What are the results of photodynamic therapy for Barrett's esophagus?
B.F. Overholt, M. Panjehpour (Knoxville)
Barrett's esophagus (BE) is characterized by replacement of normal esophageal mucosa with metaplastic columnar epithelium [1, 2]. BE is common, occurring in an estimated 10-13% of patients with chronic gastroesophageal reflux disease [3-5] but its frequency may be 20 times higher . The mucosal change is felt to represent the result of the body's defensive mechanism to protect the esophagus against repeated acid or bile reflux. BE is associated with an increased occurrence of mucosal dysplasia and adenocarcinoma in the specialized glandular mucosa [3, 4, 7-12] with a 30-52 fold increase in the occurrence of esophageal cancer compared to the normal population '4, 13]. Surgical resection is recommended for selected cases with severe dysplasia and for adenocarcinoma. However, esophagectomy is associated with a significant morbidity and mortality [14-17]. Photodynamic therapy (PDT) offers an alternative non-surgical therapy that eliminates dysplasia and superficial cancer and reduces Barrett's mucosa while reducing risks and costs [18-21]. We describe herein our results using PDT in 55 patients with BE and dysplasia and/or superficial cancer.
All patients referred for PDT had repeated endoscopic examinations and biopsies demonstrating BE and dysplasia. Eighteen patients had 20 superficial cancers less than 1.5 cm in size. In patients with endoscopic abnormalities and biopsy proven adenocarcinomas, only those with lesions that were interpreted Tis-1, N-0, M-04 by endoscopic ultrasound (EUS) and computed tomography were included in the study. However, one patient with a T-2 lesion was treated since he was not a surgical candidate. The study was approved by the Institutional Review Board of the Thompson Cancer Survival Center.
Following thorough informed consent, porfimer sodium 2 mg/kg (Photofrin® Quadra Logics Technologies, Inc., Vancouver, BC, Canada) was injected intravenously. Forty-eight hours later 630 nm light was endoscopically delivered using a cylindrical diffuser or a windowed centering esophageal balloon [22-24]. Forty-two (42) patients received 1 PDT treatment (1-2 light exposures), 10 patients received two treatments and 3 patients received three treatments to eliminate the dysplasia and/or superficial cancer.
An argon-pumped dye-laser (Lambda Plus, Coherent, Palo Alto, CA) tuned to a wavelength of 630 nm was used to provide endoscopic delivery of the red light to the esophagus. The light was focused into a 200 micron extension fiber. A 400 micron 1.5 - 2.5 cm cylindrical diffuser, was attached to the extension fiber. The diffuser was inserted through the endoscope into the esophageal lumen to deliver light intraluminally to the targeted area. Power density was 400 mW/cm of diffuser providing energy density of 100-250 J/cm from the diffuser to tissue depending on whether the initial or follow-up treatment was being delivered.
A 3 or 5 cm windowed, 25 mm diameter esophageal centering balloon was used as the sole method to deliver the light to esophageal mucosa in 15 patients and in conjunction with the diffuser in 25 patients. Fifteen patients were treated with the diffuser alone. The deflated balloon was passed into the esophagus, endoscopically positioned and inflated. More recent balloons incorporate a standard guidewire system, thereby greatly improving the ease and safety of insertion (Figure 1). A small endoscope (Fujinon EVG-FP; Wayne, NJ) was passed alongside the shaft of the centering balloon to verify and monitor the position of the balloon during PDT. A rigid 2.5 cm diffuser was used with the 3 cm balloon and a flexible 5 cm diffuser (QLT PhotoTherapeutics, Vancouver, B.C., Canada) with the 5 cm centering balloon was passed into the centering channel of the respective balloon. Light dose varied from 150 J/cm with the 5 cm balloon up to 300 J/cm with the 3 cm balloon.
All patients were sedated with meperidine and midazolam during the procedure. Oxygen was administered by nasal catheter at 2-3 l/min. Atropine sulfate 0.4 mg and glucagon 1 mg were given intravenously if secretions or esophageal motility were excessive.
EUS was performed with the Olympus MB-3 instrument (Olympus Corporation of America, New Hyde Park, NY). Endoscopies were performed with the Fujinon EVG-CT and EVG-FP (Wayne, New Jersey) and Olympus GIF-100 and GIF-1T100 endoscopes. In all patients, careful endoscopic measurements from the dental margin were obtained to precisely determine the location of the squamocolumnar junction, gastric folds and the area targeted for PDT delivery.
