Proc. Nati. Acad. Sci. USA
Vol. 83, pp. 7064-7068, September 1986
Selective ablation of atheromas using a flashlamp-excited dye laser
at 465 nm
MARTIN R. PRINCE*t, THOMAS F. DEUTSCH*, ASCHER H. SHAPIROt, RANDALL J. MARGOLIS*,
ALLAN R. OSEROFF*, JOHN T. FALLON4, JOHN A. PARRISH*, AND R. Rox ANDERSON*
*Wellman Laboratory and the tDepartment of Pathology, Massachusetts General Hospital, Harvard Medical School, Boston, MA 02114;
and the tDepartment of Mechanical Engineering, Massachusetts Institute of Technology, Cambridge, MA 02139
Contributed by Ascher H. Shapiro, June 9, 1986
ABSTRACT Ablation of human atheromas with laser methanol/distilled water (1:1, vol/vol). This laser generated
pulses that had only a small effect on normal artery tissue was 1-,u sec pulses (full width at half-maximum) at a repetition rate
shown in vitro in air and under saline using 1-psec pulses at 465 of 1 Hz with energies up to 1.5 J per pulse. Laser wavelength,
nm from a flashlamp-excited dye laser. At this wavelength, measured with a grating monochromator (Bausch and Lomb
there is preferential absorption in atheromas due to caroten- 33-86-76), ranged from 459 to 470 nm with the maximum at
oids. The threshold fluence for ablation was 6.8 ± 2.0 J/cm2 465 nm. Pulse width was measured with a reverse-biased
for atheromas and 15.9 ± 2.2 J/cm2 for normal aorta tissue. silicon photodiode (E.G. & G., Salem, MA, FND 100Q). The
At a fluence of 18 J/cm2 per pulse, the ablated mass per unit laser mirrors were adjusted prior to each experiment until
of energy ranged from 161 to 370 ,ug/J for atheromas and from output burn patterns on exposed polaroid film were circular
50 to 74 1Lg/J for normal aorta tissue. Ablation products and uniform. Spot size was determined by measuring burn
consisted of cholesterol crystals, shredded collagen fibers, and patterns on polaroid film placed just in front of specimens.
small bits of calcific material. Most debris was less than 100Mum The energy per pulse was measured with an energy meter
in diameter, but a few pieces were as large as 300 Mum. (Scientech, Boulder, CO, 36-5002 and 38-0101). Pulse-to-
High-speed photography of ablation in air suggested explosive pulse energy variation was ±5%.
ejection of debris, caused by vapor formation, at speeds on the Specimens. Specimens of atherosclerotic human aortas
scale of 300 m/sec. Histological analysis showed minimal with mostly hard or soft raised, yellow plaques were obtained
thermal damage to residual tissue. These data indicate that from the morgue and used within 5 days of death. Lesions
selective laser ablation of atheromas is possible in vitro. that were hard and white or entirely covered with pink fibrous
tissue were not included in the study. All specimens were
The ability to deliver laser energy to arteries via fine optical rinsed with normal saline to wash off blood and postmortem
fibers offers the possibility of unclogging obstructed arteries clot, tightly wrapped in plastic and stored refrigerated but
without the trauma of surgery or general anaesthesia (1). This allowed to equilibrate to room temperature just before
procedure, called laser endarterectomy or laser angioplasty, irradiation. Aortas from a total of 16 cadavers were used: 7
is now being evaluated in early clinical trials (2, 3). Its use for measuring ablation thresholds, 3 for histological analysis,
thus far, however, has been limited, apparently due to perfo- 4 for measuring ablated mass, and 2 for high-speed photog-
rations, aneurysms, and other forms of inadvertent laser dam- raphy.
