Lai 2003.

Study characteristics
Methods	2‐step prospective randomised controlled trial comparing clinical and surgical staging in previously untreated locally advanced cervical carcinoma Step 1: clinical staging (arm A) versus surgical staging (arm B) Step 2: LAP versus EXP approach Single‐centre trial in Taiwan
Participants	61 women with histologically confirmed FIGO stage IIB to IVA cervical cancer. 29 in clinical staging arm, 32 in surgical staging of which 17 in EXP arm and 15 in LAP arm In the clinical staging arm the median age was 53 years, range 33 to 70 years. In the surgical staging arm it was 56 years, range 36 to 69 years FIGO stage in the clinically staged arm was; IIB: 18 (62%) women, IIIA: 1 (3%), IIIB: 10 (35%). In the surgical staged arm it was; IIB: 15 (47%), IIIA: 2 (6%), IIIB: 15 (47%) The primary cervical tumour was > 6 cm in 5/29 (17%) of those in the clinically staged arm and 8/32 (25%) in the surgically staged arm 27 (93%) of the women in clinical staging arm and 27 (84%) in the surgical staging arm had squamous cell carcinoma. The remaining women in each arm had adenocarcinoma or adenosquamous carcinoma Tumour grade was not reported All women had a PS of 0 to 2, but numbers for each PS was not reported by treatment arm
Interventions	Interventions: Clinical staging (tumour size and FIGO stage was measured by pelvic examination, with either a contrast enhanced CT scan or an MRI scan to determine nodal status) Surgical staging (EXP or LAP approach) Further randomised comparison of EXP with LAP approach in the surgical staging group EXP: The skin incision could be either midline vertical or lateral "J", and so on. The retroperitoneum was exposed by rolling the peritoneum medially till the psoas muscle and iliac vessels were visualised. The bifurcation of the aorta, the inferior vena cava, the ovarian vessels, inferior mesenteric artery, and the ureters were identified. Any enlarged or suspicious nodes were excised, or biopsied if unresectable. The nodal tissue from the level of common iliac bifurcation to the inferior mesenteric artery was removed. Lymph nodes submitted for pathology were separately submitted by location (para‐aortic or pelvic) and laterality. Dissection above the level of the inferior mesenteric artery was restricted to those cases with palpable suspicious nodes. The caudal and cephalic extent of lymph node dissection and unresectable nodes were outlined with clips. Sampling of lower pelvic nodes could be performed if indicated or at the discretion of the responsible surgeon (if lateral J incision was used, only unilateral pelvic nodes were assessed). Intraperitoneal lavage cytology was performed via a small opening after completion of the staging procedures LAP: The woman was placed in the supine position in the Trendelenburg position. Pneumoperitoneum was obtained with a Veress needle insertion at the midpoint between the umbilicus and xyphoid process; a 5‐trocar technique was used. Intraperitoneal lavage cytology and biopsy of any suspicious area was done after initial LAP overview of the whole abdomen. The extent of lymph node dissection was similar to that of EXP All women randomised to receive LAP staging were operated on under the collaboration of attending surgeons from the Divisions of Gynecologic Oncology and Gynecologic Endoscopy, Department of Obstetrics and Gynecology. EXP staging procedures were performed by fellows under supervision or by attending gynaecologic oncologists
Outcomes	Overall survival Progression‐free survival Per cent improvement in para‐aortic node metastasis detection Feasibility and efficacy of both surgical staging methods; operation time, blood loss and lymph node yield Toxicity and survival of extended‐field radiation plus chemotherapy in the management of surgical staging documented para‐aortic lymph node metastasis The median follow‐up time of surviving women was 58 months
Notes	Participant accrual was stopped after interim analysis due to significantly worse progression‐free survival in the surgical staging arm. Data were analysed after a further 24 months' follow‐up. The research team planned to randomise 120 women, to ensure a > 80% power to detect a 20% improvement in detecting para‐aortic lymph node metastasis. The interim analysis was after enrolment of 61 women There was no significant difference between the clinical and surgical staging groups in terms of women who received concurrent chemotherapy (15/32 (47%) versus 19/29 (65%), P value = 0.198) but the time from randomisation to starting radiotherapy was significantly longer for surgical staging (median 20 days (range 10 to 46 days) versus 11 days (range 2 to 52 days), P value = 0.001) Para‐aortic lymph node metastasis were identified in 11 of the 32 women in the surgical staging group; 6 of 15 (40%) in the LAP staging group and 5 of 17 (29%) in the EXP staging group ≥ grade 3 treatment‐related toxicities occurred in 13/29 (45%) women in the clinical staging group and 12/32 of those in the surgical staging group (LAP 7/15 (47%) and EXP 5/17 (29%)) 2 women in the surgical staging group and 1 in the clinical staging group died of treatment‐related complications without recurrent or persistent disease. Recurrent or persistent disease occurred in 10/29 (34%) of those clinically staged and 21/32 (66%) women who were surgically staged (P value = 0.021). Local pelvic recurrence occurred in 5/29 (17%) women who were staged clinically and 12/32 (38%) of those staged surgically. Distant recurrence, with or without pelvic recurrence, occurred in 5/29 (17%) of those staged clinically and 9/32 (28%) of those staged surgically Median time to recurrence was 14.3 months (range 6.2 to 23.9 months) for those clinically staged and 6.1 months (range 1.4 to 45.9 months) for those staged surgically At 96 months from first enrolment, 11/29 (38%) of the women in the clinical staging arm and 22/32 (69%) in the surgical staging arm had died
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	"Those who were eligible and had signed informed consent were at first randomized to receive either arm A or arm B, then patients on arm B were randomized between LAP and EXP. Stratification factor was tumor size (break point at 6 cm)"
Allocation concealment (selection bias)	Unclear risk	Not reported
Blinding (performance bias and detection bias) All outcomes	Unclear risk	Not reported
Incomplete outcome data (attrition bias) All outcomes	Low risk	% analysed: 61/61 (100%)
Selective reporting (reporting bias)	Low risk	Trial appears to have reported all pertinent outcome data types. Although it would have been beneficial to have reported QoL it seemed very unlikely that this would have been selectively reported
Other bias	Unclear risk	Insufficient information to assess whether an additional source of bias may have been present

CT: computerised tomography; EXP: extraperitoneal; FIGO: International Federation of Gynaecology and Obstetrics; LAP: laparoscopic; MRI: magnetic resonance imaging; PS: performance status; QoL: quality of life.