A follow-up endoscopy was performed at 48 hours to determine if additional light treatment was required for small areas of esophageal mucosa that were not significantly damaged from the initial light treatment. Subsequent endoscopies were performed at 1 week, 6 and 12 months. Mucosal staining with 5-8 cc 2% Lugol's solution  was used in some patients at 6 months post-treatment to separate squamous epithelium from small and sometimes visually indistinguishable minute patches of Barrett's mucosa. Biopsies (usually 4 quadrant, "jumbo" biopsies every 2 cm) of healed mucosal areas that demonstrated squamous epithelium were evaluated for underlying glandular mucosa. In
28 patients a Nd:YAG laser with contact probe (CL-60, Surgical Lasers Technologies, Oaks, PA) was used to destroy small areas of residual Barrett's mucosa that were identified visually or by using Lugol's staining in PDT treated segments at 3-6 months. In several recent patients, small untreated areas of Barrett's mucosa were treated with the Nd:YAG laser during the 1 week endoscopic follow-up.
Figure 1. 3, 5 and 7 cm windowed esophageal centering balloons.
All patients were treated as outpatients. However, one heart transplant patient was hospitalized for cardiac monitoring and two patients early in the study were hospitalized 48 hours after PDT due to development of atrial fibrillation. An additional patient was hospitalized for treatment of a large pleural effusion. Outpatients received intermittent intravenous fluids for 5 days following PDT. All patients were treated with omeprazole, 20 mg twice a day for the first three months and then once a day. Hydrocodone bitartrate elixir and fentanyl transderm-25 were used for control of chest discomfort post PDT and prochlorperazine suppositories 25 mg were used for control of post procedure nausea. Patients were advised to avoid direct exposure of skin to sunlight for 4 weeks following injection of Photofrin® and then to gradually expose their skin to sunlight.
With the diffuser (non-balloon) system, extensive injury to exposed mucosa was observed along with destruction of dysplastic epithelium and small cancers. Treated areas demonstrated a combination of moderate to marked edema, erythema, small erosions and exudate overlying damaged mucosa. However, there were occasional small areas of mucosa that were not damaged, presumably due in part to the effects of esophageal motility, cardiac and respiratory movement on the position of the probe and from shadowing due to prominent esophageal folds around the diffuser probe in the non-distended esophagus.
The centering balloon system provided more uniform circumferential light delivery to esophageal mucosa and eliminated irregular light delivery and the "hill and valley" effect due to the shadowing by esophageal folds in the nondistended esophagus. Nonetheless, small untreated areas were occasionally seen following balloon PDT requiring a second light exposure using the diffuser system.
In all patients, full repair of the damaged mucosa required approximately 3 months. Healing was associated with mucosal repair or with the appearance of granulation tissue in areas where the mucosa had been totally destroyed. Epitheliation occurred next. Sites of previous Barrett's mucosa were replaced completely or partially with squamous epithelium beginning at 2-3 months. Lugol's staining was used to identify patches of residual Barrett's mucosa from squamous epithelium in treated areas as sometimes it was difficult to determine what was squamous with erythema versus residual Barrett's. Multiple biopsies of squamous areas, including "jumbo" biopsies in most, demonstrated no underlying glandular mucosa in 43 patients. However, in 2 patients who healed with squamous epithelium, esophageal biopsies revealed minimal foci of sub-squamous Barrett's non-metaplastic epithelium.
Complete elimination of Barrett's mucosa was found in 16 patients following PDT. Esophageal biopsies in these patients demonstrated no Barrett's epithelium. Reduction of the amount of the surface area of Barrett's mucosa was noted in all cases. In other patients, an estimated 75-80% of treated mucosa was replaced with squamous epithelium. In addition, in the 29 patients who received PDT beginning at the squamocolumnar junction and extending distally, the squamocolumnar junction migrated distally an average distance of 4.8 cm (range 0-15.5 cm) including 4 patients with less than 3 cm of Barrett's mucosa prior to PDT. In patients receiving PDT targeted to early cancers only, the treated areas were replaced entirely with squamous epithelium.
Following PDT, 42 of the 55 (76%) patients had conversion of dysplastic/malignant Barrett's mucosa to normal squamous or Barrett's mucosa without dysplasia (Table I). Of these, 33 required one PDT treatment (1-2 light exposures per treatment), 6 required
2 treatments and 3 patients required 3 treatments.
