This is a randomized noninferiority multicenter trial. Patients will be stratified according to the participating hospital. Patients will be randomized to one of the treatment arms.

Arm A: Patients will receive surgical resection, including Ivor Lewis esophagectomy or McKeown esophagectomy and systematic lymphadenectomy.

Arm B: Patients will receive 5-fluorouracil (5-FU) and cisplatin-based chemotherapy concurrently with radiotherapy. Patients will receive cisplatin (45~60mg/m2) intravenously over 1 hour on day 1 and receive 5-FU (3,200 ~ 4,000mg/m2) intravenously for 4 to 5 days. Treatment will repeat every 3 weeks for 2 courses. Patients will receive a total of 45 Gy irradiation (5 days a week for 5 weeks).

Patients will be followed at 3 and 6 months after randomization, then every 6 months for following 2 and half years (up to 3 years after randomization), and 4 and 5 years after randomization. After 5 years, annual follow-up is scheduled up to 10 years after randomization.

We will analyze the results primarily with the intention-to-treatment(ITT) analysis, and then secondarily with the per-protocol(PP) analysis as well.

• Compare 1-year, 2-year, 3-year, 5-year, 6-year, 7-year, 8-year, 9-year, 10-year overall survival between concurrent chemoradiotherapy and esophagectomy among patients who require additional treatment for esophageal squamous cell carcinoma endoscopically resected but found to have pT1b (submucosal invasion) or lymphovascular invasion
• Compare 1-year, 2-year, 3-year, 5-year, 6-year, 7-year, 8-year, 9-year, 10-year disease-free survival between concurrent chemoradiotherapy and esophagectomy in the same study population
• Compare quality of life between concurrent chemoradiotherapy and esophagectomy in the same study population
• Compare treatment-related adverse event between concurrent chemoradiotherapy and esophagectomy in the same study population

Inclusion Criteria:

1. Age ≥ 19 years and < 80 years
2. Histologically confirmed squamous cell carcinoma of the esophagus
3. Clinical stage as cT1N0M0 (AJCC/UICC 7th Edition) according to upper GI endoscopy or endoscopic ultrasound and chest computed tomography (CT) scans
4. Pathologic examination after endoscopic submucosal dissection confirmed the presence of submucosal invasion (pathologic T1b) or lymphovascular invasion
5. For participants with multiple lesions, all of them should be resected with endoscopic submucosal dissection and at least one lesion should have pathologic submucosal invasion (pT1b) or lymphovascular invasion
6. Participants has adequate hematologic function, as evidenced by an absolute neutrophil count (ANC) ≥1000/µL, hemoglobin ≥8 g/dL, and platelets ≥85,000/µL
7. Participants has adequate hepatic and renal function as defined by aspartate transaminase (AST) and alanine transaminase (ALT) ≤ 2.5 times the upper limit of normal ; a total bilirubin ≤1.5 times the upper limit of normal ; creatinine clearance ≥ 30mL/min/1.73m2
8. Participants should agree to participate in the study and sign the informed consent form

Exclusion Criteria:

1. Cervical esophageal cancer (proximal to 20cm from incisor teeth)
2. Regional lymph node metastasis (cN+) or distant metastasis (cM1) are suspected or confirmed on chest CT scans or positron emission tomography (PET)/CT scans (Equivocal results will be regarded as no metastasis. However, it can also perform a biopsy if necessary (optional))
3. Recurrent esophageal cancer
4. Uncontrolled systemic disease which makes participants medically unfit for additional treatment (esophagectomy or concurrent chemoradiotherapy) such as congestive heart failure, interstitial lung disease, severe pulmonary emphysema or chronic renal failure
5. Gastric conduit is not available for esophageal reconstruction (ex.: previous history of gastrectomy)
6. Synchronous or metachronous multiple cancers (within the past 3 years) with the exclusion of skin cancer, well differentiated thyroid cancer, carcinoma in situ, early cancer achieving curative endoscopic resection, or low grade prostate cancer (Gleason Score≤6)

Outcome assessors and data analysists will be blinded