age, to adjacent and underlying normal tissues (2-6). Ablation Thresholds. To measure the threshold fluence
Angioscopic (7) and radiographic visualization or other tech- (incident energy per unit area) for ablation, the laser beam
niques (8) may help in aiming laser radiation at plaque. Laser was made uniform by focusing into a step-index, multimode,
endarterectomy would be much safer, however, ifthe radiation quartz optical fiber with a 1-mm core, approximately 1 m long
preferentially ablated atheroma and thrombus. In the ideal case, (Math Associates, Westbury, NY, QSF 1000). The output
laser light would be delivered uniformly inside an obstructed end of this fiber was placed flush against aortic specimens
artery, and only the obstructing material would be ablated. submerged in saline. The laser energy was increased at
Preferential ablation of atheromas has previously been ac- increments of 5-10 mJ until laser pulses appeared to roughen
the tissue surface. This energy was then measured by aiming
complished by first staining atheromas with a dye and then the fiber into the energy meter and recorded as the threshold
ablating with laser radiation that is absorbed by the dye (9, 10). energy to initiate tissue removal. Threshold fluence was
A simpler method of achieving preferential ablation might be to calculated as this energy divided by the area of the 1-mm
take advantage of existing, endogenous, preferential absorption diameter fiber core. Threshold fluence determinations were
in the atheromas. The waveband from 450 to 500 nm has been repeated for five normal and five atheromatous spots on each
identified as having a 2-fold preferential absorption in of the seven aortas for a total of 70 measurements.
atheromas due to carotenoid pigments (11). In this study we Histology of Preferential Ablation. To examine the selective
have characterized preferential ablation of atheromas, in vitro, ablative effect of this laser histologically, the laser output was
using a flashlamp-excited dye laser emitting radiation in this focused with a 20-cm focal-length quartz convex lens to 2-mm
waveband at an appropriate pulse width and fluence. diameter spots on 20 specimens from three cadavers.
Atheromatous and adjacent normal regions were irradiated
METHODS with multiple pulses at a rate of 1 Hz and a fluence of 18 J/cm2
Radiation Source. Laser radiation at 465 nm was provided per pulse either in air or submerged under 1 cm of saline.
by a flashlamp-excited dye laser (Candela SLL 500 Coax) Following irradiation, the irradiated areas were marked with
with 1.5 x 10-4 M coumarin 445 (Exciton C445) in india ink, fixed in formalin, processed routinely, and stained
with hematoxylin and eosin.
The publication costs of this article were defrayed in part by page charge Measuring Mass Ablated. Specimens were mounted verti-
payment. This article must therefore be hereby marked "advertisement" cally at the edge of an analytic balance (Mettler, Hightstown,
in accordance with 18 U.S.C. §1734 solely to indicate this fact. NJ, AE163) such that ablation debris landed off the balance.
Medical Sciences: Prince et al. Proc. Natl. Acad. Sci. USA 83 (1986) 7065
A humidified chamber surrounding the specimen minimized
interference from air currents and weight loss from evapo- 0
ration. Specimen mass was recorded automatically at 2 Hz by
a computer (IBM PC) interfaced to the balance (Mettler 20
Interface Option 011). All specimens were cut from human
aortas using a 6-mm biopsy punch so that they all had the ell
same surface area for water evaporation. A 2.5-mm spot on E
each specimen was irradiated with 10 pulses, each at 0.9 J (18 09
J/cm2 per pulse) delivered at 1-sec intervals. This was 0
repeated for 7-10 normal and atheromatous aortic specimens To
from each of four aortas for a total of 66 measurements. 0
Balance readings were also taken before and after laser 4)
irradiation to correct for drift due to water evaporation. IL A
High-Speed Photography. The ablation process was pho- IL 0 A
tographed using a 1-tusec electronic flash pulse and a 35-mm A
camera aimed perpendicular to the laser beam, with the I'
shutter open. A yellow filter (Corning 3-370) excluded the
465-nm laser radiation. The room was darkened sufficiently 1 2 3 A
to allow exposure of the film (Kodak Plus'X) only by the
electronic flash. A delay circuit permitted triggering the flash II
at times ranging from 10 to 2500 ,usec, as measured with a 1 2 3 4 5 6 7 Mean
photodiode (Centronics, Mountainside, NJ, BPX65), after Aorta
the laser pulse. Specimens from two atheromatous aortas
were photographed during ablation at 25 J/cm2 per pulse over FIG. 1. Scatter plot of the ablation threshold fluence for each
2.5-mm spots. aorta. Normal human aorta (circles) required 2-3 times more energy
Analysis of Ablation Products. Ablation debris was collect- per unit area for ablation than atheromas (triangles). Bars on the solid
ed by placing a microscope slide in the path of the laser symbols for the means indicate ± 1 SD.