Thirteen (13) patients persisted with dysplasia. Of these, 3 patients with Tis cancers and 6 with HG dysplasia converted to LG. Three (3) with HG persisted unchanged although HG was histologically reclassified as focal. These patients received PDT only once. Eight (8) with high grade dysplasia converted to low grade (LG) dysplasia. One with a T-2 cancer converted to LG and 1 LG persisted following PDT. Of those with persistent dysplasia, 11 had received 1 PDT treatment and 2 had 2 treatments. All 13 patients with persistent dysplasia are scheduled for retreatment with PDT or focal Nd:YAG laser therapy to small areas of residual Barrett"s.
Full dose (300 J/cm) PDT treatment produced moderate chest pain and dysphagia lasting for approximately 5-7 days following the last light treatment but this gradually improved. Esophageal strictures developed in 29 (53%) patients. Four (4) were severe requiring multiple dilations but all eventually responded to dilation. Two (2) patients developed atrial fibrillation requiring hospitalization. Return to sinus rhythm was accomplished within hours of initiation of medical therapy in both patients. No myocardial injury was documented. Small unilateral or bilateral pleural effusions were found 48 hours following PDT in the majority of patients who underwent post-procedure chest X-rays. One patient required thoracentesis.
A minimally invasive treatment resulting in the reduction of Barrett's mucosa and the ablation of dysplasia and superficial cancers is desirable. Certainly, an alternative to esophagectomy as a treatment modality is needed. Although combined chemotherapy and radiation or trimodality therapy with radiation, hyperthermia and chemotherapy have been reported to produce promising results in some cases with stage I esophageal cancer [26-29], no information exists about the possible use of these modalities for early adenocarcinoma or for high-grade dysplasia in BE.
In an attempt to reduce the extent of Barrett's mucosa and therefore the risk of cancer, investigators have used medical [30-34] or surgical therapy [15, 17, 25, 35-37] but the results have been controversial  and have not consistently reduced either dysplasia or the extent of Barrett's mucosa. Others have used argon or Nd:YAG laser thermal ablation of Barrett's mucosa followed by omeprazole therapy for acid suppression with generally favorable, although limited, results [21-24, 39].
Photodynamic therapy (PDT) combined with acid suppression as a treatment that reduced Barrett's mucosa and eliminated dysplasia and superficial cancer was first described in 1993 . This report expands that experience to 55 patients followed for 6-66 months following PDT. All patients were maintained on long-term omeprazole to achieve acid suppression in order to allow mucosal repair in an anacid environment. Extensive mucosal ablation was observed after PDT. Follow-up endoscopic findings and biopsies demonstrated a reduction in the extent of Barrett's mucosa in all patients with replacement of an estimated 75-80% of the treated mucosa by squamous epithelium. In patients treated from the squamocolumnar junction distally including patients with short segments of Barrett's mucosa, the replacement with squamous epithelium was associated with a relocation of the squamocolumar junction distally an average of 4.8 cm.
All patients underwent multiple biopsies of areas with squamous epithelium regrowth following PDT. Two demonstrated minute fragments of non-metaplastic Barrett's mucosa but the remaining 43 patients demonstrated squamous mucosa. It is important to note that Lugol's staining should be used to identify squamous mucosa for purposes of biopsies done to determine the presence or absence of subsquamous Barrett's mucosa or the presence of small, sometimes visually indistinct, islands of residual Barrett's following PDT. Lugol's staining will also identify large and small patches of residual Barrett's mucosa that can be targeted for thermal or repeat PDT destruction.
Sixteen patients have had complete endoscopic and biopsy disappearance of Barrett's mucosa following PDT. Nd:YAG laser ablation was used during the later phase of this study in 8 of these 16 patients to destroy small residual islands of Barrettt's mucosa. Squamous epithelial regrowth over these islands was noted and biopsies confirmed the absence of Barrett's mucosa. Based on these preliminary findings, it is our belief that PDT alone or in combination with thermal ablation can eliminate Barrett's mucosa in many patients with Barrett's esophagus. This approach is the subject of current investigations.
