Su1498 What Are the Risk Factors for Misdiagnosing Invasion Depth for Superficial Esophageal Squamous Cell Carcinoma by Narrow Band Imaging and Endoscopic Ultrasonography?

Abstract

Su1498 What Are the Risk Factors for Misdiagnosing Invasion Depth for Superficial Esophageal Squamous Cell Carcinoma by Narrow Band Imaging and Endoscopic Ultrasonography? Masaki Ominami*, Yasuaki Nagami, Masatsugu Shiba, Tomoko Obayashi, Junichi Okamoto, Kunihiro Kato, Yuji Nadatani, Shusei Fukunaga, Satoshi Sugimori, Noriko Kamata, Mitsue Sogawa, Hirokazu Yamagami, Tetsuya Tanigawa, Kenji Watanabe, Toshio Watanabe, Kazunari Tominaga, Yasuhiro Fujiwara, Tetsuo Arakawa Gastroentelogy, Osaka City University Graduate School of Medicine, Osaka, Japan Background: It is vital to estimate carcinoma invasion depth to determine indications for treating superficial esophageal squamous cell carcinomas (ESCCs). Although narrow band imaging (NBI) and endoscopic ultrasonography (EUS) are useful for determining superficial ESCC invasion depth, few reports have shown the risk factors of misdiagnosis by such modalities. Aim: This study aimed to clarify the usefulness and risk factors for misdiagnosing superficial ESCC invasion depth using NBI and EUS. Methods: Between October 2008 and April 2011, 47 patients (39 men; mean age, 67.4 years) with 52 superficial ESCCs who underwent NBI and EUS before ESD or surgery were included in this study. Patients who had received chemotherapy or radiotherapy were excluded. Three endoscopists reviewed the recorded endoscopic images collected using NBI and EUS and evaluated the invasion depth as EP/LPM, MM/SM1 (submucosal invasion 200 m), or SM2 (deeper than SM1). We used the Japan Esophageal Society classification of NBI composed of vessel shape and width, and avascular area. The diagnostic accuracies of NBI and EUS were compared with the pathology. Multivariate analysis was performed to clarify the risk factors of invasion depth misdiagnosis. The cut-off values of major diameter and circumferential extension were calculated using receiver operating characteristic curve analysis. Results: The average accuracies of NBI and EUS were 76.9% (40/ 52) and 67.3% (35/52), respectively. Through NBI, the following were correctly diagnosed: 31 of 32 (96.9%) EP/LPM lesions, 6 of 14 (42.9%) MM/SM1 lesions, and 3 of 6 (50.0%) SM2 lesions. Overestimation occurred in 8 lesions and underestimation occurred in 4 lesions. Through EUS, the following were correctly diagnosed: 31 of 34 (91.2%) EP/LPM lesions, 2 of 6 (33.3%) MM/SM1 lesions, and 2 of 12 (16.7%) SM2 lesions. Overestimation occurred in 11 lesions and underestimation occurred in 6 lesions. Multivariate analysis showed that major diameter (odds ratio [OR], 1.65; 95% confidence interval [CI], 1.03-2.64; p 0.04) and circumferential extension (OR, 1.49; 95% CI, 1.04-2.15; p 0.03) were the risk factors of NBI misdiagnosing invasion depth. In contrast, no significant differences were noted for EUS. The cut-off values calculated on major diameter and circumferential extension were 28 mm (95% CI, 0.56-0.86; area under the curve [AUC], 0.71) and 0.50 (95% CI, 0.57-0.87; AUC, 0.72), respectively. The average accuracies of NBI and EUS were 91.7% (22/24) and 79.2% (19/24) or 64.3% (18/28) and 57.1% (16/28), respectively, in the lesions 28 mm or 28 mm of the major diameter. Conclusion: Major diameter and circumferential extension are risk factors of misdiagnosing invasion depth for superficial ESCC using NBI endoscopy. NBI might be useful for diagnosing superficial ESCC invasion depth, especially in cases with major diameters 28 mm.