radiation, 2 cm from the specimen. The debris on the slide
was examined while fresh, and again after being fixed and the threshold fluence for each individual was higher for
stained with hematoxylin and eosin. Debris was also collect- normal than for atheromatous tissue. The average threshold
ed by placing a saline-filled Petri dish just in front of the for atheromas was 6.8 2.0 J/cm2 (mean SD) in the seven
specimen (oriented vertically) being ablated; it was collected aortas studied (Table 1), while for normal regions on the same
on 5-aum filter paper (Millipore SM 5.0 am), fixed in 95% aortas the average threshold was 15.9 ± 2.2 J/cm2.
ethanol, and stained with hematoxylin, orange/green, and Fig. 2 is a cross section of two atheromatous and two
eosin to evaluate particle size. adjacent normal regions of human aorta given equal expo-
sures to the 465-nm laser radiation at a fluence of 18 J/cm2 per
RESULTS pulse, which is well above the threshold for ablating
atheromas but near the threshold for ablating normal aorta.
For each specimen there was a threshold fluence above Although there is substantial removal of atheroma to form
which ablation occurred and below which the tissue was craters greater than 1 mm deep, the normal tissue shows only
grossly unaffected. Near the threshold, the ablative effect a slightly roughened surface. This selective ablation occurred
was slight roughening of the surface, but at 2-3 times the both in air and under saline with radiation direct from the
threshold value, craters were formed having the same diam- laser or delivered via a quartz optical fiber. All soft, yellow
eter as the laser beam and up to several hundred micrometers atheromas as well as red, brown/black, or green thrombus
deep. When the fiber was in contact with tissue, laser pulses were readily ablated at this fluence. Pipe stem-like yellow,
at or above the threshold fluence were always associated with calcific plaques were also ablated in saline, whereas in air
a "recoil" force that could be felt while holding the fiber and they were usually ablated to a depth of only about 0.5 mm,
that increased with increasing fluence. With calcific lesions, exposing bright, white material that could not be ablated.
laser pulses above threshold were associated with a loud Normal tissue irradiated at 18 J/cm2 per pulse was slightly
popping noise and a flash of white light, like an electrical roughened and in some regions turned slightly brown at the
spark, at the site of irradiation that was easily seen through surface. No gross charring was seen on any of the specimens.
goggles that excluded the laser radiation. Fig. 1 shows that Fig. 3 shows a typical ablation crater made by 10 pulses of
Table 1. Summary of selective ablation
Value (mean ± SD)
Atheroma Normal aorta Significance* Ratio
Absorption at 470 nmt 54 ± 12 26 ± 4 2.4 x 10-5 2.2
Threshold fluencet 6.8 ± 2.0 15.9 ± 2.2 1.0 x 10-5 2.6
Aorta 1 161 ± 37 74 ± 11 8.4 x 10-6 2.2
Aorta 2 196 ± 98 53 ± 9 3.5 x 10-4 3.7
Aorta 3 370 ± 126 73 ± 17 4.5 x 10-5 5.0
Aorta 4 288 ± 122 50 ± 20 1.3 x 10-5 5.7
Average 253 62 4.0
*P value calculated with Student's t test.
7066 Medical Sciences: Prince et al. Proc. Natl. Acad. Sci. USA 83--(1986)
tissue irradiated under saline often could not be 4ininguished
from adjacent unirradiated normal aorta.
Measurements of the mass ablated by 10 pulses-of 465-nm
radiation at 18 J/cm2 per pulse showed that for the four aortas
studied, every atheroma was ablated more than every spec-
imen of normal tissue. Fig. 4 shows how mass varies with
time during ablation for atheromatous and normal aorta from
the aorta with greatest preferential ablation. Ablation rates
(see Table 1) ranged from 161 to 370 pug/J for atheromas and
from 50 to 74 ;kg/J for normal aorta. The atheromas also had
well-defined ablation craters following irradiation; normal
tissue was occasionally slightly browned but never ablated
more than a fraction of a millimeter. In many atheroma
specimens the atheroma was completely removed down to
FIG. 2. Cross sectional view of four sites on human aorta the underlying normal tissue, where the ablation process
irradiated with 15 pulses of 465-nm laser radiation at 18 J/cm2 per appeared to stop.
