Endoscopic fluorescence detection of intraepithelial neoplasia in Barrett's esophagus after oral administration of aminolevulinic acid

BACKGROUND AND STUDY AIMS: Barrett's esophagus is strongly associated with adenocarcinoma. Early malignant transformation of the Barrett's mucosa is often not visible endoscopically and may remain undetected until the invasive adenocarcinoma stage. Endoscopic surveillance is currently carried out on random four-quadrant biopsies at 1-2 cm intervals. Endoscopic fluorescence detection of protoporphyrin IX induced by 5-aminolevulinic acid can identify premalignant lesions. This study evaluates endoscopic fluorescence detection in patients having Barrett's esophagus and compares the results to those of standard endoscopy with random four-quadrant biopsies. PATIENTS AND METHODS: The study included 30 examinations in 28 patients (22 men, 6 women; age range 37-78 years, mean age 60 years,), with five patients having known intraepithelial neoplasia. A dose of 20 mg/kg of 5-aminolevulinic acid was given orally 5 hours before examination. Random four-quadrant biopsies were performed 4-6 weeks before endoscopic fluorescence detection. RESULTS: Of the biopsies taken during the endoscopic fluorescence detection procedure, 28 % (23/81) were true positives. More than one-third of the false-positive results were due to inflammation. None of the 97 control biopsies taken on nonfluorescing areas during endoscopic fluorescence detection were dysplastic. Endoscopic fluorescence detection showed low-grade intraepithelial neoplasia in five patients which was not diagnosed with random four-quadrant biopsies, while random four-quadrant biopsies alone showed three low-grade intraepithelial neoplasias that were invisible during endoscopic fluorescence detection. All high-grade intraepithelial neoplasias or adenocarcinomas (2/2) were detected with both methods. CONCLUSIONS: Fluorescence detection achieved a similar performance when compared with four-quadrant random biopsy, but resulted in fewer biopsies (81 for endoscopic fluorescence detection vs 531 for random four-quadrant biopsies)