Treatment trends, risk of lymph node metastasis, and outcomes for localized esophageal cancer

Background Endoscopic resection is increasingly used to treat localized, early-stage esophageal cancer. We sought to assess its adoption, characterize the risks of nodal metastases, and define differences in procedural mortality and 5-year survival between endoscopic and surgical resection in the United States. Methods From the National Cancer Data Base, patients with T1a and T1b lesions were identified. Treatment patterns were characterized, and hierarchical regression methods were used to define predictors and evaluate outcomes. All statistical tests were two-sided. Results Five thousand three hundred ninety patients were identified and underwent endoscopic (26.5%) or surgical resection (73.5%). Endoscopic resection increased from 19.0% to 53.0% for T1a lesions (P <. 001) and from 6.6% to 20.9% for T1b cancers (P <. 001). The strongest predictors of endoscopic resection were depth of invasion (T1a vs T1b: odds ratio [OR] = 4.45; 95% confidence interval [CI] = 3.76 to 5.27) and patient age of 75 years or older (vs age less than 55 years: OR = 4.86; 95% CI = 3.60 to 6.57). Among patients undergoing surgery, lymph node metastasis was 5.0% for T1a and 16.6% for T1b lesions. Predictors of nodal metastases included tumor size greater than 2cm (vs. <2cm) and intermediate-/high-grade lesions (vs low grade). For example, 0.5% of patients with low-grade T1a lesions less than 2cm had lymph node involvement. The risk of 30-day mortality was less after endoscopic resection (hazard ratio [HR] = 0.33; 95% CI = 0.19 to 0.58) but greater for conditional 5-year survival (HR = 1.63; 95% CI = 1.07 to 2.47). Conclusions Endoscopic resection has become the most common treatment of T1a esophageal cancer and has increased for T1b cancers. It remains important to balance the risk of nodal metastases and procedural risk when counseling patients regarding their treatment options.