pulse and 1-,usec pulse width. The two craters in the atheroma (Right)
are 1-1.5 mm deep, while the two irradiated areas of normal aorta High-speed flash photographs of the ablation process in
show only roughening of the surface with minimal ablation. The Fig. 5 show that 10 Asec after exposure to a single pulse of
black material is india ink used to mark the irradiated sites; no char laser radiation, a mist of fine particles emanates from the
was present. The scale is in mm. ablation site at a velocity of at least 300 m/sec. At 150 Msec,
some larger droplets and a large plume are visible. At 550
465-nm radiation at 18 J/cm2 per pulse. The edges are ,usec, the plume is gone, but there is a jet-like stream of
irregular but show only traces of coagulation (the black fluid-like material. At 2.5 msec (data not shown), debris is
material is the india ink used to mark the site of irradiation). still apparent in front of the specimen. This pattern was
Specimens ablated in air showed a zone of coagulation about consistently photographed on all specimens of atheroma
100 ,m thick lining the ablation crater; this was also present studied.
in some regions that were not irradiated, but seen only Ablation debris, shown in Fig. 6, consisted of whole and
sporadically in specimens ablated under saline. Histology on fragmented cholesterol crystals, small pieces of calcific
older specimens (greater than 24 hr postmortem) showed material, and fibers that appear shredded and occasionally
some autolysis in the media but no other differences from the hypereosinophilic. No smooth muscle or elastin was appar-
fresher specimens. When specimens were irradiated with ent. The particles collected on the filter paper were mostly
20-50 pulses, ablation virtually stopped at the media or under 100 ;km in maximum diameter with a few as large as 300
fibrous/sclerotic material underlying the more fatty or grum- ,um.
ous part of the plaque. Irradiation of normal aorta in air with
10 pulses at 18 j/cm2 per pulse produced ablation up to a DISCUSSION
maximum depth of 200 ,um with the formation of irregular
crater edges and a similar zone of thermal damage. Normal This study shows the feasibility of selectively ablating
atheromas with laser radiation that minimally affects normal
tissue, in vitro. The 1-,usec pulses of 465-nm radiation readily
ablated both fibrofatty and yellow calcific lesions in air and
under saline. The 2 to 6 times preferential mass ablation
measured for atheromas was significant (P < 0.0001) and
probably represents an underestimate of the differential
volume removed because fatty plaque is less dense than
normal aorta. Cross-sections of irradiated sites showed only
Er ,-2 -
U -4 - ASE
FIG. 3. Crater in atheromatous aorta made by ablation with 10 FIG. 4. Quantification of mass ablated by 465-nm laser radiation.
pulses of 465-nm laser radiation at 18 J/cm2 per pulse and 1-gsec Irradiation with 10 pulses in 10 sec at 18 J/cm2 per pulse and 1-,usec
pulse width. The crater edges are irregular but show little evidence pulse width ablates 3.3 mg of atheroma but only 0.66 mg of normal
of thermal injury. The black material is india ink, and the debris in aorta tissue. The mass loss before and after irradiation is due to water
the crater is an artifact of preparation. (Hematoxylin and eosin stain; evaporation and is compensated for by extrapolating the slope of the
bar = 1 mm.) curve, after ablation, back to the initiation of irradiation.
-m-a- Sciences: Prince et al. Proc. Natl. Acad. Sci. USA 83 (1986) 7067
FIG. 5. High-speed flash photographs of atheroma ablation by 1-tMsec pulses of 465-nm laser radiation. (A) 10 Musec; (B) 150 Msec; and (C)
500 ,u sec after irradiation. (Bar = 1 cm.)
slight roughening of the surface of normal aorta with a series may be possible to identify which plaques are most suscep-
of pulses that removed 1-2 mm of atheroma. tible to selective ablation by their color. Systemic carotenoid
The 465-nm wavelength was chosen, because there is content is related to dietary carotenoids and supplemental
preferential absorption in atheromas at that wavelength due (3-carotene in the diet increases carotenoid levels of xantho-
to the presence of carotenoid pigments (11). Since carot- mas (12). Consequently, supplemental carotenoids might
enoids are responsible for the yellow color of atheromas, it enhance selective ablation of atheromas at 465 nm.
