Ann Thorac Surg 1999;68:1763-1769
© 1999 The Society of Thoracic Surgeons
Original Articles
Photodynamic therapy for esophageal lesions: selectivity depends on wavelength, power, and light dose
Jolanda van den Boogert, MDa,
Hugo J. van Staveren, PhDa,
Ron W.F. de Bruin, PhDa,
Jan H.R. Eikelaara,
Peter D. Siersema, PhDa,
Richard van Hillegersberg, PhDa
a Laboratory for Experimental Surgery, Erasmus University, Rotterdam, The Netherlands
Address reprint requests to Dr van den Boogert, Laboratory for Experimental Surgery, Erasmus University, PO Box 1738, 3000 DR Rotterdam, The Netherlands
e-mail: vandenboogert{at}heel.fgg.eur.nl
Presented at the Thirty-fifth Annual Meeting of The Society of Thoracic Surgeons, San Antonio, TX, January 25–27, 1999.
 |
Abstract
|
|---|
Background. Photodynamic therapy with 5-aminolevulinic acid–induced photosensitization could selectively eliminate esophageal epithelial lesions. This study aimed at optimizing laser parameters for 5-aminolevulinic acid photodynamic therapy of the normal rat esophagus.
Methods. Sixty rats received 200 mg/kg 5-aminolevulinic acid orally and were illuminated 3 hours later with either 633 or 532 nm light (n = 30 for each group) through an endoesophageal balloon catheter. Rats received either 8.3 or 25 J/cm diffuser, applied with a 33, 100, or 300 mW/cm diffuser. During illumination, tissue fluorescence measurements and light dosimetry were done. Rats were sacrificed at 48 hours after photodynamic therapy.
Results. During illumination, protoporphyrin IX fluorescence declined faster when a higher power output was used. Fluence rate at the esophageal surface was highest for 633-nm light. At 532 nm, light caused less damage to the epithelium and muscle than 633-nm light. Illumination with 33 mW resulted in selective epithelial ablation, whereas illumination with 300 mW caused muscle damage with minor epithelial damage.
Conclusions. The assumed selective epithelial damage of 5-aminolevulinic acid photodynamic therapy in the esophagus largely depends on the combination of wavelength, power, and light dose applied. Most selective epithelial damage was found when low-power 633-nm light was used.
|
Introduction
|
|---|
Since 1970 the incidence of esophageal adenocarcinoma has increased, in Europe and the United States, at a rate greater than that of any other malignancy [1]. Barrett’s epithelial metaplasia is a premalignant condition that causes a 30- to 50-fold greater risk of esophageal adenocarcinoma [2]. The current treatment option for patients with Barrett’s esophagus with high-grade dysplasia (HGD) or esophageal adenocarcinoma is esophageal resection. For Barrett’s esophagus with HGD, however, this treatment is controversial because the mortality and morbidity rates associated with esophageal resection is judged to be too high for premalignant lesions [3]. 5-Aminolevulinic acid (ALA)–induced, protoporphyrin IX (PpIX)-mediated photodynamic therapy (PDT) is an experimental therapy for Barrett’s esophagus. Pharmacokinetic studies showed selective PpIX fluorescence in esophageal mucosa (squamous or Barrett’s) compared with submucosa and muscularis after ALA administration [4, 5]. Wavelength-dependent activation of accumulated PpIX ideally selectively destroys Barrett’s epithelium. In clinical studies, ALA-PDT has been shown to eliminate Barrett’s epithelium followed by re-epithelialization with squamous epithelium [6, 7]. However, incomplete ablation leading to remaining islands of residual Barrett’s epithelium underneath regenerated squamous epithelium has been described [6, 7].
The present study aimed to find illumination parameters for selective and complete epithelial damage by varying the power output and total light dose applied. A pharmakokinetic study showed similar ALA-induced photosensitization in squamous and Barrett’s epithelium, so the normal rat esophagus was chosen to compare lesions in a standardized and reproducible model [4]. Because PDT damage to the underlying muscle layer or nerve tissue has been found to result in esophageal dilatation, we compared 633-nm light with 532-nm light, which penetrates tissue less deeply [8, 9]. In addition, fluorescence was measured during illumination, and light dosimetry was done.
|
Material and methods
|
|---|
Animals
The experimental protocol was approved by The Committee on Animal Research of Erasmus University. Seventy male Wistar rats (Harlan CPB, Zeist, The Netherlands), weighing 250 to 300 g were used. They had free access to tap water and food (AM II, Hope Farms, Woerden, The Netherlands). Photodynamic therapy was done using intramuscular ketamine and xylazin anesthesia.
