摘要：Abstract Aim To investigate whether adding bevacizumab to neoadjuvant carboplatin-paclitaxel (CP) helps achieve optimal debulking, measured by complete resection rate (CRR) at interval debulking surgery (IDS), in patients with initially unresectable International Federation of Gynecology and Obstetrics stage IIIC/IV ovarian, tubal or peritoneal adenocarcinoma. Methods Multicentre, open-label, non-comparative phase II study. Ninety-five patients randomised (2:1) to receive four cycles of neoadjuvant CP ±3 concomitant cycles of bevacizumab 15 mg/kg (BCP) followed by IDS. Primary objective is to evaluate the CRR at IDS in the BCP group (reference CRR rate defined as 45% CRR). A stopping rule based on bevacizumab-related adverse events (AEs) of special interest was implemented. Results In the BCP group (N = 58), IDS was performed in 40 (69%) patients, of whom 85% had a complete resection. The CRR of this group was therefore 58.6% (34 patients), statistically over pre-defined 45%. The CRR in the CP group was 51.4%: 22 (60%) patients underwent IDS (85% had a complete resection). Grade ≥3 adverse events occurred in 62% of the BCP-treated patients and 63% of the CP-treated patients: mainly blood and lymphatic, gastrointestinal and vascular disorders, without more toxicity with BCP. Postoperative complications (mainly wound, infectious and gastrointestinal complications) occurred in 28% and 36% of the patients, respectively. The pre-specified safety stopping rule was not reached. Conclusion The primary objective was met as the CRR with BCP was significantly higher than the reference rate. Bevacizumab may be safely added to a preoperative program in patients deemed non-optimally resectable, whatever the final surgical decision. Bevacizumab's role in this setting should be further investigated.