Lasers Med Sci 2002, 17:154–164Ownership and Copyright© 2002 Springer-Verlag London Limited

In Vivo Study of Intradermal Focusing for TattooRemoval

X.H. Hu1, W.A. Wooden2, S.J. Vore3, M.J. Cariveau1, Q. Fang1 and G.W. Kalmus4

1Department of Physics, 2Department of Surgery, Brody School of Medicine, 3Department of Comparative Medicine, BrodySchool of Medicine, 4Department of Biology, East Carolina University, Greenville, North Carolina, USA

Abstract. Delivery of intradermally focused nanosecond laser pulses with small energy as an alternatetechnique applicable to clinical procedures in dermatological and plastic surgery is an area of relatively newinterest with multiple potential applications. We assessed this approach on common tattoo pigments in dermisin an in vivo study using a wavelength of 1064 nm. Paired micropigs were tattooed with standard blue, black,green and red pigments. The tattoos were allowed to mature and then treated by 12 ns pulses in a focused beamof 11.4� cone angle. Visual observation and histological analysis of biopsies were performed to evaluateresults. Significant reduction in pulse energy and collateral damage was achieved with pulse energy rangingbetween 38 to 63 mJ. Blue and black tattoos were found to respond well from a clinical standpoint. The depthdependence of tissue response and pigment redistributions at 1 hour, 1 week and 1 month after laser treatmentwas quantitatively analysed through biopsies and a strong relationship was demonstrated between tattooresponse and laser-induced dermal vacuolation. The optical absorption coe$cients of the four tattoo pigmentswere measured to be approximately the same and the laser-induced plasma is suggested to be responsible forthe pigment redistribution. As we hypothesised, intradermal focusing of nanosecond pulses significantlyreduced required pulse energy for tattoo ablation to about 60 mJ or less. These results stimulate a number ofadditional questions relevant not only to clinical applications but also to the understanding of thefundamental process of laser–pigment interaction in the dermis as it relates to tattoo removal.

Keywords: Intradermal focusing; Laser surgery; Tattoo removal

INTRODUCTION

There is a tremendous volume of research inboth basic and clinical sciences investigatingthe medical applications of laser technology.The development of laser systems capable ofscanning a beam of small energy and intra-dermally focused nanosecond (ns) pulses posethe question of how these systems could aug-ment or improve clinical procedures in derma-tological and plastic surgery. The potentialareas of application could include not onlytattoo removal and pigmented lesions butextend also to vascular malformations andother lesions. The process of tattooing is onethat lends itself readily to in vivo clinicalinvestigations and the potential for developingnew laser systems for more e#ective tattooremoval could be very beneficial consider-

ing the current worldwide trend of increasedpopularity of tattooing and the subsequentdesire for removal. The study of tattoo removalalso allows for the academic pursuit of basicbiophysics in understanding the complexityand fundamental problems of light distributionin the skin and the pigment ablation process.

Q-switched lasers are widely used in tattooremoval as the sources of nanosecond pulsesdue to their relatively simple system configur-ations and high reliability in comparison toother short-pulsed lasers. Although a selectivephotothermolysis model [1] has been exten-sively applied to interpret the data obtainedwith the ns pulses [2–6], very few quantitativein vivo investigations have been conductedto provide critical insight into the mechanismof the ablation process within the skin. Asa result, fundamental problems remain un-answered. For example, what is the role of thetissue or pigment absorption in the ablationprocess in relation to the strong electro-magnetic field created by the pulse during

Correspondence to: Dr Xin-Hua Hu, Department of Physics,East Carolina University, Greenville, NC 27858, USA. Tel:252-328-6476; Fax: 252-328-6314; e-mail: hux@mail.ecu.edu

the short time scales of nanoseconds?Additionally, what is the e#ect of cavitationand transient acoustics induced by the ns laserpulses on the pigments and pigment mobilis-ation? Finally, can one reduce further thecollateral tissue damage by decreasing theenergy of a ns pulse delivered into the skin?This lack of a clear understanding of the mech-anism underlying skin and pigment ablationby ns pulses has impeded the further improve-ment of the way we deliver the pulses fortreating skin lesions ever since the Q-switchedlasers were introduced into dermatologicalclinics in the 1980s [8], with a telling manifes-tation of this situation demonstrated by thefact that multiple Q-switched laser systemswith pulse energy of 1 J or larger are needed inmost clinics practising laser tattoo removalprocedures for treating di#erent coloured pig-ments [6]. Recently, we have studied the mech-anism of skin ablation by ns laser pulses [9]and skin optics [10] in vitro in our e#orts toinvestigate a new approach to delivering nspulses in treating multicolour pigmentedlesions with reduced collateral tissue damage[11]. As a prelude to our e#orts in identifyingan e$cient clinical method of treating skinlesions with short laser pulses, we studiedtattoo removal in vivo in a micropig animalmodel using a focused beam. In this paper wereport initial results from two micropigs using12 ns pulses at 1064 nm wavelength from aQ-switched Nd:YAG laser and pulse energiesbetween 38 and 63 mJ.