Experimental design
The animals were randomly allocated to two groups of 35 animals each. In each group illumination was done with either 633-nm or 532-nm light. In both groups, 30 animals received 200 mg/kg 5-aminolevulinic acid (ALA) (Sigma Chemical Company, St. Louis, MO) dissolved in phosphate buffered saline (PBS) to a volume of 1 mL in one oral gavage. In both groups, animals were allocated to six groups of 5 animals each. Illumination was done at 3 hours after ALA administration, with 8.3 or 25 J/cm diffuser applied with a 33, 100, or 300 mW/cm diffuser. Control animals (n = 5 in both 633-nm and 532-nm light groups) received PBS orally and were illuminated 3 hours thereafter with 300 mW and 25 J. Animals were sacrificed at 48 hours after PDT.
Photodynamic therapy treatment conditions
To deliver homogenous and circumferential light to the esophagus, a custom-made optically clear double-lumen balloon catheter (PTA Balloon Catheter Opta 5; Cordis, Roden, The Netherlands) was used (length, 20 mm; diameter, 3.5 mm when inflated with 0.4 mL air), with the lumen for the laser fiber exactly in the center [8]. Light was transmitted through a 400-µm fiber with a 10-mm-long cylindrical diffusing tip 760 µm in diameter (Lightstic 360; Rare Earth Medical Inc, West Yarmouth, MA). When the balloon catheter and diffuser were used as described, the illuminated area was 1.1 cm2. Fluence and fluence rates in the present study are expressed per centimeter diffuser (and thus per 1.1 cm2 surface area). Illumination with 633-nm light was done with a dye laser, Model 630 (600 Series Dye Module; Laserscope, San Jose, CA) pumped by a KTP/532 surgical laser (Laserscope, San Jose, CA).
The application of 532-nm light was done with the KTP/532 laser. With a linear diffuser, the built-in power meter of the KTP laser could not be used, and so the power emitted was measured with an integrating sphere (Optometer Model 370; Graseby Optronics, Orlando, FL). Each desired power output could be achieved by using neutral density filters. A spectroscope (WaveMate; Coherent, Auburn, CA) was used to verify the accuracy of the wavelength in both groups.
Fluorescence and dosimetry
True light fluence and PpIX spectra were monitored during PDT at 15-second intervals using a spherical isotropic probe, as described in detail previously [8, 10]. The mean autofluorescence measured in control animals was subtracted from the recorded spectra, and the remaining PpIX fluorescence peak (at 705 nm using 633-nm light and at 635 nm using 532-nm light) was integrated over a 30-nm range. Values are given relative to the start value, which was set to 1.
Histology
Animals were sacrificed at 48 hours after PDT. After inspection of the thoracic and abdominal cavities, the esophagus and a small piece of stomach was dissected and opened longitudinally. The width at the laser site was measured and the esophageal appearance noted. The esophagus was curled up from distal to proximal, fixed in formalin, sectioned and stained with hematoxylin and eosin for conventional light microscopy. Damage was scored semiquantitatively on a scale from 0 to 3 for each separate esophageal layer by two investigators masked to the treatment (Table 1). Additionally, a selectivity factor for epithelial damage was introduced and defined as the epithelial damage score divided by the muscular damage score (muscular damage score = [muscular necrosis + muscular inflammation]/2).
Statistical analysis
Fluorescence intensity, true fluence rates, and damage scores are expressed as mean ± standard error of the mean. Comparisons were made using Student t test. A difference was considered to be significant at p values less than 0.05.
|
Results
|
|---|
All animals survived until sacrifice. No abnormalities were seen in the thoracic cavity at autopsy. In the 633-nm light group at all doses with 100 mW and the two higher doses with 300 mW, a white spot (10 to 25 mm2) was apparent on the cranial surface of the liver next to the esophagus. In animals treated with 33-mW 633-nm light, a smaller (1 to 5 mm2) white spot was apparent on the liver in 7 of 10 animals (2 in the 8.3 J and 5 in the 25 J group). In controls and animals treated with 532-nm light, no macroscopic liver abnormalities were observed.
