Polterauer 2012.

Study characteristics
Methods	Prospective, multicentre study (5 specialised European centres for gynaecologic oncology) Women enrolment between February 2005 and December 2008
Participants	226 women with epithelial ovarian cancer FIGO Stages IIA to IV in whom radical cytoreductive surgery was performed and standard chemotherapy with paclitaxel and carboplatin was applied. Women having received neoadjuvant chemotherapy followed by interval debulking were excluded Mean age 57.5 year (SD 11.9) FIGO stages II, III and IV: 15 (6.6%), 174 (76.9%) and 37 (16.4%); FIGO stages IIIC and IV: 198 women (87.6%) Histological type serous/other: 194/32 NMRD: 69.4% SVRD (≤ 1 cm): 87.2% (NB: this category also includes NMRD) Austria
Residual disease details	Residual disease was defined as: Any RD (SVRD (≤ 1 cm) or LVRD (> 1 cm) Complete debulking (NMRD)
Outcomes	3‐year OS (unadjusted) with NMRD: 72.4%; minimal RD: 65.8%; gross RD: 45.2% Subgroup analysis of stages IIIC and IV: 3‐year OS (unadjusted) with NMRD 69.7% (SE 5.3%); any RD 53.6% (SE 8.3%) (P = 0.003) HR (apparently for ‘Any RD’ vs ‘No RD’, adjusted for FIGO‐stage, histological grade, histological type and age) 1.4 (95% CI 1.0 to 2.1) “Multivariable survival analysis revealed residual tumour size (p=0.04) and older women age (p =0.02) as independent prognosticators for impaired overall survival. Complete cytoreduction was predictive for a higher rate of treatment response (p=0.001) and was associated with prolonged progression‐free and overall survival (p<0.001 and p=0.001).” HR for PFS (apparently for ‘Any RD’ vs ‘NMRD’, adjusted for FIGO stage, histological grade, histological type and age) 1.6 (95% CI 1.3 to 2.1) Univariate survival analysis of categorical variables by the log‐rank test. Multiple forward stepwise Cox regression analysis.
Risk of bias (QUIPS)	1. Study participation (a‐f): low risk Adequate number of participants and description of target population. Baseline characteristics, eligibility criteria, sampling frame and period/place study took place presented clearly. 2. Study attrition (a‐e): unclear risk Unclear if patients with incomplete follow‐up were excluded before arriving at the stated sample size. Insufficient information to permit judgement. 3. Prognostic factor measurement (a‐f): unclear risk Valid and reliable measurement of RD. Multicentre design may introduce heterogeneity in measurement of RD Outcome level assessment: Outcome: overall survival 4. Outcome measurement (a‐c): low risk Valid and reliable measurement of outcome 5. Adjustment for other prognostic factors (a‐g): low risk Cohort was recruited with objective to identify and verify clinical and molecular prognostic/predictive factors in ovarian cancer. Possible confounding prognostic factors would also have been included in study. Multivariate model for OS adjusted for FIGO stage, histological grade, histology subtype and age 6. Statistical analysis and reporting (a‐d): unclear risk No conceptual framework; variable selection criteria for multivariate analysis unstated. Outcome: progression‐free survival 4. Outcome measurement (a‐c): low risk Valid and reliable measurement of outcome 5. Adjustment for other prognostic factors (a‐g): low risk Cohort was recruited with objective to identify and verify clinical and molecular prognostic/predictive factors in ovarian cancer. Possible confounding prognostic factors would also have been included in study. Multivariate models for PFS adjusted for FIGO stage, histological grade, histology subtype and age 6. Statistical analysis and reporting (a‐d): unclear risk No conceptual framework; variable selection criteria for multivariate analysis unstated
Notes	Source of funding: the European commission (FP6 Specific Targeted Research or Innovation Project) Declaration of interest: none declared Median follow‐up: 25.0 months (range: 1 to 49) Retrospective non‐randomised study. Blinding not reported (but not applicable). Adjusted HRs are derived from a prognostic model. No details on how modelling was performed, but this seems to have been done based on significance testing (and not on including putative confounders in the analysis, irrespective of statistical significance). Women and disease characteristics not reported according to debulking status. NB: possible overlap with Hofstetter 2013.