MATERIALS AND METHODS

Two female Sinclair Yucatan micropigs(Charles River Laboratories, Inc.) were used inour study because the histology and healingrate of porcine skin are similar to that ofhuman skin [4]. All manipulations involvingthe micropigs were accomplished in strictaccordance with requirements of the AnimalWelfare Act and the National Institutesof Health ‘Guide for the Care and Use ofLaboratory Animals’ and followed an AnimalUse Protocol approved by the Animal Care andUse Committee of East Carolina University.The 3-month-old pigs, weighing approximately12 kg, underwent tattooing procedures at theonset of the research project. Under generalanaesthesia, each pig received four tattooareas each consisting of four tattoo strips,70 mm long (x-axis) by 10 mm wide (y-axis),

spaced about 35 mm apart to facilitate laterlaser treatment and biopsy procedures. Inaddition, a circular spot of about 10 mm diam-eter and of the same colour was tattooed nearone end of each tattoo strip. These circularspots were not treated by laser and wereused as the baseline for evaluation of removale$cacy on the tattooed strips. The tattoo areaswere placed just behind the shoulder regionand just anterior to the hip roughly midwaybetween the spinal column and sternum. Aclinical nurse specialist experienced in theprocess performed all tattooing proceduresusing a medical tattoo unit. Four di#erentpigments (Lasting Impression I, Inc.) wereused in each tattoo area and all sites weremonitored for 10 days postprocedure with noobserved infection.

The absorption coe$cients of the tattoo pig-ments were determined by measuring the colli-mated transmittance with a cw laser beam ofpower of 0.3 mW from a Nd:YAG laser at awavelength of 1064 nm. The laser beam wasmodulated at 1 kHz and transmittance wasmeasured by a photodiode and a lock-inamplifier. All pigments were dissolved in 50%alcohol at a concentration of 1% by weightand contained in a glass cuvette of path length12.4 mm. A weak scattering component in thetransmitted light was filtered out by two 2 mmapertures, separated by a distance of 100 mm,before the photodiode. Since the tattoo pig-ment particles are mostly metallic and thushave a mass density larger than that of thealcohol, care was taken to ensure the homo-geneity of the solutions during the measure-ments. The results, shown in Table 1, wereconfirmed with absorbance measurements ofthe pigments in further diluted solutions by afactor of 8 using a UV-VIS spectrophotometer(8453E, Agilent Technologies). In comparison,the absorption coe$cient of 50% alcohol at1064 nm was measured to be only 7.38�10�3 cm�1. We found that variation in theabsorption coe$cient of the pigment solutionsat 1064 nm was less than �6%, with the blueand red absorbing less than black and green.These findings were in marked contrast to thelarge di#erence in the visible region.

Approximately 40 days after tattooing andunder general anaesthesia, the tattoo strips oneach micropig were treated with 12 ns pulsesat 1064 nm from a Q-switched Nd:YAG laser(Surelite I, Continuum). The pulse repetitionrate was set at 10 Hz and the M2 factor for thebeam quality was measured as 1.30. The laser

In Vivo Study of Intradermal Focusing for Tattoo Removal 155

beam was collimated and expanded to a diam-eter of 15 mm using a pair of spherical lensesbefore entering a specially designed focusingcup (Fig. 1). The focusing cup consisted of twoparts with the outer part firmly adjacent to atemplate on the skin and the inner part con-taining a planoconvex lens of 75 mm focallength and 25 mm diameter to produce a con-verging beam with a full cone angle of �=11.4�in the air. The inner part carrying the focusinglens was adjustable along the laser beam axis(z-axis) with a 0.1 mm precision to vary thedepth of the focal point below the pig skinsurface. To reduce the beam defocusing due tothe refractive index mismatch at the roughskin surface [12], we employed a 1 mm thick

film of 0.9% saline solution. This film wasformed within the template that was pressedagainst the skin with a rectangular opening of60�7 mm2 area centred over a tattoo strip.Dry nitrogen gas was circulated continuouslythrough the cup chamber during the treatmentto reduce moisture condensation on the focus-ing lens and the lens was cleaned betweeneach treatment procedure. The pulse energycited in Tables 2–5 was determined afterthe focusing lens from the average powermeasured before and after each treatmentusing a power meter (30-15-A-P-HE, Ophire)with the fluctuation in average power less than3%. A schematic side view of the set-up isshown in Fig. 1.