Dysplasia was eliminated in 42 of the 55 patients. Of the 13 patients with persistent dysplasia, 4 quadrant biopsies every 2 cm of the treated areas documented in 3 patients with Tis cancers and in 6 with HG dysplasia conversion to LG dysplasia. Three (3) with HG dysplasia persisted unchanged although HG was histologically reclassified as focal. These 3 patients received PDT only once. Eight (8) patients with high grade (HG) converted to low grade (LG) dysplasia following PDT. One patient with a T-2 cancer converted to LG and 1 LG persisted following PDT. Of those with persistent dysplasia, 11 had received 1 PDT treatment and 2 had 2 treatments. Retreatment with PDT or focal Nd:YAG laser therapy is planned for all 13 patients.
Eighteen patients were treated for superficial cancers. All superficial cancers were destroyed with PDT. Follow-up (mean = 22 months, range 7-62 months) revealed no recurrence. One patient developed a second primary superficial cancer in a previously untreated site and underwent a successful second course of PDT. Another patient developed a subsquamous, 2x4 mm adenocarcinoma in the center of a 5 cm site treated 6 months earlier for HG dysplasia. The site demonstrated complete overlying squamous mucosa. The patient was retreated with PDT and no tumor recurrence has been detected over an 20 month follow-up period. Our assumption is that the cancer existed at the time of PDT and was not adequately destroyed during the initial therapy. With the initial PDT, we noted complete mucosal destruction followed by complete squamous epitheliation at 6 months. Believing that the tumor was probably present during the initial treatment, we are concerned that even with deep PDT induced mucosal damage, tumor can persist. Hence the need for follow-up of treated sites for at least 1 year. We are also concerned that new techniques of ablation of non-dysplastic Barrett's mucosa, such as thermal ablation, may produce even less mucosal damage, thereby creating a potential for persistence of Barrett's mucosa or the occurrence of dysplasia or cancer in mucosa that was treated with less damaging technologies.
Follow-up of these patients is important. As noted above, one of our patients developed a sub-squamous adenocarcinoma that was detected 6 months after PDT in a treated area that showed complete squamous epitheliation. We also have noted that dysplasia and in 1 case an early carcinoma has developed in untreated areas in 8 patients between 4 months and 3 years following initial PDT therapy. These patients have subsequently been treated with another course of PDT or with Nd:YAG laser therapy. Clearly, in Barrett's patients with dysplasia or early carcinoma, the goal should go beyond elimination of the initial disease and should ultimately be the ablation of all Barrett's mucosa.
Complications following PDT included esophageal strictures in 29 patients. It appears that overlapping of treatment margins created greater tissue injury resulting in stricture formation. Typically, several dilations were needed for return of swallowing to normal but in 4 patients strictures were severe, requiring multiple dilations. This is not unexpected due to the deep injury required to abolish HG dysplasia and superficial cancer.
Improvements in light and drug dosimetry appear to reduce the incidence of strictures. A preliminary review of our current studies using 5 and 7 cm windowed centering balloons reveals an incidence of only mild stricture formation in < 20%. Longer balloons reduce the need for overlapping of treated areas, appear to reduce the occurrence of stricture formation and, in addition, dramatically reduce treatment times,
Two patients developed atrial fibrillation requiring hospitalization and medical therapy for conversion to sinus rhythm. Both patients remain in regular rhythm. In a sequential series of 12 consecutive patients undergoing esophageal PDT for Barrett's esophagus and dysplasia, post procedure evaluation of cardiac enzymes and electrocardiograms demonstrated no myocardial enzyme abnormalities or permanent electrocardiographic abnormalities .
We noted pleural effusions in most patients who underwent chest X-rays 48 hours post- PDT. An additional 14 patients were studied sequentially with X-rays 48 hours post-PDT. Six demonstrated small bilateral pleural effusions, 4 had left pleural effusions and 1 had a small effusion on the right. Three patients had normal post PDT X-rays. All patients were asymptomatic relative to pulmonary symptoms. Small, asymptomatic pleural effusions, are common following PDT for BE .
Photosensitivity was a minor problem for most patients but 3, in spite of instructions to avoid direct sunlight, developed significant cutaneous erythema and edema after sun exposure 1 week following PDT. Another patient developed severe cutaneous blisters over his ankles after 4 hours of sun exposure 2 months following PDT. He had not exposed this area of his skin to sunlight previously.
