Endoscopic ultrasound is inadequate to determine which T1/T2 esophageal tumors are candidates for endoluminal therapies.

OBJECTIVES

Esophageal endoscopic ultrasound is now regarded as essential in the staging of esophageal carcinoma. There is an increasing trend toward endoluminal therapies (ie, endoscopic mucosal resection and radiofrequency ablation) for pre-cancer or early-stage cancers because of concerns of high morbidity associated with esophagectomy. This study reviews our institutional experience with preoperative endoscopic ultrasound staging of early esophageal cancers in patients who underwent an esophagectomy to evaluate the accuracy of staging by endoscopic ultrasound and how this affects treatment recommendations.

METHODS

A prospective esophagectomy database of all patients undergoing an esophagectomy for esophageal cancer at a single high-volume institution was retrospectively reviewed for patients with early-stage esophageal cancer. This study analyzed patients with clinical Tis to T1 disease, as predicted by preoperative endoscopic ultrasound, and correlated this with the pathologic stages after esophagectomy. The surgical outcomes were evaluated to assess the safety of esophagectomy as a treatment modality.

RESULTS

From 2005 to 2011, 107 patients (93 male, 14 female) with a mean age of 66 years (range, 39-91 years) were staged by preoperative endoscopic ultrasound to have esophageal high-grade dysplasia, carcinoma in situ, or T1 cancer and underwent an esophagectomy. Tumor depth was correctly staged by endoscopic ultrasound in only 39% (23/59) of pT1a tumors (invading into the lamina propria or muscularis mucosa) and 51% (18/35) of pT1b tumors (submucosal). Of the endoscopic ultrasound-staged cT1a-lpN0 lesions, there were positive lymph nodes in 15% of pathologic specimens (2/13). Patients with pT1a-mm lesions had a 9% rate of pathologic lymph node involvement (1/11), and those with pT1b tumors had a 17% rate of lymph node spread (6/35). Esophagectomy was performed in all 107 patients with a 30-day mortality rate of less than 1% (1/107).

CONCLUSIONS

The sensitivity and specificity of endoscopic ultrasound for determining true pathologic staging are poor for early-stage esophageal cancers. Lesions thought to be cT1a-lpN0 by endoscopic ultrasound have at least pN1 disease in 15% of cases. Endoluminal therapy of these lesions based on endoscopic ultrasound undertreats a significant number of patients. Esophagectomy is still the standard therapy for early-stage esophageal cancers in the majority of patients.