A clinical version with a beam-scanningdevice is yet to be designed for this system.Therefore, in the experimental system theanimal was used as a moving target on amotor-driven translation table as opposed toscanning the laser beam. The pig was trans-lated along the longitudinal direction of thetattoo strip (x-axis) at a constant speed toensure that only one pulse was delivered perspot with a spot-to-spot distance of 100 �m overthe 60 mm length of the treated tattoo strip.After a complete pass, the table was translatedmanually by 190 �m in the y-axis followed by areversed x-axis translation. Treatment of onetattoo strip was completed with about 30passes. Digital photos and punch biopsies ofthe tattooed skin for each test strip wereobtained immediately prior to lasering and at 1hour, 1 week, 2 weeks and 1 month after lasertreatment for clinical and histological evalu-ation. The biopsies were fixed in 10% bu#eredformalin and processed by routine histological

Table 1. Properties of tattoo pigments used in the study

Pigment Blue (navy liner) Black Red (red lip liner 3) Green (green 2)

Chemicalcompositiona

Iron oxide,ultramarine blueand violet, glycerin,isopropyl alcohol,and titanium dioxide

Iron oxide,glycerin,isopropylalcohol,titaniumdioxide

(Organic) ironoxide, D & C redno. 30, D & C redno. 7, glycerin, andisopropyl alcohol

Chromiumhydroxidegreen, glycerin,isopropylalcohol, andtitanium dioxide

Absorptioncoe$cient (cm�1)b

4.97 5.63 5.08 5.52

aObtained from the pigment bottle labels.bThe absorption coe$cients of pigments (0.1 g) dissolved in 50% alcohol (10 g) were determined through measurement ofcollimated transmission at 1064 nm.

Fig. 1. Schematic side view of the experimental set-up withan intradermally focused laser beam and the micropig trans-lated by a motorised table. The insert shows the focusedbeam penetrating through a saline film on the skin with a coneangle � exaggerated for better viewing.

156 X.H. Hu et al.

procedures to obtain stained slides. Additionalgross examinations were made and photostaken three months after the laser treat-ment immediately before the animals wereeuthanised to evaluate the late phase con-dition of the treated skin areas. The system ofevaluation consisted of skin appearanceassessment and histological analysis. Visualappearance assessment was carried out in ablinded fashion by a plastic surgeon exper-ienced in the clinical treatment of tattoosusing photographs of tattoo test areas andwas subsequently confirmed by another evalu-ator. Histological analysis was conducted intwo ways: the first in analysing the actualtissue damage and the second in quantify-ing the pigment mobilisation through depthdependence.

RESULTS

Because of the relatively small cone angle of11.4� used in our study, laser-induced break-down in the saline on the skin surface in whichshockwaves were clearly audible occurredfrequently for treatments with focal depthsless than 3 mm. Consequently, the actual pulseenergies delivered inside the skin wereexpected to be smaller than those measuredbefore the saline film and cited in Tables 2–5because of the shielding e#ect of laser-inducedplasma in the saline. Previous results [12] indi-cated that for a 6ns pulse at 1064 nm with acone angle of 22� the pulse transmissionthrough a plasma induced in water decreasedto about 50% at the breakdown thresholdfrom nearly 100% below the threshold. The

Table 2. Response of blue tattoos

Tattoo ID 0C 0F 0K 0N 9C 9F 9K 9N

Pulse energy (mJ) 63 63 63 49 56 55 55 55Focal depth (mm) 5.9 5.9 4.7 4.7 4.7 7.1 2.4 2.4Petechial haemorrhagea Mod. Mod. Mod. Mod. Min. Min. Mod. Min.Tattoo removala Good Good Exc. Good Good Exc. Good GoodHypopigmentationa Min. Min. Min. Min. Mod. Mod. Mod. Mod.Index of epidermis necrosis – 1 hourb 0 0 0 0 0 0 0 0Sum index of vacuolation – 1 hourb 11 0 4 18 13 9 20 10Sum index of aggregated pigment – beforeb 34 12 17 12 34 12 17 12Sum index of aggregated pigment – 1 hourb 15 0 5 16 20 9 23 12Sum index of aggregated pigment – 1 weekb 0 0 0 0 4 0 0 0Sum index of aggregated pigment – 1 monthb 0 4 20 20 15 21 0 15

aDescriptions of the scales are given in the Results section.bMeasured from biopsy slides immediately before, 1 h after, 1 week after and 1 month after laser treatment.

