• Pathological interpretation of biopsy specimen for invasive cancer, carcinoma in-situ, or high-grade dysplasia [ Time Frame: Within 3 days of endoscopy examination ] [ Designated as safety issue: No ]
  

• Pathological interpretation of biopsy specimen for any grade of dysplasia [ Time Frame: within 3 days of endoscopy examination ] [ Designated as safety issue: No ]
  

This is a prospective blinded controlled tandem endoscopy trial conducted in a regional medical center (Lotung Poh-Ai Hospital, Taiwan). Patients with tissue diagnosis of head and neck cancer are enrolled. Endoscopy screening for esophageal lesions is first performed by using the conventional white light system. Suspicious lesions are recorded. Then the entire esophagus is examined under the NBI system by another endoscopist, who is blinded to the result of the conventional endoscopy. After the endoscopic inspection completed, all suspicious lesions are biopsied. Before each procedure, which one of the two participating endoscopists applies the conventional or NBI system is randomized. The diagnostic rate of esophageal neoplasm by the conventional white light system and that of the NBI system are compared. The endpoint is the neoplastic lesion detected on endoscopic biopsy. We considered invasive cancer, carcinoma in-situ, and high-grade dysplasia (which usually cannot be distinguished from carcinoma in situ) as the primary points. Any dysplastic lesions were considered as secondary points.