Sonographic characteristics of locoregional lymph nodes that can predict the presence of metastatic carcinoma by endoscopic ultrasound-guided fine needle aspiration in patients with carcinomas of the esophagus/gastroesophageal junction

Background: Endoscopic ultrasound (EUS) and EUS-guided fine needle aspiration (EUS-FNA) of lymph nodes (LN) are used for preoperative staging of patients with esophageal/gastroesophageal junction (E/GEJ) carcinomas. The recognition of sonographic features of LN can be variable in selecting LN for EUS-FNA. We evaluated the predictive value of sonographic features of LNs that correlate with the presence of metastatic carcinoma by EUS-FNA. Materials and methods: Patients with GIPB who underwent EUS/EUS-FNA for the preoperative staging of LNs over 29 months were studied. The EUS features including size, shape, borders, and echogenicity along with primary tumor stage were evaluated/correlated with the EUS-FNA result to determine features that predict positive EUS-FNA results. Results: 86 LNs from 74 patients with E/GEJ carcinomas were studied. FNA yielded a positive/suspicious for carcinoma diagnosis in 42 cases and a negative result in 44 cases. The sonographic features of the LNs with and without metastatic carcinoma were average short axis measurement (1.02 vs 0.73), average long axis measurement (1.413 vs 1.348), average long:short axis ratio (1.45 vs 1.2), hypoechoic cortex (97.6 vs 72.8 %), round shape (71.4 vs 29.5 %), and well-circumscribed borders (40.5 vs 31.8 %). The features most strongly predictive of a positive result were short axis (p = 0.006), hypoechoic cortex (p = 0.01), and round shape (p = 0.0005) on univariate analysis. High-primary tumor stage (T3 and T4, p = 0.07 and p = 0.27, respectively) did not predict metastatic carcinoma on EUS-FNA. On multivariate analysis, short axis measurement alone was strongly predictive of a positive EUS-FNA (p = 0.04). Conclusions: Sonographic features of LNs including size, shape, echogenicity, and border characteristics can be useful in the selection of pertinent LNs for EUS-FNA sampling in preoperative staging of E/GEJ malignancies based on their ability to predict the presence of metastatic carcinoma. While hypoechoic cortex, round shape, and short axis measurement were independent factors that strongly predicted the presence of metastatic carcinoma on EUS-FNA, short axis alone was strongly predictive of metastatic carcinoma on multivariate analysis.