Detection of mucosal recurrent nasopharyngeal carcinomas after radiotherapy with narrow-band imaging endoscopy

Purpose: This study evaluated the feasibility of screening mucosal recurrent nasopharyngeal carcinoma with narrow-band imaging (NBI) endoscopy. Methods and Materials: One hundred and six patients were enrolled. All patients underwent conventional white-light (WL) endoscopic examination of the nasopharynx followed by NBI endoscopy. Biopsies were performed if recurrence was suspected. Results: We identified 32 suspected lesions by endoscopy in WL and/or NBI mode. Scattered brown spots (BS) were identified in 22 patients, and 4 of the 22 who had negative MRI findings were histopathologically confirmed to be neoplasias that were successfully removed via endoscopy. A comparison of the visualization in NBI closer view corresponded to histopathological findings in 22 BS, and the prevalence rates of neoplasias in tail signs, round signs, and irregularities signs were 0% (0/6), 0% (0/7), and 44.4% (4/9), respectively (p = 0.048). The sensitivity, specificity, and diagnostic capability were 37.5%, 92.9% and 0.652 for WL, 87.5%, 74.5% and 0.810 for NBI, and 87.5%, 87.8%, and 0.876 for NBI closer view, respectively. NBI closer view was effective in increasing specificity compared with NBI alone (87.8% vs. 74.5%, p < 0.05), and in increasing sensitivity and diagnostic capability compared to WL alone (87.5% vs. 37.5%, p < 0.05; 0.876 vs. 0.652, p = 0.0001). Conclusions: Although NBI in endoscopy can improve sensitivity of mucosal recurrent nasopharyngeal neoplasias, false-positive (nonneoplasia BS) results may be obtained in areas with nonspecific inflammatory changes due to postradiation effects. NBI closer view not only can offer a timely, convenient, and highly reliable assessment of mucosal recurrent nasopharyngeal carcinoma, it can also make endoscopic removal possible.