Gupta 2018.

Study characteristics
Methods	Single‐centre RCT conducted in India.
Participants	633 women aged 18–65 years with histologically confirmed squamous cell carcinoma of the cervix with 1994 FIGO Stage IB2, IIA or IIB disease. The baseline characteristics did not show any significant differences between the 2 arms. Histopathological type Squamous cell carcinoma Median age 50 years in the NACT + RH group and 48 years in the CCRT group. Performance status 0: 91.8% in the NACT + RH arm and 92.4% in the CCRT arm. 1: 8.2% in the NACT + RH arm and 7.6% in the CCRT arm. FIGO Stage FIGO Stage IB2 bulky: 18% of women in the NACT + RH arm and 17.7% in the CCRT arm. FIGO Stage IIA bulky: 25.3% of women in the NACT + RH arm and 24.6% in the CCRT arm. FIGO Stage IIB bulky: 56.7% of women in the NACT + RH arm and 57.7% in the CCRT arm. Tumour grade Not specified
Interventions	NACT + RH 3 cycles of paclitaxel 175 mg/m2 and carboplatin (dosed to an area under curve of 5–6) once every 3 weeks. Women underwent clinical response assessment after the second and third cycles of chemotherapy. Women who had no response or disease progression at these times were crossed over to receive definitive CCRT, whereas responders underwent surgery 3–4 weeks after the third cycle of chemotherapy. Women assigned to the NACT + surgery group underwent Piver‐Rutledge class III radical abdominal hysterectomy, bilateral pelvic lymphadenectomy and lower para‐aortic lymph node sampling by expert gynaecological oncologists. Surgery was abandoned in women with intraoperative findings of either unresectable primary tumour or lymph node disease, and these women were treated with definitive CCRT. CCRT Women assigned to the CCRT group and those who crossed over from the NACT + surgery group received standard external‐beam radiotherapy to the whole pelvis + brachytherapy. They received an external radiation dose of 40 Gy in 20 fractions with 2 Gy per fraction and a midline shield at 20 Gy, followed by intracavitary radiation to 'point A' as follows: either 2 applications of a low‐dose rate of 30 Gy each or 5 applications of a high‐dose rate of 7 Gy each. Radiation doses were modified to respect tumour, rectal and bladder constraints. These women also received 5 cycles of cisplatin 40 mg/m2, administered once per week starting with external‐beam radiotherapy. Adjuvant treatment for both groups Women in the NACT + surgery group who underwent RH were given adjuvant therapy (radiotherapy or CCRT) as per protocol defined criteria, in accordance with published evidence. On the basis of histopathological evaluation of the surgical specimen, adjuvant chemoradiotherapy was given in the presence of any 1 of the following features: lymph node metastasis, positive surgical margins or parametrial involvement. Adjuvant radiotherapy alone was given based on the presence of any 2 of the following features: deep cervical stromal invasion, lymphonodular invasion, or tumour size > 4 cm. Women in both groups were evaluated at protocol defined time points to evaluate response, monitor for relapse and assess toxicity.
Outcomes	DFS (primary endpoint) OS (secondary endpoint) Toxicity
Notes	Written informed consent was obtained from all women before inclusion in the study. Women were randomly assigned, after stratification by stage. Between September 2003 and February 2015, 635 women were randomly assigned, of whom 633 (316 in the NACT + surgery group and 317 in the CCRT group) were included in the final analysis, with a median follow‐up of 58.5 months. There was good compliance to planned treatment in both arms, including doses and duration of chemotherapy and radiotherapy. In the NACT + surgery group, 2 (0.63%) women received CCRT, and in the CCRT group, 2 (0.63%) women underwent surgical resection (protocol non‐adherence). In the NACT + surgery group, 68 (21.5%) women crossed over (presurgery crossover and intraoperative unresectable disease) to receive definitive CCRT, 42 (13.3%) women received postoperative adjuvant chemoradiotherapy, and 31 (9.8%) women received postoperative adjuvant radiotherapy, according to protocol‐defined criteria. 5‐year DFS in the NACT + surgery group was 69.3% compared with 76.7% in the CCRT group (HR 1.38, 95% CI 1.02 to 1.87; P = 0.038), whereas the corresponding 5‐year OS rates were 75.4% ‐in the NACT + surgery group and 74.7% in the CCRT group (HR 1.025, 95% CI 0.752 to 1.398; P = 0.87). In subgroup analyses, the DFS detriment in the NACT + surgery group was statistically significant in women with FIGO Stage IIB disease, with a significant test of interaction between treatment effect and Stages IIA and IIB disease. In women with Stage IIB disease, 5‐year DFS rates in the NACT + surgery was 67.2% and CCRT group was 79.3% (unadjusted HR for DFS in the NACT + surgery group, 1.90, 95% CI 1.25 to 2.89; P = 0.003). In the NACT + surgery group, compared with the CCRT group, there was a lower rate of rectal, bladder and vaginal toxicity occurring or persisting 90 days after treatment completion (rectal: 5.7% with NACT + surgery vs 13.3% with CCRT; P = 0.002; bladder: 2.8% with NACT + surgery vs 7.3% with CCRT; P = 0.017); vaginal: 19.9% with NACT + surgery vs 36.9% with CCRT; P = 0.001). However, 24 months after treatment completion, there was no difference in rectal and bladder toxicities between groups, whereas vaginal toxicity continued to occur at a lower rate in the NACT + surgery group (12.0% with NACT + surgery vs 25.6% with CCRT; P = 0.001). Despite adequate NACT, surgery was possible in only 227 (72.15%) women in the NACT + surgery group. The 2 chemotherapy regimens were different in the 2 groups with respect to platinum and paclitaxel dose and administration schedule. Thus, the choice of platinum drug is unlikely to be a critical factor in the outcome of the study. Quality of life was not prospectively measured.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Eligible women were randomly assigned to either study group in a 1:1 ratio using a computerised block design with a block size of 4.
Allocation concealment (selection bias)	Unclear risk	Random assignment was performed by the Clinical Trials Unit of Tata Memorial Centre. Women were stratified according to clinically determined 1994 FIGO Stage IB2, IIA, or IIB disease before random assignment. There were only 4 block sizes.
Blinding of participants and personnel (performance bias) All outcomes	High risk	Study investigators and participants were not blinded to treatment allocation.
Blinding of outcome assessment (detection bias) All outcomes	High risk	Study investigators and participants were not blinded to treatment allocation.
Incomplete outcome data (attrition bias) All outcomes	Low risk	635 women combined and 29 lost to follow‐up but only 2/635 were not included in final analysis.
Selective reporting (reporting bias)	Low risk	DFS was reported in first instance without the inclusion of participants who died from any cause, but this is additionally reported (albeit not statistically significant whereas the former was significant and this is presented as main result in the abstract). However, in our review, we used the 1 including death from any cause so no bias to our review. They also included OS and other morbidity outcomes.
Other bias	Unclear risk	Insufficient information to permit judgement.