Table 3. Response of black tattoos

Tattoo ID 0A 0H 0I 0P 9A 9H 9I 9P

Pulse energy (mJ) 62 63 63 49 55 55 55 38Focal depth (mm) 5.9 5.9 5.9 4.7 3.6 5.9 2.4 1.3Petechial haemorrhagea Min. Mod. Min. Mod. Mod. Mod. Min. Mod.Tattoo removala Fair Good Mod. Mod. Good Good Mod. Mod.Hypopigmentationa Min. Min. Min. Min. Min. Mild Min. Min.Index of epidermis necrosis – 1 hourb 4 0 3 1 2 0 0 1Sum index of vacuolation – 1 hourb 34 26 41 20 0 30 18 16Sum index of aggregated pigment – beforeb 42 21 15 16 42 21 15 16Sum index of aggregated pigment – 1 hourb 23 10 14 11 0 21 12 16Sum index of aggregated pigment – 1 weekb 36 22 2 9 0 0 16 15Sum index of aggregated pigment – 1 monthb 20 3 15 0 18 16 29 21

aDescriptions of the scales are given in the Results section.bMeasured from biopsy slides immediately before, 1 h after, 1 week after and 1 month after laser treatment.

In Vivo Study of Intradermal Focusing for Tattoo Removal 157

increased loss of pulse energy due to the salinebreakdown prompted us to use focal depthslarger than 3 mm and we estimated that theactual pulse energy could be less than half ofthe values cited in Tables 2–5 for treatmentswith focal depth of 2.4 mm and 1.3 mm.

The saline film covering the skin duringablation procedures served the dual purpose ofreducing the index mismatch at the skin sur-face and tissue cooling. The laser treatmentlasted about 30 min for each tattoo strip.Regardless of tattoo colour, treated areas ofthe skin appeared white after removal of thetemplate because of the laser-induced vacuola-tion under the surface. Although no significantimmediate haemorrhage was observed in anyof the test areas, varying degrees of delayedfoci of petechial haemorrhage with little vis-

ible epidermal damage were observed in mosttattoo strips about 30 min post-treatment.Based on gross examination and review ofphotographs of the skin surface, we gradedthe degree of petechiation per treated areaof 3.4 cm2 in each tattoo strip within 1 hour oflaser treatment using a standardised scalebased on the number of foci: (1) none, (2)minimal, <10, (3) moderate, 11–25, (4) signifi-cant, >26 foci. Grading was performed withinone hour of laser treatment based on the com-bined evaluations of two investigators (WAWand XHH) and results, grouped by the tattoocolours, are listed in Tables 2–5. Comparingthe appearance of the four tattoo groups atsimilar ranges of pulse energy and focal depth,the blue and black pigments showed the mostintense petechiation or bruising, green less

Table 4. Response of green tattoos

Tattoo ID 0D 0E 0L 0Mc 9D 9E 9L 9M

Pulse energy (mJ) 63 64 64 49 55 55 55 55Focal depth (mm) 5.9 5.9 4.7 4.7 4.7 5.9 2.4 2.4Petechial haemorrhagea Mod. Mod. Mod. Mod. None Min. Mod. Mod.Tattoo removala Fair Poor Poor Fair Fair Fair Fair Mod.Hypopigmentationa Min. Min. Min. Min. Mild Min. Min. MildIndex of epidermis necrosis – 1 hourb 0 0 0 0 0 0 0 1Sum index of vacuolation – 1 hourb 0 0 0 0 0 0 0 0Sum index of aggregated pigment – beforeb 31 18 14 11 31 18 14 11Sum index of aggregated pigment – 1 hourb 13 0 15 5 15 0 24 17Sum index of aggregated pigment – 1 weekb 27 21 24 3 5 16 10 17Sum index of aggregated pigment – 1 monthb 8 27 10 7 22 15 36 15

aDescriptions of the scales are given in the Results section.bMeasured from biopsy slides immediately before, 1 hour after, 1 week after and 1 month after laser treatment.cLaser ablation without the saline film on skin surface.

Table 5. Response of red tattoos

Tattoo ID 0B 0G 0J 0O 9B 9G 9J 9O

Pulse energy (mJ) 62 63 62 49 55 55 55 52Focal depth (mm) 5.9 5.9 5.9 4.7 3.6 7.1 2.4 2.4Petechial haemorrhagea None Min. Min. Min. None None Min. NoneTattoo removala Poor Fair Poor Fair Good Fair Mod. FairHypopigmentationa Min. Min. Min. Min. Mild Mod. Mild MildIndex of epidermis necrosis – 1 hourb 0 0 0 0 0 0 0 0Sum index of vacuolation – 1 hourb 0 0 0 0 0 0 0 0Sum index of aggregated pigment – beforeb 0 0 0 0 0 0 0 0Sum index of aggregated pigment – 1 hourb 0 0 0 0 0 0 0 0Sum index of aggregated pigment – 1 weekb 0 0 0 0 0 0 0 0Sum index of aggregated pigment – 1 monthb 0 0 0 0 0 0 0 0

aThe descriptions of the scales are given in the Results section.bMeasured from biopsy slides immediately before, 1 hour after, 1 week after and 1 month after laser treatment.