Others have used PDT to treat BE. Laukka et al.  used "low dose" PDT for the treatment of BE in 5 patients. They described a mean reduction of 2.4 cm (range 1-5) in the length of the Barrett's segment in patients treated with "low dose PDT" and maintained for 6 months on omeprazole 20 mg daily. The photosensitizer used in their work was 1.5 mg/kg of hematoporphyrin derivative followed 48 hours later with photoradiation with 630 nm light using a 1.0-1.5 cm cylindrical diffuser to produce a light dose of 150-200 J/cm.
Other investigators have used thermal ablation of Barrett's mucosa with argon or Nd:YAG laser followed by omeprazole therapy for acid suppression with generally favorable, albeit limited, results. Brandt and Kauvar  also treated one patient with Barrett's mucosa with Nd-YAG laser thermal ablation and achieved transient regression of the Barrett's mucosa. The patient was maintained on omeprazole but after 14 weeks the specialized columnar mucosa recurred. In a subsequent communication suppression of acid after additional Nd:YAG therapy resulted in elimination of Barrett's mucosa . Sampliner et al.  reported a favorable result of treatment with Nd-YAG laser thermal ablation of Barrett's mucosa in one patient. Biopsies 11 months later demonstrated only squamous epithelium.
Berenson et al.  used argon laser thermal photoablation of Barrett's mucosa in 10 patients who were maintained on omeprazole 40 mg per day for purposes of acid inhibition. Patients underwent 3-12 endoscopies and had 1-8 locations ablated.
Most ablated areas were retreated one to six times. They felt squamous epitheliation following laser ablation occurred both by spread from contiguous squamous epithelium and de novo from progenitor cells within the glandular mucosa. They emphasized the importance of abolishing the glandular mucosa. Berenson further indicated that acid suppression is essential to allow restoration of squamous epithelium after mucosal ablation.
Benefits of PDT for the treatment of Barrett's dysplasia include a minimally invasive technique for ablation of dysplastic mucosa and, in some cases, superficical cancers, fewer therapeutic endoscopic sessions compared to endoscopic thermal ablative techniques, lower costs compared to surgery, and reduced morbidity and mortality when compared to short-term outcomes of surgical resection patients. Surgical costs are 1.5-4.5 times more than PDT and require a minimum of 6 weeks recovery time compared to 2-3 weeks for PDT patients (Table II). Morbidity is also less following PDT. Although the incidence of strictures in this series is high, all patients have been able to be dilated successfully. It is noted that mild strictures are found in up to 64% of surgical patients . Furthermore, our recent work with longer balloons indicates a dramatic reduction in the incidence and severity of post-PDT strictures (< 20%) presumably due to improved delivery of light energy to mucosa and less overlapping of treated areas. Most importantly, we have had no mortality in our PDT patients whereas the mortality from surgery ranges from 6-14% [14, 17, 25, 36]. We believe that patient outcomes when defined as reduced recovery times, lower morbidity, lower mortality and lower costs are improved by using PDT as the treatment for Barrett's dysplasia and/or superficial carcinoma.
The application of photodynamic therapy for endoscopic therapy for patients with BE has proved effective in the elimination of dysplastic mucosal and early, superficial cancers and in some patients, complete elimination of Barrett's epithelium. Improvements in delivery of the activating light to mucosa including centering balloons and longer optical diffusers have greatly reduced the incidence of post-PDT strictures and the time required for endoscopic therapy. Further improvements in the balloon system will occur and for example could provide a 180-270° window for precise delivery of PDT to a focal area of abnormal mucosa. Lower cost diode lasers currently in development should reduce the procedural cost and improve the use and portability of PDT. Likewise, 2nd and 3rd generation photosensitizers drugs activated by different wavelengths of light should allow more tissue selectivity and specificity and in addition much reduced photosensitivity (e.g. 1-7 days vs 4-6 weeks).
With greater selectivity of drug uptake into abnormal mucosa, laser induced fluorescence spectroscopy [46, 47] could target abnormal mucosa followed by delivery of light for photochemical or photothermal ablation - the "smart laser." Photosensitizer drug-induced fluorescence differences between normal and diseased mucosa could allow the smart laser to treat areas of abnormal mucosa with the appropriate wavelength to destroy the diseased mucosa. This could also occur without the drug by using laser-induced tissue autofluorescence and a Nd:YAG laser for thermal ablation.