Fluorescence
When PpIX fluorescence was plotted versus illumination time, fluorescence in the 633-nm group declined faster when a higher power output was used (Fig 1A). When fluorescence was plotted versus the applied energy, fluorescence was increasingly bleached with decreasing output power (Fig 1B). When 33-, 100-, or 300-mW, 633-nm light was used, PpIX fluorescence after 8.3 J was not fully bleached, because fluorescence after 25 J was significantly lower than after 8.3 J (p = 0.01, p = 0.04, and p = 0.002, respectively). The inversed fluorescence versus the applied energy shows a linear relationship (Fig 1C) as was found previously by Robinson and associates [11]. The slope of the straight lines is a measure of the bleaching rate (with respect to the applied energy). The slopes of the 33-, 100-, and 300-mW lines differed significantly (p < 0.02) and were 0.106 ± 0.005, 0.083 ± 0.005, and 0.046 ± 0.005 J-1, respectively.
View larger version (24K):
[in this window]
[in a new window]
|
Fig 1. Mean integrated protoporphyrin IX (PpIX) fluorescence at 705 nm during 633-nm illumination plotted against illumination time (A) and applied energy (B). The error bars indicate standard errors of the mean. In panel C, the reciprocal of the normalized protoporphyrin IX fluorescence is plotted as a function of light dose. The lines are least-square best fits of the data. Energy is applied with 33 (dotted line), 100 (solid line), or 300 mW (dashed line).
|
|
In the 532-nm groups, PpIX fluorescence (versus illumination time) declined fastest with 300-mW output (data not shown). The fluorescence after 25 J was significantly lower than after 8.3 J for the 100- and 300-mW groups (p = 0.03 and p = 0.04, respectively). Fluorescence intensity after long illumination (25 J) with 33 mW did not differ from the intensity after short illumination (8.3 J) (p = 0.11). Again, for a certain energy applied, the fluorescence is bleached most for the lowest output power.
Dosimetry
In the 633-nm light group, during optimal irradiation the measured, true, fluence rate (scattered plus nonscattered light) at the surface was 2.4 ± 0.05 times higher than the calculated irradiance (nonscattered light), with a minimum of 1.9 and a maximum of 3.0. In the 532-nm light group the true fluence rate was 1.3 ± 0.05 times higher (less than using 633-nm light, p < 0.001) than the irradiance, with a minimum of 1.0 and a maximum of 2.0.
Histology after 633-nm light illumination
No histologic abnormalities were found in the control group. Most damage to the epithelium was observed in animals treated with 33 mW, although the difference was not significant compared with animals treated with 100 mW, 25 J (Table 2). When 33 mW and a total light dose of 25 J were used, complete epithelial ablation was achieved in 4 of 5 animals. In 1 animal, one epithelial cell layer remained. With 33 mW and a total light dose of 8.3 J, complete epithelial ablation was achieved in 3 of 5 animals. Damage to the submucosa was comparable in all groups; only illumination with 300 mW and 8.3 J caused less damage. Damage to the muscularis propria was more severe in the 25-J groups than in the 8.3-J groups (p 
0.01). Given a fixed total light dose, lowering the power from 300 mW to 33 mW resulted in more severe epithelial damage than muscular damage and thus in a higher selectivity factor (for 8.3 J, 100 versus 33 mW, p = 0.02; 300 versus 33 mW, p = 0.04; for 25 J, p = 0.049 and p = 0.04, respectively). Most selective epithelial damage was found after application of 8.3 J with 33 mW (selectivity factor = 3.7 ± 0.7, p 0.04 versus all other 633-nm illumination schemes), and most complete epithelial damage after application of 25 J with 33 mW (selectivity factor = 1.1 ± 0.1) (Fig 2). Other treatment regimens showed less damage to the epithelium than to the muscularis propria (selectivity factor < 1).
