Introduction: Computed tomography (CT) and endoscopic ultrasound (EUS) are part of the regular staging protocol in esophageal cancer. The value of the two methods was assessed in patients with early cancer in Barrett's esophagus.
Methods: One hundred consecutive patients (median age 64 yr, interquartile range [IQR] 58-72) with suspected early cancer in Barrett's esophagus who were referred to our hospital for endoscopic therapy were prospectively included in a standardized staging program with upper gastrointestinal endoscopy, EUS (7.5 MHz in all cases plus 12.5 or 20 MHz for elevated and/or depressed lesions), CT of the chest and upper abdomen, and abdominal ultrasonography. The results were summarized in accordance with the TNM classification. On the basis of the lymph node findings on CT and/or EUS, the patients were assigned to three categories: C1, no suspicious lymph nodes; C2, paraesophageal lymph nodes < or =1 cm in size at the tumor level, lymph nodes > or =1 cm in size not at the tumor level in the mediastinum or celiac trunk; and C3, paraesophageal lymph nodes > 1 cm in size at the tumor level. The EUS and CT findings were checked every 6 months in patients who underwent endoscopic treatment. Surgical resection was scheduled in operable patients if staging showed a T category higher than T1 and/or the lymph node staging was assessed as C3. Patients with suspected submucosal infiltration underwent diagnostic endoscopic resection, and if submucosal involvement was confirmed were referred for surgery.
Results: The median follow-up period was 25 months (IQR 19.5-30.0). The T category diagnosed with CT was < or = T1 in all patients. On EUS, the T category was classified as T1 in 92% of cases (N = 92) and as > T1 in 8% (N = 8, p < 0.05). Enlarged lymph nodes (C2 and C3) were detected in 45% of the patients. Significantly more C2 lymph nodes were diagnosed with EUS than CT (28 vs 19, p < 0.05). Lymph nodes at the level with the highest suspicion, C3, were detected using CT in only three of nine cases. Sensitivity of CT for N staging was not acceptable compared with EUS (38%vs 75%). No extranodal metastases were found on CT.
Conclusions: In suspected early cancer in Barrett's esophagus, EUS is superior to CT for T staging and N staging. As CT had no influence on the TNM classification in any of these patients, it may be possible to dispense with this method as a staging procedure in patients with cancer in Barrett's esophagus. By contrast, EUS is required in order to differentiate between patients with cancer in Barrett's esophagus in whom endoscopic therapy is suitable and those in whom surgical treatment is required.