Since laser-induced thermal ablation of materials requires
A a threshold of energy absorption, the threshold fluence for
ablation depends on optical absorption (13). Atheroma ab-
sorption at 465 nm is higher than that of normal vascular
tissue; accordingly, the atheroma ablates at a lower fluence.
Preferential ablation was achieved with fluences that were
substantially above threshold for plaque but near threshold
for normal artery. In addition to the optical differences
between atheromatous and normal artery, there are also
differences in thermal and mechanical properties that may
affect ablation thresholds and ablation rates (14). According-
ly, further enhancement of selective ablation might be
achieved by taking advantage of differential thermal and
mechanical phenomena. Presumably ablation proceeds with
least risk at some optimal exposure conditions that are not
necessarily those used for this study. One possibility for
optimizing dosimetry in vivo might be to measure the thresh-
old fluence for each lesion being treated through detection of
the "recoil" force associated with ablation.
Selective ablation may also be affected b.y laser pulse width
(15, 16). Ideally, the pulse width should-be sufficiently short
to prevent excessive diffusion of thermal energy from the
plaque to adjacent normal tissues. The thermal relaxation
B time for a distance comparable to the- characteristic optical
penetration depth at 465 nm in plaque is about 10 msec (17).
The fat-soluble carotenoid chromophores in plaque, howev-
er, are probably concentrated in small fat droplets dispersed
throughout the plaque. The thermal relaxation time for 1- to
20-,m droplets is of the order of 2-800 ,sec. Thus, the 1-,usec
a, pulse width used in these studies prevented significant
thermal diffusion during the period of irradiation. A further
study of ablation at various pulse widths is, however,
Because contours of the ablation craters were irregular,
0 ablation could not be characterized by measuring crater
diameter and depth. Nevertheless, it was apparent that the
.Jr,~~~~r~ plaque was removed to a depth much greater than that of
adjacent normal aorta tissue. The zone of coagulation seen
lining the craters made by ablation in air was probably mostly
due to desiccation rather than thermal effects because it was
present in some areas that were not ablated and was seen only
FIG. 6. Ablation debris. (A) Intact ana shattered cholesterol occasionally on specimens ablated under saline. The tenden-
crystals. (Bar = 100 um.) (B) Collagen fragments and small bits of cy for ablation to slow down at the fibrous/sclerotic base of
calcific material. (Hematoxylin and eosin stain, bar = 200 ,m.) Note the plaque or residual media is consistent with its pink
that only a few collagen fragments have the deep staining quality of (nonyellow) color. This is a useful selective effect particularly
denatured collagen. when the normal tissue underlying a plaque is completely
7068 Medical Sciences: Prince et al. Proc. Natl. Acad. Sci. USA 83 (1986)
obliterated and the sclerotic or fibrotic tissue at the base of many helpful discussions. This work was supported by the Arthur 0.
the plaque represents the final barrier to perforation. and Gullan M. Wellman Foundation, a Heede Fellowship, and Grant
The high speed photographs of Fig. S suggest that the 1 F32 HL07377-01 from the National Institutes of Health.
ablation is an explosion-like process occurring on the order 1. Marcuz, R., Martins, J. R. M., Tupinamba, A., Lopes, E. A.,
of msec and similar to processes reported for other materials Pena, A. F., Cravalho, V. B. & Decourt, L. V. (1980) Arq.
(18). In the early stages (Fig. 5A), the configuration is like that Bras. Cardiol. 34, 9-12.
of a jet of vapor, containing condensed droplets in the form 2. Ginsburg, R., Wexler, L., Mitchell, R. S. & Profitt, D. (1985)
of a fog, escaping from a nozzle at sonic speed and expanding Radiology 156, 619-624.
to a pressure lower than that at the nozzle exit. This would 3. Geschwind, H. J., Boussignac, G., Teisseire, B., Benhaiem,
be consistent with rapid phase changes from liquid to vapor N., Bittoun, R. & Laurent, D. (1984) Br. Heart J. 52, 484-489.
at foci where laser energy is absorbed. In Fig. 5B, what 4. Crea, F., Abela, G. S., Fenech, A., Smith, W., Pepine, C. J. &
appear to be either liquid droplets or torn tissue fragments are Conti, C. R. (1986) Am. J. Cardiol. 57, 171-174.