View this table:
[in this window]
[in a new window]
|
Table 2. Histopathologic Changes of the Esophageal Wall 48 Hours After 5-Aminolevulinic Acid Photodynamic Therapy in Different Groups Treated With 633-nm Lighta
|
|
View larger version (124K):
[in this window]
[in a new window]
|
Fig 2. Histologic characteristics of a rat esophagus at the photodynamic therapy site after 633-nm illumination with 33 mW and 8.3 J (A) and 25 J (B). The arrow in panel A indicates the beginning of the illuminated area (from arrow to the right), showing very selective and complete loss of epithelium, whereas the muscle layer was not damaged. Long illumination (B) caused, besides loss of epithelium, severe damage to the muscularis propria. (E = epithelium; K = keratin layer; LP = lamina propria; M = muscularis propria; MM = muscularis mucosae; SM = submucosa.) (Hematoxylin and eosin, original magnification x 40.)
|
|
Histology after 532-nm light illumination
No histologic abnormalities were found in the control group. A light dose of 8.3 J did not cause any epithelial or muscular damage and caused only slight edema and inflammation of the submucosa (Table 3). Only in the groups treated with 25 J applied with 33 or 100 mW, was complete epithelial ablation found in 1 and 2 animals, respectively. Most selective epithelial damage was achieved with 33 mW (selectivity factor = 2.0 ± 0.7).
View this table:
[in this window]
[in a new window]
|
Table 3. Histopathologic Changes of the Esophageal Wall at 48 Hours After 5-Aminolevulinic Acid Photodynamic Therapy in Different Groups Treated With 532-nm Lighta
|
|
Histology after 633-nm versus 532-nm light illumination
Comparing 633- with 532-nm treatment regimens with the same output power and energy, 532-nm light caused less damage to the epithelium and muscularis propria than 633-nm light (Fig 3). Analysis of the different 532-nm and 633-nm groups showed that illumination with 33 mW (8.3 J in the 633-nm group and 25 J in the 532-nm group) caused the most selective epithelial damage.
View larger version (150K):
[in this window]
[in a new window]
|
Fig 3. Histopathologic section of a rat esophagus at the laser site after 633 nm (A) and 532 nm (B) light irradiation with 100 mW and 8.3 J. The 633-nm illumination caused considerably more damage than 532-nm illumination. Magnification as in Figure 2. Abbreviations as in Figure 2.
|
|
|
Comment
|
|---|
Our results indicate that the selectivity of ALA-PDT for inducing esophageal epithelial damage depends largely on the combination of wavelength, output power, and light dose. Most selective epithelial damage was achieved with 33 mW illumination, whereas illumination with 300 mW caused damage to the underlying muscularis propria with less damage to the epithelium. Very selective damage to the epithelium was found after using 633-nm light illumination with 33 mW and 8.3 J.
For improving the results of PDT, most attention has been paid to the development of a photosensitizer that accumulates selectively in (pre)malignant lesions. 5-Aminolevulinic acid has been a great step forward for PDT of superficial (epithelial) lesions. Until recently, less attention has been given to the influence of illumination parameters. Incomplete destruction of Barrett’s epithelium after ALA-PDT in patients and the proved importance of the time interval between ALA administration and illumination raises the question of whether adjusting the illumination protocol can further improve the results [6–8].
Because green light (532 nm) penetrates tissue less deeply than red light (633 nm), we expected to find more superficial and thus more selective damage to the epithelium using 532-nm light [12]. Indeed, we found no macroscopic liver damage in the green light groups. In general however, comparison of the damage between the 633- and 532-nm groups with the same power and energy delivered showed that 532-nm light caused less epithelial damage than 633-nm light. We choose 532-nm light because PpIX extinction coefficients and the ratio of photon energies are approximately equal at 532 nm and 633 nm. Equivalent rates of photo absorption are therefore achieved with approximately equal energy fluence rates at 532 nm and 633 nm [12]. However, the optical absorption and scattering of tissues also vary at different wavelengths. The in vivo dosimetric results showed that for a given output power the in situ measured fluence rate for 532-nm light is lower than that for 633-nm light (fluence rate multiplication factor ß = measured fluence rate/calculated irradiance, ß532 = 1.3 and ß633 = 2.4). Dosimetry in this study suggests a matching true light dose for 532 and 633 nm illumination with a two times higher output power or a two times longer illumination for 532-nm compared with 633-nm light. Based on the results of a pilot study, we had chosen a factor of three in varying the incident light dose. When the damage in the 633-nm light group was compared with the damage in the 532-nm light group with a comparable true light dose (three times longer illumination or a three times higher power output), there were no differences in epithelial or muscular damage or in selectivity factor.
