Search

Narrow band imaging (NBI) has been used in the gastrointestinal endoscopy field as a novel endoscopic imaging technique and has contributed to improved qualitative diagnosis and detection of lesions. However, there are only a few studies of use of NBI in the gynecology field. We applied NBI in flexible hysteroscopy at our outpatient clinic and evaluated the utility of NBI hysteroscopy for diagnosis of malignant endometrial lesions by comparison of the sensitivity and specificity between white light alone and white light+NBI using hysteroscopic video images. The subjects were 65 patients with a suspected endometrial lesion in the uterine cavity. These patients underwent flexible hysteroscopy using NBI in addition to conventional white light. Video images from 65 patients were edited into two groups, white light alone (WL group) and white light+NBI (NBI group) (130 images in total). Computerized block randomization of the order was then performed. Four raters independently diagnosed the images without use of other clinical information. Using the pathological diagnosis as the gold standard, we evaluated the sensitivity and specificity of diagnosis of atypical endometrial hyperplasia (AEH) or endometrial carcinoma compared between the WL and NBI groups. The sensitivity of diagnosis of AEH or endometrial carcinoma was numerically higher in the NBI group for all raters, and the average sensitivity was significantly higher in the NBI group compared to the WL group (78.6% vs. 63.7%, P<0.001). The specificity for each rater and the average specificity were comparable between the two groups. Compared to white light hysteroscopy, NBI hysteroscopy had a higher sensitivity for diagnosis of AEH or endometrial carcinoma without loss of specificity. This suggests that NBI hysteroscopy may be more useful than white light hysteroscopy for endoscopic diagnosis of malignant endometrial lesions.

Narrow band imaging (NBI) for detection of blood vessels and microstructures on the mucosal surface is used in gastrointestinal endoscopy since it can improve qualitative diagnosis and detection of lesion. However, there are no studies on flexible hysteroscopy using NBI. We performed flexible hysteroscopy with NBI for outpatients to investigate the sensitivity and specificity of endoscopic diagnosis of malignant endometrial lesions. Of patients who attended our hospital for suspected lesions in the uterine cavity between April 2009 and May 2010, 104 subjects underwent hysteroscopy with NBI, in addition to white light. Using the pathological diagnosis as the gold-standard, we evaluated the sensitivity and specificity of NBI hysteroscopy for detecting atypical endometrial hyperplasia (AEH) or carcinoma. The results were also compared with historical data (n=209) for conventional hysteroscopy using white light only in 2008. The sensitivities were 97.2% [95% confidence interval (95% CI): 90.3-99.7%] and 82.6% (95% CI: 74.4-89.0%) for NBI hysteroscopy and conventional hysteroscopy, respectively. The 95% CIs for the two methods did not overlap and the sensitivity of lesion detection was higher with NBI hysteroscopy. Specificities were comparable, 90.6% (95% CI: 75.0-98.0%) and 85.1% (95% CI: 76.3-91.6%) between the methods. NBI hysteroscopy has increased sensitivity for detection of atypical endometrial hyperplasia (AEH) or carcinoma. A comparison with historical data suggested that NBI may be useful for diagnosis of malignant endometrial lesions. As far as we are aware, this is the first evaluation of flexible hysteroscopy with NBI for diagnosis of malignant endometrial lesions.