Davidson 2019.

Study characteristics
Methods	Multicentre retrospective and single‐centre prospective cohort Prospective data collection was to explore minimally‐invasive surgery following NACT
Participants	All participants received NACT followed by interval debulking surgery for an advanced ovarian, fallopian tube or primary peritoneal cancer At Duke, information on women receiving NACT was collected retrospectively between January 2000 and September 2013 and prospectively (with subject informed consent after October 2013). At the Ohio State University and the University of Oklahoma, subjects were identified retrospectively. Women at all 3 institutions were included if they were diagnosed prior to 30 June 2017 to allow for at least 12 months of post‐diagnosis follow‐up. N = 282 participants with advanced ovarian, fallopian tube or primary peritoneal cancer Median age: 63.9 (range: 34.1 to 84.8) Race: Caucasian – 229 (81.2%) FIGO: IIIC – 114 (40.4%); IV – 101 (35.8%); presumed AOC – 57 (20.2%); unknown stage – 10 (3.5%) Histology: serous – 227 (80.5%); undifferentiated – 4 (1.5%); endometrioid – 1 (0.4%); mixed – 5 (1.8%); clear cell – 5 (1.8%); NOS – 21 (7.5%); unknown – 15 (5.3%) Ascites: 88 (31.2%)
Residual disease details	Carboplatin and paclitaxel: 87.2% Median NACT cycles: 4 (range: 2 to 10) Indication for NACT: disease volume ‐ 80 (28.4%); comorbidities ‐ 19 (6.7%); both ‐ 29 (10.3%) Median surgery duration, minutes: 194 (range: 45 to 459) Determination of resectability: diagnostic laparoscopy – 78 (27.7%); clinical exam/imaging – 196 (69.5%); unknown – 8 (2.8%) Surgical approach at IDS: laparoscopy only – 27 (9.6%); laparoscopy converted to laparotomy – 26 (9.2%); exploratory laparotomy only – 221 (78.4%) Median surgical complexity score: 2 Surgical complexity score: Low (0 to 3): 193 (68.4%) Moderate (4 to 7): 80 (28.4%) Complex (8 to 9): 9 (3.2%) Intraoperative complications: 23 women (8.7%). Bowel injuries (including serosal injuries) (n = 16); bladder (n = 6); vascular injuries (n = 6). Postoperative complications were seen in 62 women (22%) prior to hospital discharge and included: Ileus/small bowel obstruction: 26 (9.2%) Pulmonary issues: 12 (4.3%) Altered mental state: 10 (3.6%) Wound cellulitis/haematoma, UTI and cardiac concerns: 5 (1.8%) Re‐operation: 1 (0.4%) 32 (11.3%) experienced complications after discharge and within 30 days of surgery 18 (6.4%) re‐admitted. Data for reasons for re‐admission available for n = 7: infectious complications (n = 3), gastrointestinal dysmotility (n = 3), acute renal failure related to urinary retention (n = 1). 2 required re‐operation during re‐admission. 1 underwent re‐operation in outpatient setting for wound debridement. Optimal cytoreduction defined using two methods: NMRD (described in study as RD0) (n = 165/271; 60.9%) or SVRD ≤ 1 (n = 228/271; 84.1%). The latter definition is used in multivariable analysis. NMRD: 165 (60.9%) SVRD: 63 (23.2%) LVRD 1 cm to 2 cm: 6 (2.2%) LVRD > 2 cm: 37 (13.7%) Missing (n = 11)
Outcomes	Disease‐specific overall survival (DSS) defined as time from completion of adjuvant chemotherapy to death due to cancer Median disease‐specific overall survival (DSS): 24.8 months Median DSS in RD ≤ 1: 25 months Median DSS in RD > 1: 23.5 months Multivariable Cox PH for DSS adjusted for ASA score, age, SCS and major morbidity: LVRD > 1 cm (vs SVRD ≤ 1 cm): HR 1.7 (95% CI 1.1 to 2.8), P = 0.03 No deaths within 30 days of IDS
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 Adequate cut‐off for residual disease used. Multicentre design may introduce heterogeneity in measurement of RD. Outcome level assessment: Outcome: disease‐specific survival 4. Outcome measurement (a‐c): high risk Overall survival not used as outcome. Rather, disease‐specific survival was used. Disease‐specific survival (DSS) defined as time from completion of adjuvant chemotherapy to death due to cancer. 5. Adjustment for other prognostic factors (a‐g): high risk Age arbitrarily categorised; ASA score dichotomised. Model predicting DSS adjusted for ASA score, age, SCS, presence of major morbidity. Few of these were deemed important prognostic factors. 6. Statistical analysis and reporting (a‐d): unclear risk No conceptual framework; data driven based on P values of univariate associations Outcome: progression‐free survival Not reported
Notes	—