with a given fluence, irregularity of ablation, thickness of the necrotic zone, ablation shrinkage, and inflammation. To assess the ability of each model to measure these differences, we chose to perform resurfacing of test spots with 4 regimens combining carbon dioxide and erbium laser treatment, the 2 most popular laser treatments used for resurfacing.
To date, no attempt has been made to systematically assess the validity of ex vivo skin as a model for use in laser research. The objectives of this study are to examine the histological differences between skin treated by laser in vivo and ex vivo and to provide tentative guidelines for evaluating the histological findings from the 2 models.
Ten patients scheduled to undergo cervicofacial rhytidectomy were enrolled in this laser study approved by the Stanford University Institutional Review Board. Two lasers were used. The Luxar LX-20SP Novapulse carbon dioxide laser (10,600 nmol/L) with SureScan computer pattern generator (Luxar Corp, Bothell, Wash) was used in the superpulse mode at 6 W and 16 Hz with a pulse duration of 730 microseconds. The 7.6×6.4-mm computer pattern generator parallelogram was used. This laser delivered a fluence of 4.7 J/cm2. The ESC Derma-20 Erbium:YAG laser (2940 nmol/L) (ESC Medical Systems, Yokneam, Israel) was used with a 6-mm spot size at 14 W and 8 Hz with a pulse duration of 350 microseconds to deliver 1.7 J per pulse and a fluence of 4.7 J/cm2. These parameters were selected to provide identical fluences for the laser systems being compared in this study.
Each patient was examined preoperatively to determine Fitzpatrick skin types22 and to grade the degree of facial skin actinic damage and rhytidosis using a 5-point scale. One hour prior to rhytidectomy, each patient received an injection of 2 mL of 2% lidocaine in the subcutaneous plane of the right preauricular region; dermal infiltration was avoided. Single or combination laser energy was applied to 4 separate target areas as follows: (1) carbon dioxide laser (4 passes), (2) carbon dioxide laser (2 passes) followed by erbium:YAG laser (4 passes), (3) erbium:YAG laser (8 passes), and (4) erbium:YAG (4 passes) followed by carbon dioxide laser (2 passes).
The skin was wiped after each pass with the carbon dioxide laser.
The in vivo skin (treated 1 hour prior to excision) was excised at the outset of the rhytidectomy procedure to minimize tissue trauma and minimized to reduce potential edema formation. Immediately after excision, each laser site was identified, separated from adjacent skin, and placed in an individually labeled container in 10% formaldehyde solution. An additional untreated specimen was submitted as a control.
An untreated crescent of skin was similarly excised at the outset of the procedure, immediately transferred to the back table, patted with saline-dampened gauze, patted dry with gauze, and treated with a laser in the same way the in vivo skin was treated; it was then transferred to 10% formaldehyde solution. The rhytidectomy procedure was completed in the usual manner.
One histopathology technician prepared all specimens on the same day. Hematoxylin-eosin staining was used. Each slide was labeled with a study code to allow later identification of each specimen by the type of laser used and the interval of treatment. The study dermatopathologist (B.M.E.) remained blinded to this code throughout the histological evaluation process. The code was revealed after all histopathologic data were recorded and ready for analysis.
Control specimens were histologically graded using a 5-point scale for degree of melanocyte atypia, hypertrophy, hyperplasia, epidermal atypia, polarity abnormalities, and parakeratosis: Article
Dermal elastosis was graded. An optical micrometer was used to determine the thickness of the epidermis, papillary dermis, and reticular dermis.
Specimens treated 1 hour prior to excision (in vivo) and specimens treated following skin resection (ex vivo) were graded for irregularity of ablation, collagen injury, and inflammation. The optical micrometer was used to determine the thickness of the thermal necrotic zone and the distance from the ablated surface to the dermal-subcutaneous junction.
Finally, the laser changes in the in vivo and ex vivo specimens were compared. The t test was used to determine the significance of differences between the 2 groups. Differences at P≤.05 were considered statistically significant.
Ten patients (7 men, 3 women; mean±SD age, 49.7±6.6 years) were enrolled as candidates for rhytidectomy and participation in this laser study. All had Fitzpatrick skin type II or III (mean±SD, 2.3±0.5). Gross actinic skin damage scores ranged from 1 to 4, with a mean of 2.5±1.0. The observed degree of facial rhytidosis scores ranged from 1 to 4, with a mean of 2.2±0.8.
Ninety sites on 10 patients were examined histologically. Each patient had skin samples harvested at 4 sites that were treated with the 4 laser regimens 1 hour prior to rhytidectomy and 4 sites that were treated with laser following resection at rhytidectomy. One additional site on each patient was examined histologically as a control.
The 10 control specimens demonstrated mild to moderate melanocyte atypia and hypertrophy but minimal hyperplasia on histological examination. There were mild to moderate epidermal atypia and polarity abnormalities, minimal parakeratosis, and marked dermal elastosis. Mean control skin layer thickness values were 65 µm for the epidermis, 55 µm for the papillary dermis, and 1360 µm for the reticular dermis.
