Surg Endosc Although eyond the ot th ades the nd TA 12.4.7,8].the role of EUS in the detection of regional invades the adjacent organs).The criteria for malig nodal metastasis still is controversial.This large po ston lymph nodes included a round shape,a diameter exceeding erative study aimed to compare the accuracy of EUS 10 mm,hypoechogenecity.and well-demarcated borders staging ith that of PET and staging using postopera [9].Celiac lymph nodes were defined by their location staging of PET/CT imaging Methods All scans were performed using a PET/CT system(Philips Gemini,DA best,Ne lands)wit 2-(Huorne-18)-fuoro Patients 2deoypehcoDC Aherleaa6hfsL nosis of hion -nr trast CT scan was performed before PET.and the resultin tively enrolled in Seoul National University Hospital to data were used to create an attenuation correction map for undergo a pre perative staging protocol that included EUS. whole-body P and chest CI scan cor the skull base to the mid thigh region. patients emission the staging p dimensional tations.In the and extens ive regional lymph node dissection.The Insti- walitative assessment of the images foci with increased tutional Review Board of the Seoul National University FDG uptake compared with the surrounding structures Hospital approved this study,and written informed con- were defined as malignancies.For quantitative analysis of from all the patients All information FDG uptake.the maximal standardized uptake value (SUVmax)was assessed EUS imaging CT imaging All EUS examinations were performed by a single expe All CT scans were performed with a Genesis High-Speed (S.G.K.),who had a high volume of CT(GE Medical Systems,Milwaukee,WI,USA).Scans S.The exa min a radial 370 14 mi/kg of with 12 MHz asse ssed the depth of 120 kVn and 230 mAs The in ing data were re (T stage)whereas imaging with 5 MHz detected the structed with 7-mm collimation The CT seans were per malignant-appearing perigastric or mediastinal lymph formed from the lower neck to the lower pole of the nodes (N stage).c liac nodes,or hepatic metastases (M kidney.Images were evaluated for the presence of primary stage).Minimal alloo Ito avoid tumo tumor,locoregional and/ metast mpus)were sed for mm)were cons ered pos sitive fo uperficial lesion or malignant strictur metastasis The HFUS examinations were performed after water instillation into the esophageal lumen and suctioning of Pathologic stage luminal air.If the echoendoscope could not pass the e patients did notn the proxima proce Pathologic metastas cation system ng to th The EUS criteria used for T staging were as follows:TI cancer staging 0.The ional lymph nodes included (tumor invades the mucosa or submucosa but not the paratracheal posterior and anterior mediastinal subcarinal muscularis propria).T2(tumor invades but does not extend paraesophageal,pericardial,inferior pulmonary ligament, SpringerAlthough several reports have compared the perfor￾mance of EUS, CT, and PET in staging esophageal cancer [2, 4, 7, 8], the role of EUS in the detection of regional nodal metastasis still is controversial. This large postop￾erative study aimed to compare the accuracy of EUS staging with that of PET and CT staging using postopera￾tive pathologic staging as a gold standard. It also aimed to determine the impact of EUS in preoperative locoregional staging of resectable esophageal cancer. Methods Patients From September 2005 to July 2008, patients with a diag￾nosis of biopsy-proven esophageal cancer were prospec￾tively enrolled in Seoul National University Hospital to undergo a preoperative staging protocol that included EUS, whole-body PET/CT, and chest CT scan concomitantly. All patients eligible for surgical resection who had no distant metastasis or extensive adjacent organ invasion in the staging protocol were allocated to esophageal resection and extensive regional lymph node dissection. The Insti￾tutional Review Board of the Seoul National University Hospital approved this study, and written informed con￾sents were obtained from all the patients. All information was collected and analyzed retrospectively. EUS imaging All EUS examinations were performed by a single expe￾rienced endoscopist (S.G.K.), who had a high volume of EUS. The examination was performed using a radial array echoendoscope (Olympus GF-UM 2000;Tokyo, Japan) with the patient under sedation using midazolam. Imaging with 12 MHz assessed the depth of primary tumor invasion (T stage), whereas imaging with 5 MHz detected the malignant-appearing perigastric or mediastinal lymph nodes (N stage), celiac nodes, or hepatic metastases (M stage). Minimal balloon inflation was used to avoid tumor compression. In addition, high-frequency ultrasonic (HFUS) probes (UM-2R, 20 MHz; Olympus) were used for superficial lesion or malignant stricture. The HFUS examinations were performed after water instillation into the esophageal lumen and suctioning of luminal air. If the echoendoscope could not pass the stricture, the patients did not undergo the dilation proce￾dure, and staging was assessed from the proximal tumor margin. The EUS criteria used for T staging were as follows: T1 (tumor invades the mucosa or submucosa but not the muscularis propria), T2 (tumor invades but does not extend beyond the muscularis propria), T3 (tumor invades the adventitia but not the adjacent organs), and T4 (tumor invades the adjacent organs). The criteria for malignant lymph nodes included a round shape, a diameter exceeding 10 mm, hypoechogenecity, and well-demarcated borders [9]. Celiac lymph nodes were defined by their location within a 2-cm area from the celiac trunk and indicated distant metastasis. PET/CT imaging All scans were performed using a PET/CT system (Philips Gemini, DA best, Netherlands) with 2-(fluorine-18)-fluoro- 2-deoxy-D-glucose (FDG). After at least a 6-h fast, images were obtained 60 min after intravenous injection of 555– 740 MBq (15–20 mCi; 0.22 mCi/kg) of 18FDG. A noncon￾trast CT scan was performed before PET, and the resulting data were used to create an attenuation correction map for PET. Sections 5 mm thick were obtained at 50 mA and 120 kVp from the skull base to the mid thigh region. The PET scan was performed at a 5-min emission acquisition per image level, and the images were recon￾structed into three-dimensional representations. In the qualitative assessment of the images, foci with increased 18FDG uptake compared with the surrounding structures were defined as malignancies. For quantitative analysis of FDG uptake, the maximal standardized uptake value (SUVmax) was assessed. CT imaging All CT scans were performed with a Genesis High-Speed CT (GE Medical Systems, Milwaukee, WI, USA). Scans were obtained after 1.4 ml/kg of a nonionic contrast medium (Ultravist 370, Schering, Berlin, Germany) had been injected intravenously. The scanning parameters were 120 kVp and 230 mAs. The imaging data were recon￾structed with 7-mm collimation. The CT scans were per￾formed from the lower neck to the lower pole of the kidney. Images were evaluated for the presence of primary tumor, locoregional and/or distant lymph node metastasis, and distant organ metastasis. The lymph nodes with the shortest diameter ([8 mm) were considered positive for metastasis. Pathologic stage Pathologic staging was determined using the tumor node metastasis (TNM) classification system according to the American Joint Committee on Cancer (AJCC) esophageal cancer staging [10]. The regional lymph nodes included paratracheal, posterior and anterior mediastinal, subcarinal, paraesophageal, pericardial, inferior pulmonary ligament, Surg Endosc 123