158 X.H. Hu et al.

and red the least. These observations agreewell with the acute histological responses ofthe skin tissues showing epidermal necrosisand laser-induced vacuolation in biopsiestaken 1 hour after the laser treatment and areconsistent both with previous results [3,5] andour own clinical experience in tattoo removalusing commercial Q-switched laser systems.Surprisingly, the close correlation of the acuteresponses with the pigment colour is in strongcontrast with the fact that the absorption coef-ficients of the tattoo pigments dispersed insolutions are nearly the same at 1064 nm (seeTable 1), suggesting that e#ects of presentlyundefined factors other than pigment absorp-tion alone are involved in ns laser pulseremoval of tattoos.

The clinical appearance of each tattoo stripfor tattoo ablation was evaluated by compar-ing the photos of untreated baseline pigmentspots to each treated tattoo strip 1 monthpost-treatment using a subjective scale of: (1)poor, <20%, (2) fair, 20–40%, (3) moderate,40–60%, (4) good, 60–80%, (5) excellent, >80%clearance of pigmentation compared to base-line immediate post-laser control. Gradingscores were assigned based on review andconsensus of two investigators (WAW andXHH) and results, grouped by the tattoo col-ours, are listed in Tables 2–5. Visual evalu-ation of clinical tattoo ablation e$cacy asassessed through these comparisons demon-strated a removal grading for blue pigmenttattoos ranging from good to excellent. Thesame grading for black test sites ranged fromfair to good whereas the degree of removal forred and green pigments was judged less thanfor either blue or black, with green gradingslightly better than red.

A final index used in evaluating the grossvisual appearance of test sites was the degreeof observable hypopigmentation. An exit clini-cal examination of the micropigs was con-ducted 3 months post-treatment and late phasephotos were taken to complete the study.Detailed evaluation of these photographsrevealed no evidence of hypertrophic scars inany of the treated fields at this late time pointregardless of the acute response to treatment.Similarly, we found no incidences of hyperpig-mention in any treated area although thepotential for this in the pig model used in thisstudy has not been defined. Hypopigmentationwas noted at this time point and was gradedas: (1) minimal, >30%, (2) mild, 30–60%, (3)moderate, >60% compared to control. Results

of hypopigmentation scoring were againachieved by consensus of two of the investi-gators, and are also presented in Tables 2–5.Treated fields in one pig with freckling wereobserved to develop the common freckling seenwith maturation of the pigs 3 months after thelaser treatment. The pigmentation seemed tobe in a natural and organised form and crossedover into the zones of treatment in a grosslynormal appearance, indicating the continuedpresence of some normal, physiologically func-tioning melanocytes within the basal layer ofthe epidermis. From these observations weconcluded that the treatment of pigmentedlesions with a focused beam of 12 ns pulse withenergy in the region of 50 mJ may cause amoderate amount of petechiation and/or bruis-ing in the acute response of the skin but iswithin the relative safe limits for protectingthe basal layer near the epidermis–dermisjunction as indicated by 66% of reviewedsamples being rated as minimal, i.e. less than30% hypopigmentation, for this index.

Biopsies taken immediately before lasertreatment, 1 hour, 1 week and 1 month aftertreatment have been quantitatively analysedfor histological parameters by two investi-gators (MJC and GWK) in a double-blindstudy. To evaluate the tissue response andtattoo pigment redistribution according to thedepth in the dermis or the distance from theepidermis–dermis junction, an ocular grid with20�20 squares was used in an optical micro-scope to divide each biopsy slide into an epi-dermal layer and seven dermal layers from theepidermis–dermis junction at 40� magnifi-cation. The squares of the ocular grid in eachlayer correspond to 0.12 mm squares on thebiopsy. Highly localised laser-induced necrosisin the epidermis and vacuolation in the dermiswere observed in biopsies taken at 1 hour afterlaser treatment. A necrosis index was definedto measure the acute damage in the four sub-layers of epidermis (stratum corneum, stratumgranulosum, stratum spinosum and stratumbasale), which is equal to the number of dam-aged sublayers ranging from 0 to 4. Using theocular grid, we further defined an index ofvacuolation based on the number of squarescontaining vacuolations for each dermis layerconsisting of 20 squares. Both epidermal necro-sis and dermal vacuolation were found to dis-appear in biopsies taken 1 week or later afterthe laser treatment. Granules of aggregatedtattoo pigments of sizes on the order of 10 �mwere observed to exist in the upper dermis of

