Cetina 2013.

Study characteristics
Methods	RCT conducted in Mexico. Unclear whether this was a single or multicentre trial.
Participants	Between May 2004 and June 2009, 211 eligible women with a histological diagnosis of untreated FIGO Stage IB2–IIB cervical cancer with no evidence of cancer in para‐aortic lymph nodes (as evaluated by CT scan). 100 participants were assigned to brachytherapy (brachytherapy) and 111 to RH. Aged 18–70 years. Women were ineligible for the study if they had previously received chemotherapy or radiotherapy. No difference in baseline characteristics between arms. Median age 44 years (range 23–66 years) in the brachytherapy group and 45 years (range 25–62 years) in the RH group. Karnofsky's performance status Median score was 90 in both arms (range 80–100 in the brachytherapy arm and 70–100 in the RH arm). Mean tumour size 32 mm in both arms (range 12–64 mm in the brachytherapy arm and 12–81 mm in the RH arm). Table 1 in Cetina 2013 stated that the median tumour size in each group was 32 cm but we assumed this should have been 32 mm. FIGO Stage FIGO Stage IB: 218 (18%) women in the brachytherapy arm and 18 (16%) in the RH arm. FIGO Stage IIA2: 12 (12%) women in the brachytherapy arm and 11 (10%) in the RH arm. FIGO Stage IIB: 70 (70%) women in the brachytherapy arm and 82 (74%) in the RH arm. Histopathological type Squamous cell cancer: 83 (83%) women in the brachytherapy arm and 100 (90%) in the RH arm Adenocarcinoma: 14 (14%) women in the brachytherapy arm and 8 (7%) in the RH arm. Adenosquamous carcinoma: 3 (3%) women in the brachytherapy arm and 3 (3%) in the RH arm. Haemoglobin Median haemoglobin concentration: 13.3 g/dL (range 10.1–18 g/dL) in the brachytherapy group and 12.8 g/dL (range 10–16 g/dL) in the RH group.
Interventions	Brachytherapy after external‐beam chemoradiotherapy with gemcitabine + cisplatin Immediately after completion of external‐beam radiotherapy and chemotherapy, women in the brachytherapy group underwent low‐dose rate brachytherapy. An brachytherapy dose of 30–35 Gy was delivered to 'point A' to result in a cumulative dose of 80–85 Gy combining external‐beam radiotherapy and brachytherapy. The cumulative external‐beam radiotherapy and brachytherapy dose to 'point B' (the pelvic wall) was 55–65 Gy. RH after external‐beam chemoradiotherapy with gemcitabine + cisplatin Within 4–6 weeks after the EBRT and chemotherapy, women in the RH group were submitted to type III RH and bilateral pelvic lymph node dissection and para‐aortic lymph node sampling if the multidisciplinary team judged the disease could be resected obtaining margins free of disease. Postoperative low‐dose rate brachytherapy was mandated in the surgical arm women only if the surgical specimen revealed positive surgical margins and was administered within 4 weeks after surgery at a median dose of 30 Gy to the vaginal mucosa delivered to a depth of 0.5 cm. External‐beam chemoradiotherapy with gemcitabine + cisplatin External‐beam radiotherapy 50.4 Gy to the entire pelvic region in 28 sessions of 1.8 Gy/day, 5 days/week, over the 6 weeks of chemotherapy.
Outcomes	OS PFS Pathological response: defined as complete with the absence of viable malignant cells in the surgical specimen Operative complications: defined as intraoperative including: bladder, ureteral, bowel and vascular injuries and estimated blood loss > 1000 mL Early postoperative and long‐term complications: defined as any adverse event occurring within or after 30 days from surgery Late complications: including proctitis, cystitis and hydronephrosis Toxicity to chemoradiotherapy with cisplatin and gemcitabine (not reported by treatment arm)
Notes	Median length of follow‐up was 36 months (3–80 months). Dose modification was not allowed for any of the drugs. In the brachytherapy arm, 86 (86%) completed treatment as per protocol. 3 (3%) women abandoned treatment during external‐beam radiotherapy and 11 (11%) had residual tumours after external‐beam radiotherapy that prevented application of brachytherapy. In the surgical arm, 86/111 (77%) women received the full intervention. 15 (13.5%) abandoned treatment during external‐beam radiotherapy; 7 (6%) were judged not to be resectable and 3 (3%) had medical contraindication to surgery. No differences between the median dose and days to complete external‐beam radiotherapy between arms. Both arms received a median of 5 cycles (1–6) of cisplatin + gemcitabine. Among the 86 women who received surgery, 62 (72%) had pathological complete response and 24 (28%) had pathological partial response. In 16/24 women, there were only microscopic foci. The median number of pelvic and para‐aortic lymph nodes removed was 30 (range 8–60) in the brachytherapy arm and 14 (range 3–32) in the surgical arm. 9 women (10%) had positive pelvic lymph nodes (median 3; range 1–7) and 6 of these also had positive para‐aortic lymph nodes (median 2; range 1–4), and they received para‐aortic radiotherapy. PFS rates in the brachytherapy arm were 75% vs 72% in the surgical arm. OS was 76% in the brachytherapy vs 74.5% in the surgical arm. In the univariate analysis, none of the factors analysed (time to complete external‐beam radiotherapy, < 45 vs > 45 days; histology, squamous vs non‐squamous; clinical Stage IB2–IIA vs IIB; age < 50 vs > 50 years and haemoglobin > 12 vs > 12 g/dL) for either PFS or OS were significant. The multivariate analysis also showed none to be significant. Toxicity to chemoradiotherapy: the most frequent toxic effects were haematological and gastrointestinal. Grade 3 leukopenia occurred in 30% and neutropenia in 25% of women. Acute complications in the surgical arm Median hospital stay was 5 days (range 4–6 days). Median surgical time was 4 hours (range 4–6 hours). Median blood loss was 450 mL (range 150–1600 mL). 12 women (14%) were transfused. 3 women (3.5%) had vascular laceration; 1 (1.5%) had a urethral tear and 2 (2%) had section of the ureter. 1 woman (1.5%) had wound dehiscence and 1 (1.5%) woman had infection in the surgical wound. Late toxicity In the brachytherapy arm, grade 1 and 2 proctitis and cystitis were registered in nearly half of the women; however, grade 3 occurred in only 2% and grade 4 in only 2%. In the surgical arm, 6 women had infection after 30 days from surgery. 3 (3.4%) women had unilateral lymphocysts that required treatment with percutaneous drainage in 2 and lymphocyst resection and drainage in 1. In addition, 2 (2.3%) women had uretero‐cutaneous fistulae treated with surgery and double J‐stent positioning.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Not reported. Quote: "After signing an informed consent, women were stratified according to FIGO stage IB2–IIA or IIB and randomly assigned before chemoradiation".
Allocation concealment (selection bias)	Unclear risk	Not reported.
Blinding of participants and personnel (performance bias) All outcomes	High risk	Impossible to blind participants and clinicians to these interventions.
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not reported.
Incomplete outcome data (attrition bias) All outcomes	Low risk	211/211 (100%) eligible women were analysed for efficacy and toxicity and survival outcomes were analysed using appropriate statistical techniques. 18/211 women were lost to follow‐up.
Selective reporting (reporting bias)	Low risk	All pertinent outcomes appeared to have been reported by the trial authors.
Other bias	Unclear risk	Insufficient information to assess whether an additional risk of bias existed.