visible, and in Fig. 5C a high-speed jet-like train of liquid 5. Lee, G., Ikeda, R. M., Chan, M. C., Lee, M. H., Rink, J. L.,
droplets or tissue fragments is seen. This provides strong Reis, R. L., Theis, J. H., Low, R., Bommer, W. J., Kung,
A. H., Hanna, E. S. & Mason, D. T. (1985) J. Am. Coll.
evidence that the ablation is due not only to vaporization but Cardiol. 56, 181-185.
also to liquefaction and mechanical failure accompanied by 6. Gerrity, R. G., Loop, F. D., Golding, L. A. R., Ehrhart, A. &
ejection of bulk material. Vaporization of at least some tissue Argenyi, Z. B. (1983) J. Thorac. Cardiovasc. Surg. 85,
fluid is expected because the energy required for ablation 409-421.
(2700-6200 J/g) is comparable to the latent heat of vapor- 7. Spears, J. R., Marais, H. J., Serur, J., Pomerantzeff, O.,
ization for water (about 2500 J/g). Geyer, R. P., Spizener, R. S., Weintraub, R., Thurer, R.,
Explosive laser ablation of arterial obstructions may carry Paulin, S., Gerstin, R. & Grossman, W. (1983) J. Am. Coll.
some risks. Pressure waves of sufficient amplitude might Cardiol. 1, 1311-1314.
disrupt adjacent, unirradiated normal tissue. Although most 8. Kitrell, C., Willett, R. L., Santos-Pacheo, C., Ratliff, N. B.,
Kramer, J. R., Malk, E. G., Feld, M. S. (1985) Appl. Opt. 24,
of the debris generated was small, some fragments were large 2280-2281.
enough to occlude distal small arteries and arterioles. Not all 9. Abela, G. S., Normann, S., Cohen, D., Feldman, R. L.,
artery obstructions are yellow and thus may not ablate well Geiser, E. A. & Conti, C. R. (1982) J. Am. Coll. Cardiol. 50,
with blue laser pulses. In addition, ablation of some yellow 1199-1205.
plaques may be complicated if they are covered with a 10. Murphy-Chutorian, D., Kosek, J., Mok, W., Quay, S.,
nonyellow fibrous cap. Huestis, W. & Mehigan, J. (1985) Am. J. Cardiol. 55,
In summary, it is possible to ablate human atheromas in 1293-1297.
vitro with laser pulses that spare normal vascular tissue. The 11. Prince, M. R., Deutsch, T. F., Mathews-Roth, M. M.,
selectivity demonstrated might be enhanced by a more Margolis, R., Parrish, J. A. & Oseroff, A. R. (1986) J. Clin.
Invest. 78, 295-301.
rigorous determination of the optimal wavelength, fluence, 12. Blankenhorn, D. H. (1960) Ann. Intern. Med. 53, 944-954.
and pulse width, as well as by pretreating patients with oral 13. Ready, J. F. (1982) Proc. IEEE 70, 533-544.
carotenoids. In vivo application of this method will require 14. Welch, A. J. (1984) IEEE J. Quantum Electron. QE20,
developing a laser catheter that can deliver a relatively 1471-1481.
uniform fluence inside occluded arteries. Evaluation of 15. Deckelbaum, L. I., Isner, J. M., Donaldson, R. F., Laliberte,
selective atheroma ablation will depend on studies of effica- S. M., Clarke, R. H. & Salem, D. N. (1986) J. Am. Coll.
cy, acute complications, and healing after selective laser Cardiol. 7, 898-908.
endarterectomy. 16. Grundfest, W. S., Litvack, I. F., Goldenberg, T., Sherman,
T., Morgenstern, L., Carroll, R., Fishbein, M., Forrester, J.,
The authors thank Harold (Doc) Edgerton for assistance with the Margitan, J., McDermid, S., Pacala, T. J., Rider, D. M. &
high-speed photography; Candela Corp. for loan of a flashlamp- Laudenslager, J. B. (1985) Am. J. Surg. 150, 220-226.
excited dye laser; and Franz Hillenkamp, Myron Wolbarsht, 17. Anderson, R. R. & Parrish, J. A. (1983) Science 220, 524-527.
Rayhaneh Akhavan, Steve Murray, Jay Walsh, and Chris Shea for 18. Gagliano, F. P. & Paek, U. C. (1974) Appl. Opt. 13, 274-279.