Existing and future PDT drugs can and will be activated by different wavelengths of light. For example, sodium porfimer is currently used with red light (630 nm) to capitalize on the deeper tissue penetration of red light. However, the drug can also be activated by green light (340 nm). Green light has very shallow penetration and is absorbed almost completely by hemoglobin and mucosa. Theoretically, green light could be used to destroy mucosa without producing significant submucosal damage  thereby reducing the incidence of stricture formation. Other photosensitizers such as etiopurpurins, benzoporphyrins, purpurins, phthalocyanines and chlorin derivatives are being studied for use in PDT [49, 50]. These drugs are activated by longer wavelengths of light which penetrate and damage tissue more deeply than red light. They will have to be evaluated carefully for esophageal PDT due to the potential of esophageal strictures associated with deeper tissue injury. Aminolevulinic acid (ALA) also holds promise for treating Barrett's mucosa since it can be administered orally and is associated with only mucosal injury . A word of caution is needed however. High grade dysplasia is not infrequently associated with intramucosal carcinoma. If the green light or ALA injury did not produce destruction of mucosa deep enough to totally ablate the abnormal mucosa, one would predict the persistence of subsquamous Barrett's mucosa and the persistence of carcinomas. More investigational work is needed in this area but the concept is appealing since injury limited to the mucosa would likely be associated with less toxicity and stricture formation.
PDT will become a standard "first line" therapeutic option for treatment of dysplastic epithelium and superficial mucosal cancers in the gastrointestinal track and, in fact, an all organ systems accessible by activating light. PDT in combination with other therapeutic techniques including thermal or ultrasound ablation, chemotherapy or radiation therapy will further expand our capabilities to effectively treat and cure patients with precancerous tissue or early cancers. The use of the power of light in the diagnosis and treatment of human disease is beginning.
In conclusion, in a group of 55 patients with Barrett's esophagus and dysplasia, we have reported the ablation of 20 superficial cancers in 18 of these patients, the elimination of dysplasia in 42 of the 55 patients, the complete elimination of Barrett's mucosa in 16 patients and the reduction of the extent of Barrett's mucosa in all patients who were treated with PDT and were maintained on long term omeprazole. Repeated PDT sessions were required in some patients to accomplish elimination of dysplasia. Esophageal strictures occurred in 53% of patients but were treated satisfactorily with dilation. Current results using 5 and 7 cm centering balloons has shown a dramatic reduction in the incidence of stricture formation. Mucosal injury with these balloons is less than with diffusers or shorter balloons that require overlapping of treated areas. We are carefully following the patients treated with longer PDT balloons to evaluate the long term effects on dysplasia. Lugol's staining is an important technique to identify residual patches of Barrett's mucosa following PDT. Small residual patches of Barrett's mucosa can be successfully destroyed with Nd:YAG laser therapy. Our results indicate that PDT alone or in combination with thermal ablation can eliminate superficial cancers, dysplasia and Barrett's mucosa in many patients with BE.
The authors acknowledge the significant contributions to this work provided by Lee Whisman, Joe DeCosta, F. Paul Buckley, III and Donna Edwards. Photofrin® and laser fibers were provided by Quadra Logics Technologies, Inc., Vancouver, BC, Canada.
16. Reid BJ, Weinstein WM, Lewin KH, et al. Endoscopic biopsy can detect high-grade dysplasia or early adenocarcinoma in Barrett's esophagus without grossly recognizable neoplastic lesions. Gastroenterology 1988;94:81-90.
18. Overholt BF, Panjehpour M. Barrett's esophagus: photodynamic therapy for ablation of dysplasia, reduction of specialized mucosa and treatment of superficial esophageal cancer. Gastrointest Endosc 1995;42:64-70.
19. Overholt BF, Panjehpour M. Photodynamic therapy in Barrett's esophagus: reduction of specialized mucosa, ablation of dysplasia and treatment of superficial esophageal cancer. Semin Surg Oncol 1995;11:372-376.
29. Seitz JF, Biovannini M, Padaut-Cesana J, et al. Inoperable nonmetastatic squamous cell carcinoma of the esophagus managed by concomitant chemotherapy (5-fluorouracil and cisplatin) and radiation therapy. Cancer 1990;66:214-219.
51. Gossner L, Sroka R, Stolte M, et al. Photodynamic therapy of superficial squamous cell cancer and severe dysplasia of the esophagus: preliminary results using 5-aminolevulinic acid. Gastrointest Endosc 1996;43:A336.