Accuracy of the endoscopic evaluation of esophageal involvement in esophagogastric junction cancer

BACKGROUND: Esophageal involvement length (EIL) is a promising indicator of metastasis or recurrence in mediastinal lymph nodes (MLNs) in adenocarcinoma of the esophagogastric junction (EGJ). This study aimed to elucidate the accuracy of the preoperative endoscopic evaluations of EIL and its clinical significance in this disease. MATERIALS AND METHODS: In total, 75 patients with Siewert type II (N = 53) or III (N = 22) adenocarcinoma of the EGJ, who underwent surgical resection without preoperative therapy between 1995 and 2016 were enrolled. We retrospectively examined the accuracy of the preoperative endoscopic evaluations of EIL (preoperative EIL), compared to the pathologically evaluated EIL. Finally, we investigated the association between preoperative EIL and metastasis or recurrence in MLNs. RESULTS: The accuracy of the preoperative EIL within a 1-cm interval was only 53.3%. Among patients with discordance between the pre- and postoperative evaluations, 68.6 % had the underestimation in the preoperative EIL. pN1–3 (OR = 5.85, 95% CI: 1.03–33.17) and undifferentiated histologic type (OR = 2.52, 95% CI: 0.89–7.14) were potential risk factors for the discordance. Regarding metastasis or recurrence in MLNs, preoperative EIL of 2–3 cm (OR = 10.41, 95% CI: 1.35–80.11) and >3 cm (OR = 8.33, 95% CI: 1.09–63.96) were independent predictors. CONCLUSION: Although the accuracy of the endoscopic evaluations of EIL is insufficient with many underestimations, EIL should be assessed in preoperative staging because of significant predictive power for metastasis or recurrence in MLNs.