Studies using hematoporphyrin derivatives as a photosensitizer and 514-nm versus 630-nm light drew different conclusions, because some used equivalent rates of photon absorption [9, 13–15], whereas others did not account for the differences in optical properties of tissue at these wavelengths [12]. The former studies concluded that 514-nm light has equal or more effect and toxicity than 630-nm light [9, 13–15]. The latter studies concluded that 514-nm light had less effect than 630-nm light [12].
In the present study, equal doses of incident light led to the finding that 532-nm light is less effective than 633-nm light because of the lower true in vivo light doses. In clinical ALA-PDT practice, the assumption that the illumination time can be shortened by using 514-nm instead of 633-nm light to achieve equal damage is based on the higher PpIX excitation coefficient at 514 nm but does not take into account the lower true in vivo light dose at 514 nm. The present results can be extrapolated to indicate that the incident light doses must be approximately equal for 514-nm and 633-nm ALA-PDT to achieve equal damage at the surface. Our results further show the importance of monitoring both the true light dose and the fluorescence in a clinical setting, as both the fluence rate (due to the different light scattering properties of tissue) and concentration of the photosensitizer vary among patients.
Results of PDT could be improved by varying the power output for a given total light dose. In both the 532- and 633-nm groups, most selective epithelial damage was achieved using low output power. This finding is consistent with two other studies of ALA-PDT on normal mouse skin and transplanted rat tumors [11, 16]. Similar results were found using hematoporphyrin derivatives in tumors in mice [17–19]. This is probably due to the fact that for achieving an equal light dose, the duration of the illumination increases with decreasing power. Rapid PpIX photobleaching and photochemical oxygen consumption during high irradiances can lead to relative hypoxia and reduced PDT effect. Long illumination implies high oxygen supply (oxygen saturation of blood x vascular flow x time). Thus, there seems to be a critical combination of total light dose (low enough to cause minor muscular damage) and illumination time (long enough to supply oxygen needed for epithelial damage) to achieve selective epithelial ablation.
Although ALA-induced photosensitization occurs almost exclusively in the esophageal mucosa, almost all animals in the present study had damage to the submucosa or muscularis, which suggests that the photodynamic threshold to induce damage to the submucosa and muscle is lower than that to the epithelium [4, 5, 8]. As previously described by us, a difference in vascular supply, sensitivity to singlet oxygen radicals, or PpIX metabolism might explain this observation [8].
Besides the selectivity of the epithelial damage, the completeness of the epithelial ablation is important. In the present study the optimal illumination scheme was between 8.3 and 25 J applied with 33 mW of 633-nm light. One animal in the 25 J group did not have complete epithelial ablation. A second treatment therefore could ensure complete elimination of the Barrett’s epithelium in patients.
In summary, we demonstrated the importance of light dosimetry, fluorescence measurements, and, moreover, the choice of the illumination protocol for ALA-PDT of esophageal lesions. Using equal doses of incident light, illumination with 532-nm light was less effective than 633-nm light because of the lower true light dose. The application of low-power 633-nm light caused selective epithelial damage when a low light dose was given and caused complete epithelial damage when a high light dose was given. With these optimized illumination parameters, efficient and safe esophageal ALA-PDT can be performed.
|
Acknowledgments
|
|---|
This study was supported by a grant from The Netherlands Digestive Disease Foundation (WS 95-10). The authors thank Mrs Coby J. F. Peekstok for histologic staining.
|
References
|
|---|
-
Blot W.J., Devesa S.S., Kneller R.W., Fraumeni J.F., Jr Rising incidence of adenocarcinoma of the esophagus and gastric cardia. JAMA 1991;65:1287-1289.
