Methods	NSGO-EC-9501 (in collaboration with EORTC −55991 - amendment 1) was a phase III open, randomised I national clinical trial
Participants	320 women from 13 Nordic centres with surgical stage I, II, IIIA (positive peritoneal fluid cytology only), or IIIC (positive pelvic lymph nodes only) endometrial cancer participated. Women had histologically proven endometrial cancer of one of the following types: FIGO grade 3 (poorly differentiated endometrioid adenocarcinoma) with infiltration ≥ 50% of the myometrial thickness or any serous, clear cell or anaplastic carcinoma and women had to have had an endometrial cancer confined to the corpus cervix, regional lymph nodes or peritoneal fluid cytology. They were treated by hysterectomy, bilateral salpingo-oophorectomy and external beam radiotherapy. The trial recommended CT-based computer-aided 3-dimensional dose planning and four field technique. The prescribed dose to the target volume was at least 44 Gray. If other fractionations than 2.0 Gray five times per week were used, the dose was converted to a 2 Gray equivalent dose according to the linear quadratic formula. Brachytherapy was included as a part of the standard radiotherapy procedure in some centres and the decision to deliver brachytherapy was made before randomisation Women were excluded if there was: Any macroscopic postoperative residual tumour. Spread of disease on the surface or outside the uterus (Exceptions: macroscopically radically excised pelvic lymph nodes or positive peritoneal fluid cytology allowed). Macroscopic tumour involvement of the cervix at preoperative examination. Para-aortic lymph node involvement. Squamous carcinoma or small cell carcinoma with neuroendocrine differentiation. Preoperative irradiation. Previous or concurrent malignant disease except for curatively treated carcinoma in situ of the cervix or basal cell carcinoma of the skin. Active infection or other serious underlying medical condition, which might prevent the patient from receiving treatment or to be followed Uncontrolled or potentially active site of pelvic infection (e.g. fistula or abscesses). Inadequate bone marrow, liver, or kidney function (WBC < 3.5×109/L, neutrophils < 1.5×109/L, platelets < 100×109/L S-creatinine > upper normal value, total S-bilirubin > 2 × upper normal value, estimated GFR < 50 ml/min according to Cockroft-Gault or Jeliffe for patients who will be treated with cisplatin. Previous extensive abdominal surgery or other condition that might give a substantial increase in the risk for complications from the radiation treatment or chemotherapy. Longer interval than 7 weeks between surgery and randomisation
Interventions	Women treated by Surgery and radiotherapy were allocated no additional treatment or chemotherapy. Chemotherapy was given before or after completion of the radiation. The decision about sequence had to be taken before randomisation. The following combinations were used: AP: Doxorubicin 50 mg/m2 (epirubicin 75 mg/m2) and cisplatin 50-75 mg/m2 q 3 weeks. Cisplatin could be substituted with carboplatin AUC 5-6. TP: Paclitaxel 175 mg/m2/3 hours and carboplatin AUC 5-6 q 3 weeks. TEC: Paclitaxel 175 mg/m2/3 hours, epirubicin 60 mg/m2 (doxorubicin 40 mg/m2), and carboplatin AUC 5 q 3 weeks TAP: Doxorubicin 45 mg/m2 and cisplatin 50 mg/m2 day 1, paclitaxel 160 mg/m2/3 hours day 2, G-CSF 5 μg/kg/day starts on day 3 and continues for at least 10 days q 3 weeks (GOG-177 (46)). At least four cycles were planned.
Outcomes	The primary end-point was to compare progression-free survival of patients treated with either radiation alone or radiation and chemotherapy given sequentially. The secondary end-points were overall survival, toxicity and pattern of progression
Results (OS & PFS)	see the combined EORTC and NSGO table
Results (site of recurrence)	see the combined EORTC and NSGO table
Toxicity	see the combined EORTC and NSGO table
Notes	For follow-up, patients were seen 3 and 6 months after end of treatment, thereafter every 6 months for 5 years. Survival was calculated from the date of randomisation until the date of death. The progression-free survival was calculated as time elapsed from date of randomisation to date of progression or death of disease, whichever was the first registered event. Patients who died of non-disease-related causes were censored at time of death. Toxicity was graded according to CTC 1999
Risk of bias
Bias	Authors’ judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Block randomisation was performed centrally by the study office at Linkoping University Hospital with stratification for centre and histology
Allocation concealment (selection bias)	Low risk	The data analysis was organised by the data centre at the NSGO. Although the statistician’s awareness of the group allocation has not been formally stated, it will have been of the highest quality and will have observed the usual rules
Blinding (performance bias and detection bias) All outcomes	High risk	It was not possible to conceal the treatment allocation from the participants
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	The trial has not yet been reported
Selective reporting (reporting bias)	Low risk	There is no suggestion of selective reporting
Other bias	Low risk	There is no suggestion of bias
Main problems with the trial as perceived by this analysis	Low risk	The main issues are that the study population were a heterogenous group with very variable risk factors including some women with grade 1 disease, some with clear cell cancers and some with stage 3