In Vivo Study of Intradermal Focusing for Tattoo Removal 159

biopsies with blue, black and green tattoos. Toquantify the distribution of the aggregatedpigments in the dermis, we defined a pigmentindex as the number of squares containingaggregated pigments for each layer coveringthe dermis at the 40� magnification. Thevacuolation and pigment indices of the sevengrid layers within the dermis were then addedto yield a sum index for each biopsy specimen.We list the necrosis index for the epidermisand sum indices of vacuolation and aggregatedpigments for the dermis from biopsies obtainedat di#erent times relative to the laser treat-ment in Tables 2–5. It should be noted that apigment index of 0 does not exclude the exist-ence of tattoo pigments. For biopsies with redtattoo, the pigment was widely dispersed withsizes less than 5 �m and could only be ident-ified under higher magnification (400�),resulting in these specimens receiving a valueof 0 for pigment index at the 40� parameter.Figure 2 presents two micrographs of blue andred tattoo biopsies taken 1 hour after lasertreatment to show the di#erent forms of tattoopigment aggregation and vacuolations at

magnifications of 100� and 400�. Under the400� magnification, we have observed thecoexistence of macrophages showing evidenceof actively phagocytosing the laser-dispersedtattoo pigment. There is no indication thatthese macrophages migrate any appreciabledistance from the site of laser activity.

Significant correlation was observed fromthe depth dependence of the pigment distri-bution and the vacuolation in the dermis in allbiopsies. Two representative groups of blue,black and green tattoo biopsies are shownin Fig. 3 with pulse energy of 63 mJ and focaldepth of 5.9 mm and in Fig. 4 with 55 mJand 2.4 mm as examples. In biopsies of redand green tattooed tissues no laser-inducedvacuolation was seen, indicating that thelaser-induced breakdown either did not occuror was much weaker than that in blue andblack. This is consistent with the di#erentdegrees of the postprocedural petechial haem-orrhage, caused by either breakdown of orthermal damage to the blood vessels, noted inthe independent clinical examination of theskin surface. Comparing the pigment index of

Fig. 2. Histological micrograph images of biopsies taken 1 hour after laser treatment with the arrow indicating the tattoopigments. (a) The blue tattoo (ID: 0C) (100×), bar=100 �m; (b) the same slide as (a) (400×), bar=25 �m; (c) the red tattoo (ID:9B) (100×); bar=100 �m; (d) the same slide as (c) (400×), bar=25 �m.

160 X.H. Hu et al.

biopsies taken before, 1 hour and 1 month afterthe laser treatment, we found two changes thatmay be responsible for the tattoo removal, i.e.the overall decrease in pigment density in eachlayer and the pigment mobilisation into deeperlayers. It was noted that for blue tattoos thepigment reduction at 1 month after treatmentwas evident in nearly all layers whereas forblack, the corresponding pigment reductionoccurred most prevalently in those layershaving relatively large vacuolation indices.We further noted that black tattoos exhibiteda trend of tattoo reduction in the shallowlayers and an increase in the deeper layers asshown in Fig. 5.

DISCUSSION

In the current approach of treating pigmentedlesions by ns pulses, large pulse energies up toand even exceeding 1 J must be used whichfrequently results in considerable collateral

tissue damage [6]. We proved in this initial invivo study that the pulse energy can be signifi-cantly reduced to between 40 and 60 mJ at1064 nm wavelength through intradermalfocusing. The team hopes to pursue a series ofsubsequent experiments with the ultimate aimbeing the development of a compact laser sys-tem possessing beam scanning capability fortreating di#erent coloured pigments using asingle wavelength and shortened treatmenttimes as an alternative to the existingmethods. We would like to point out that thequestion as to how this type of Q-switchedintradermally focused laser system could beapplied in an e$cient clinical environment isyet to be answered. Likewise, a number ofquestions arising from the optical proper-ties of the dermis and laser interactionswith pigments within the dermis will alsoneed to be examined further. Also to berefined, is the potential of developing aclinically applicable scanning system, whichwould allow the comparison in clinical

Fig. 3. Depth dependence of the aggregated pigment index from four biopsies immediately before, 1 hour and 1 month after thelaser treatment and the vacuolation index 1 hour after the laser treatment with laser pulse energy of 63 mJ and focal depth of5.9 mm: (a) blue tattoo (ID: 0C); (b) black tattoo (ID: 0H); (c) green tattoo (ID: 0E).

In Vivo Study of Intradermal Focusing for Tattoo Removal 161

applications to standard Q-switched laser sys-tems commercially available in dermatologicalclinics.