-
Cameron A.J., Ott B.J., Payne W.S. The incidence of adenocarcinoma in columnar-lined (Barrett’s) esophagus. N Engl J Med 1985;313:857-859.[Abstract]
-
Wang K.K. Barrett’s esophagus. Compr Ther 1994;20:36-43.[Medline]
-
Van den Boogert J., Houtsmuller A.B., De Rooij F.W.M., De Bruin R.W.F., Siersema P.D., Van Hillegersberg R. Kinetics, localization and mechanism of 5-aminolevulinic acid-induced porphyrin accumulation in normal and Barrett’s-like rat esophagus. Laser Surg Med 1999;24:3-13.
-
Mäkinen K., Grönlund-Pakkanen S., Tiirikainen M., Nuutinen P., Kuusisto A., Alhava E. Protoporphyrin-IX distribution and photodynamic effect in rat esophagus after aminolaevulinic acid administration. Scand J Gastroenterol 1997;32:633-637.[Medline]
-
Laukka M.A., Wang K.K. Initial results using low-dose photodynamic therapy in the treatment of Barrett’s esophagus. Gastrointest Endosc 1995;42:59-63.[Medline]
-
Barr H., Sheperd N.A., Roberts D.J.H., Tan W.C., Krasner N. Eradication of high-grade dysplasia in columnar-lined (Barrett’s) oesophagus by photodynamic therapy with endogenously generated protoporphyrin IX. Lancet 1996;348:584-585.[Medline]
-
Van den Boogert J., Van Hillegersberg R., Van Staveren H.J., et al. Timing of illumination is essential for effective and safe photodynamic therapy. Br J Cancer 1999;79:825-830.[Medline]
-
Van Gemert J.C., Berenbaum M.C., Gijsbers G.H.M. Wavelength and light-dose dependence in tumour phototherapy with haematoporphyrin derivative. Br J Cancer 1985;52:43-49.[Medline]
-
Van Staveren H.J., Marijnissen J.P.A., Aalders M.C.G., Star W.M. Construction, quality assurance and calibration of spherical isotropic fibre optic light diffusers. Lasers Med Sci 1995;10:137-147.
-
Robinson D.J., De Bruijn H.S., Van der Veen N., Stringer M.R., Brown S.B., Star W.M. Fluorescence photo bleaching of ALA-induced protoporphyrin IX during photodynamic therapy of normal hairless mouse skin. Photochem Photobiol 1998;67:140-149.[Medline]
-
Foster T.H., Gibson S.L., Raubertas R.F. Response of photofrin-sensitised mesothelioma xenografts to photodynamic therapy with 514 nm light. Br J Cancer 1996;73:933-936.[Medline]
-
Bellnier D.A., Prout G.R., Lin C.W. Effect of 514 nm argon ion laser radiation on hematoporpyrin derivative-treated bladder tumor cells in vitro and in vivo. J Natl Cancer Inst 1985;74:617-625.
-
Nseyo U.O., Whalen R.K., Lundahl S.L. Canine bladder response to red and green light whole bladder photodynamic therapy. Invest Urol 1993;41:392-396.
-
DeLaney T.F., Sindelar W.F., Tochner Z., et al. Phase 1 study of debulking surgery and photodynamic therapy for disseminated intraperitoneal tumors. Int J Radiat Oncol Biol Phys 1993;25:445-457.[Medline]
-
Hua Z.X., Gibson S.L., Foster T.H., Hilf R. Effectiveness of delta-aminolevulinic acid-induced protoporphyrin as a photosensitizer for photodynamic therapy in-vivo. Cancer Res 1995;55:1731-1732.
-
Gibson S.L., Foster T.H., Feins R.H., Raubertas R.F., Fallon M.A., Hilf R. Effects of photodynamic therapy on xenografts of human mesothelioma and rat mammary carcinoma in nude mice. Br J Cancer 1994;69:473-481.[Medline]
-
Pe M.B., Ikeda H., Inokuchi T. Tumour destruction and proliferation kinetics following periodic, low power light, haematoporphyrin oligomers mediated photodynamic therapy in the mouse tongue. Eur J Cancer B Oral Oncol 1994;30B:174-178.
-
Sitnik T.M., Henderson B.W. The effect of fluence rate on tumor and normal tissue responses to photodynamic therapy. Photochem Photobiol 1998;67:462-466.[Medline]
Top
Abstract
Introduction