Comparison of 5 criteria (irregularity of ablation, ablation shrinkage, inflammation, remaining thickness, and thickness of necrotic zone) between the in vivo and ex vivo treated specimens revealed significant differences between the 2 models, even when the means of all laser treatments were considered (Figure 1). Irregularity of ablation was minimal in the in vivo group (0.25) but very significant in the ex vivo group (3.0),3 with a 12-fold greater degree of irregularity. The degree of ablation shrinkage was higher in the in vivo group (1.18) than in the ex vivo group (1.03). The ex vivo model overestimated the remaining thickness (1359 µm in the ex vivo specimens vs 1280 µm in the in vivo specimens). This is secondary to lesser amounts of skin removal (118 vs 234 µm). It also underestimated the thickness of the necrotic zone (51.3 µm in the ex vivo group vs 71.3 µm in the in vivo group). Only inflammation demonstrated minimal differences, with 1.48 µm in the in vivo group and 1.43 µm in the ex vivo group.
Comparison between the in vivo and ex vivo models was even more striking when assessed for each laser regimen individually (Table 1). Irregularity of ablation was vastly higher in the ex vivo specimens for all laser treatments. Irregularity was 13 times higher for carbon dioxide laser treatment alone (P<.001), 4 times higher for erbium:YAG laser treatment alone (P<.05), 16.5 times higher for carbon dioxide laser treatment followed by erbium:YAG laser treatment (P< .005), and 17.5 times higher for erbium:YAG laser treatment followed by carbon dioxide laser treatment (P<.001).
The ex vivo model underestimated the thickness of the necrotic zone by 40% (P<.05) for erbium:YAG laser treatment followed by carbon dioxide laser treatment. The ex vivo model's underestimation of ablation shrinkage by 24% (P=.05) approached statistical significance for this laser treatment regimen. For treatment with the erbium:YAG laser alone, differences in inflammation and thickness of the necrotic zone approached statistical significance (P=.08).
Rapid advances in laser technology have necessitated the development of human skin models to compare the efficacy and safety of the various laser regimens. To date, the vast majority of these studies have been done on skin that has already been removed from the body (the ex vivo model).9, 18 Although more recently published reports use in vivo models (human and animal), none of these studies has systematically compared in vivo and ex vivo models.2, 6-8 Thus, ex vivo models have been the gold standard without ever having been systematically compared with in vivo models. The present study demonstrates significant differences between these 2 models that suggest that the ex vivo model is limited in its ability to predict laser-induced changes in living skin.
Comparison of the in vivo and ex vivo models demonstrates statistically significant differences in their measurement of key laser-induced histological changes. This was found to be true with a high degree of statistical significance for irregularity of ablation (P<.001). The difference in the thickness of the necrotic zone was significant for erbium:YAG laser treatment followed by carbon dioxide laser treatment (P<.05); differences approached statistical significance for several other histological measurements.
The extreme differences in irregularity of ablation that were found in all laser regimens suggest that ex vivo laser testing is an unreliable tool. With such irregular ablation, different microtome slices from the same specimen may be drastically different. Indeed, this 1 factor alone may lead to the differences between in vivo and ex vivo testing seen in the other 4 variables (ablation shrinkage, inflammation, thickness remaining, and necrotic zone thickness).
We have previously described the use of the preauricular crescent of skin excised in rhytidectomy for in vivo laser testing with excellent results and a high level of patient acceptance.13 Stuzin et al1 and Cotton et al2 also had excellent results using preauricular skin models. Despite aging and actinic changes, preauricular skin retains several characteristics needed for laser testing. These include a layered configuration with preservation of dermis and epidermis, and preauricular skin is representative of skin that is routinely lasered and left in place. Eyelid skin that is removed in blepharoplasty, by contrast, is of much poorer quality, lacking the thickness and uniformity this type of testing requires. In addition, it models only eyelid skin and cannot necessarily be extrapolated to predict the effects of the laser on the thicker, more sebaceous facial skin. The present study found patients very accepting of this technique, and the present data support the continued use of this model in laser testing.
Significant differences exist between the changes observed when lasers are tested on skin on the back table vs living skin (Figure 2, Figure 3, Figure 4, and Figure 5). The most striking difference is the highly significant degree of irregularity of ablation in the ex vivo model. Indeed, highly irregular ablation produces variability between histological slices from the same specimen, thus limiting the ex vivo model's ability to test other variables, as well as ablation irregularity. Therefore, the present data support the in vivo preauricular model as a possible gold standard for laser testing in the future.
Skin treated with lasers ex vivo does not accurately predict changes found in skin treated with lasers in vivo and removed later. Thus, lasering on the back table is not as reliable as the in vivo model for testing laser and tissue interaction.
Irregularity of ablation may be as much as 17.5 times greater in ex vivo vs in vivo models (P<.001). This fact limits the reliability of histological analysis in the ex vivo model.
The preauricular skin in vivo model should be the gold standard for laser testing because this skin is the most representative of facial areas in laser resurfacing.
Accepted for publication May 17, 1999.
Corresponding author: David Greene, MD, Division of Otolaryngology/Head and Neck Surgery, Cleveland Clinic Florida, 3000 W Cypress Ave, Ft Lauderdale, FL 33309 (e-mail: Dgreene@post.harvard.edu).