Because of the strong light scatteringdescribed in pig skin [10], we expected that thedistribution of pulse energy inside porcineskin would become much wider than that ofthe direct beam in the cases of no scattering[13]. Consequently, pulses with energies in theregion of 50 mJ were used in this study toachieve a visible acute response that was muchlarger than the threshold estimated at a fewmillijoules from our previous study on thesurface ablation in isolated pig skin withsimilar beam configurations [9]. Histologicalanalyses of biopsies taken 1 hour after lasertreatment have shown that the epidermis in allbiopsies remained essentially intact, except forhighly localised damage in one strip withblack tattoos. This can be attributed to thecombination of small pulse energy and thesaline film used for index matching and tissuecooling. Due to the high density of ablatedspots per tattoo strip, large laser fluences,

between 200 J/cm2 and 340 J/cm2 are accumu-lated during the half-hour treatment. Theseresults show that collateral tissue damageis primarily determined by the pulse energyif the thermal relaxation of absorbed energyis allowed. This is consistent with pre-vious results on ablating ocular tissues bynanosecond pulses [14].

Comparing the results of acute responses tolaser treatment from gross observation andbiopsy histology on a total of 32 treated tattoos(see Tables 2–5), we observed a strong relationbetween the degree of tattoo removal and theextent of long-lasting vacuolation observablein biopsies at 1 hour after laser treatment fortattoo removal. Previous studies have shownthat the long-lasting vacuolation in soft bio-logical tissues is a result of shock wavesand acoustic transients by the laser-inducedplasma [14,15]. The observations thus suggestthat pigment breakdown and tattoo removalcan be e$ciently achieved through intra-dermal focusing with reduced pulse energy.From the layer dependence of the aggregated

Fig. 4. Depth dependence of the aggregated pigment index from biopsies immediately before, 1 hour and 1 month after the lasertreatment and the vacuolation index 1 hour after the laser treatment with laser pulse energy of 55 mJ and focal depth of 2.4 mm:(a) blue tattoo (ID: 9K); (b) black tattoo (ID: 9I); (c) green tattoo (biopsy ID: 9M).

162 X.H. Hu et al.

pigment indices from biopsies immediatelyprior to and 1 month after laser treatment(Figs 3–5) we also noticed that in the blue andblack tattooed skin, treatment had eitherreduced the amount of pigment granules in allgrid layers and/or redistributed the pigmentgranules deeper into the dermis. This providesdirect microscopic evidences that tattooremoval by short laser pulses is, to a certaindegree, due to pigment mobilisation. Since thepigment redistribution could also be accom-

plished through the thermal damage and cor-responding recovery process in skin, it may beresponsible for tattoo removal in some cases ofgreen and red tattoos.

Two models have been proposed to explainthe soft tissue ablation by ns laser pulses invisible and near-infrared spectral regions: (1)the selective-photothermolysis model for skinand pigment ablation [1]: and (2) the plasmaablation model for ocular tissue ablation inwhich tissue absorption is negligible [16,17].The fundamental di#erence between the twomodels lies in whether the ablation process iscaused by a localised plasma induced by thestrong electromagnetic field of ns pulses.Although the tissue absorption dominatesthe laser–tissue interaction described by thephotothermolysis model, its role in the existingplasma ablation model remains unknown[16,17]. In a previous study we reported experi-mental results on the surface ablation of por-cine skin using ns pulses at 1064, 532, 266 and213 nm, in which the optical absorption coef-ficient of the skin increases by two orders ofmagnitude as the laser wavelength decreases,and we found that the ablation data canonly be satisfactorily explained by a plasma-mediated ablation model [9]. The theoreticalmodelling of the above experimental results, tobe published elsewhere, suggested to us thatthe initial seed electrons are generated intissue chromophores through a thermal ionis-ation pathway and energised by the strongelectric field of the ns pulse to initiate subse-quent avalanche ionisation that leads to theformation of plasma. Therefore, the ablationthresholds in these cases of substantial lightabsorption are determined mainly by the gen-eration of initial seed electrons and thus thelocal temperature reached in tissue chromo-phores. Since the local temperature dependson both the chromophore absorption cross-section and the packing configuration of thechromophores, we expect that the ablationthresholds of di#erent tattoo pigments dependon the form of pigment aggregates in additionto the pigment absorption. Based on the abovediscussion it is plausible to attribute the leastresponse of the red tattoo pigments to the nslaser pulses to the very dispersed distributionof the red pigments which facilitate the di#u-sion of the localised thermal energy in thesepigments and consequently increase the corre-sponding ablation threshold. It seems clearthat better comprehension of these complexprocesses will require further quantitative

Fig. 5. Distribution of aggregated black pigments (ID: 9H) inbiopsies taken (a) before, (b) 1 hour after, and (c) 1 monthafter the laser treatment (100×), bar=100 �m.

In Vivo Study of Intradermal Focusing for Tattoo Removal 163

studies to achieve a clear understanding ofthe fundamental mechanism underlying thetreatment of pigmented lesions by ns pulses.

