Chiu 2005.

Methods	RCT
Participants	Country: China Number randomised: 81 Postrandomisation exclusions: 1 (1.2%) Number analysed: 80 Average age: 62 years Females: 14 (17.3%) Inclusion criteria Mid‐ or lower‐thoracic oesophageal cancers that were confirmed on histology to be a squamous cell carcinoma deemed to be resectable. Exclusion criteria Participants who had distant metastasis to solid visceral organs or local invasion into trachea, descending aorta, or recurrent laryngeal nerve. Participants > 75 years or who had a serious premorbid condition or a poor physical status that compromised a thoracotomy. Participants with compromised cardiac function or creatinine clearance less than 50 mL/min were ineligible.
Interventions	Participants were randomly assigned to 2 groups Group 1: chemoradiotherapy (N = 36) Further details: external radiotherapy total of 50 to 60 gray given in 25 to 30 fractions over 5 to 6 weeks and two 3‐weekly cycles of cisplatin 60 mg/m² on day 1 and day 22 and 5‐fluorouracil chemotherapy 200 mg/m²/day from day 1 to day 42 Group 2: oesophagectomy (N = 44) Further details: 2‐ or 3‐stage oesophagectomy
Outcomes	Outcomes reported were short‐term mortality, long‐term mortality, short‐term adverse events, short‐term serious adverse events, long‐term recurrence, and length of hospital stay
Cancer stage/histology	Not reported (must be deemed resectable and have no metastases)/squamous cell carcinoma
Tumour location	Mid‐ or lower‐thorax
American Society of Anaesthesiologists (ASA) physical status score	Not reported
Notes	We attempted to contact the study authors in February 2015 Reason for postrandomisation drop‐out: initially deemed resectable but later considered unresectable Patients were followed up at 6 to 8 week intervals in the 1st year and at 3‐month intervals thereafter Median follow‐up: 1.5 years
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Comment: this information was unavailable.
Allocation concealment (selection bias)	Unclear risk	Comment: this information was unavailable.
Blinding of participants and personnel (performance bias) All outcomes	High risk	Comment: it was impossible to blind the participants and healthcare providers.
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Comment: this information was unavailable.
Incomplete outcome data (attrition bias) All outcomes	High risk	Comment: there was an imbalance in postrandomisation exclusions.
Selective reporting (reporting bias)	Low risk	Comment: the trial reported all important outcomes.
Source of funding	Low risk	Quote: "This study was supported by the Research Grant Council (RGC) of Hong Kong Special Administrative Region, China".
Other bias	Low risk	Comment: there was no other source of bias.