In summary, we have carried out an initialin vivo study of tattoo removal in micropigs by12 ns pulses in a converging beam of 11.4� coneangle at 1064 nm wavelength. We achievedsignificant tattoo pigment mobilisation for theblue and black pigments with a nominal rangeof pulse energy from 38 to 63 mJ and found astrong relation between the degree of tattooremoval and laser-induced vacuolation in thedermis. Furthermore, we established that thevariation between optical absorption of tattoopigment was too small to be responsible for thelarge di#erence in the responses between theblue/black and green/red tattoos. These resultsare very encouraging for possible developmentof a new generation of compact Q-switchedlasers for dermatological and plastic surgerywith high e$ciency diode-laser pumping andautomated scanning delivery system. We arecurrently preparing to study the use of a con-verging beam with a larger cone angle up to30� which is expected to further reduce thepulse energy, collateral tissue damage and,more importantly, to achieve better responsesfrom the green and red tattoos by exceedingtheir respective breakdown thresholds withoutcausing breakdown in the saline film atsmaller focal depths.

ACKNOWLEDGEMENTS

The authors thank Kenneth Jacobs for constructing thetranslation table and the focusing cup for this project andPat Vore, RN, CPRSN for providing tattooing expertise.

REFERENCES

1. Anderson RR, Parrish JA. Selective photothermolysis:precise microsurgery by selective absorption of pulsedradiation. Science 1983;220:524–7.

2. Anderson RR, Margolis MJ, Watanable S, Flotte T,Hruza GJ, Dover JS. Selective photothermolysis ofcutaneous pigmentation by Q-switched Nd:YAG laserpulses at 1064, 532, and 355 nm. J Invest Dermatol1989;93:28–32.

3. Taylor CR, Anderson RR, Gange RW, Michaud NA,Flotte TJ. Light and electron microscopy analysis oftattoos treated by Q-switched ruby laser. J InvestDermatol 1991;97:131–6.

4. Fitzpatrick RE, Goldman MP, Ruiz-Esparza J. Use ofthe alexandrite laser (755nm, 100ns) for tattoo pigmentremoval in an animal model. J Am Acad Dermatol1993;28:745–50.

5. Zelickson BD, Mehregan DA, Zarrin AA, Coles C,Hartwig P, Olson S et al. Clinical, histologic, andultrastructural evaluation of tattoos treated withthree laser systems. Lasers Surg Med 1994;15:364–72.

6. Kilmer SL: Laser treatment of tattoos. Dermatol Clin1997;15:409–17.

7. Chan HH, Lam LK, Wong DSY, Leung RSC, Ying SY,Lai CF et al. Nevus of Ota: a new classification basedon the response to laser treatment. Lasers Surg Med2001;28:267–72.

8. Reid WH, Miller ID, Murphy MJ, Paul JP, Evans JH.Q-switched ruby laser treatment of tattoos; a 9-yearexperience. Br J Plas Surg 1990;43:663–9.

9. Hu XH, Fang Q, Cariveau M, Pan X, Kalmus GW:Mechanism study of porcine skin ablation by nano-second laser pulses at 1064, 532, 266 and 213 nm. JQuant Electron 2001;37:322–8.

10. Du Y, Hu XH, Cariveau M, Ma X, Kalmus GW, Lu JQ.Optical properties of porcine skin dermis between900 nm and 1500 nm. Phys Med Biol 2001;46:167–81.

11. Hu XH. E$cient use of Q-switched lasers in thetreatment of cutaneous lesions. SPIE Proc 1995;2395:586–91.

12. Nahen K. and Vogel A: Plasma formation in water bypicosecond and nanosecond Nd:YAG laser pulses –Part II: transmission, scattering, and reflection. IEEEJ Sel Top Quantum Electron 1996;2:861–71.

13. Lu JQ, Hu XH, Dong K: Modeling of the rough-interface e#ect on a converging light beampropagating in a skin tissue phantom. Appl Opt 2000;39:5890–7.

14. Vogel A, Schwelger P, Frieser A, Asiyo MN,Birrngruber R. Intraocular Nd:YAG laser surgery:light-tissue interaction, damage range, and reductionof the collateral e#ects. IEEE J Quantum Electron1990;26:2240–60.

15. Juhasz T, Hu XH, Turi L, Bor Z: Dynamics of shockwaves and cavitation generated by picosecond laserpulses in corneal tissue and water. Lasers Surg Med1994;15:91–8.

16. Stern D, Schoenlein RW, Puliafito CA, Dobi ET,Biringruber R, Fujimoto JG. Corneal ablation by nano-second, picosecond, and femtosecond lasers at 532 and625 nm. Arch Ophthalmol 1989;107:587–92.

17. Kennedy PK: A first-order model for computation oflaser-induced breakdown thresholds in ocular andaqueous media. 1. Theory. IEEE J Quantum Electron1995;31:2241–9.

Paper received 1 November 2001;accepted after revision 14 December 2001

164 X.H. Hu et al.