You are here
Efficacy and safety of bevacizumab-containing neoadjuvant therapy followed by interval debulking surgery in advanced ovarian cancer: Results from the ANTHALYA trial
European Journal of Cancer 2017;70:133-142
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.
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.
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.
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.
- We present findings from the ANTHALYA study in women with advanced ovarian cancer.
- The complete resection rate at interval debulking surgery was examined.
- Bevacizumab-containing neoadjuvant therapy achieved 58.6% complete resection rate at interval debulking surgery.
- The incidence of perioperative complications with bevacizumab was acceptable.
- The role of bevacizumab in this setting should be further investigated.
Keywords: Bevacizumab, Interval debulking surgery, Neoadjuvant chemotherapy.
Primary debulking surgery followed by adjuvant platinum- and taxane-based chemotherapy is the cornerstone of treatment for ovarian cancer. In patients with advanced ovarian cancer, neoadjuvant chemotherapy followed by interval debulking surgery (IDS) offers similar overall survival (OS) benefits as primary debulking surgery and is better tolerated , , and . An EORTC study investigated debulking surgery followed by six courses of platinum-based chemotherapy versus three courses of platinum-based neoadjuvant chemotherapy followed by IDS and three further courses of platinum-based chemotherapy . Residual disease was <1 cm in 41.6% of the patients after primary debulking and in 80.6% of the patients after IDS. Median OS was similar between groups. Adverse event (AE) rates and mortality tended to be higher after primary debulking surgery. Based on these results, neoadjuvant carboplatin and paclitaxel (CP) chemotherapy is the standard treatment before IDS in patients with non-optimally resectable ovarian cancer. Regardless of which debulking strategy is used, the absence of residual tumour after surgery is the most important independent prognostic factor for OS.
Two randomised phase III trials (GOG218  and ICON7 ) in the first-line ovarian cancer setting reported that adding bevacizumab to CP, followed by maintenance bevacizumab monotherapy, was more effective than chemotherapy alone at improving response rates and progression-free survival after initial surgery. For patients with non-optimal initial surgery who are at high risk of progression, the addition of bevacizumab to CP also improved OS .
In this study, we hypothesised that adding bevacizumab to neoadjuvant CP chemotherapy would help to achieve optimal debulking rate at IDS, as measured by the complete resection rate (CRR), in patients with initially unresectable International Federation of Gynecology and Obstetrics (FIGO) stage IIIC/IV ovarian, tubal or peritoneal adenocarcinoma, without increasing the incidence of postoperative complications.
2. Patients and methods
2.1. Study design
The Avastin Neoadjuvant Therapy in patients with Advanced ovarian cancer initialLY unresectAble (ANTHALYA), multicentre, open-label, non-comparative phase II study, randomised patients 2:1 to receive four cycles of neoadjuvant CP chemotherapy with (BCP group) or without (CP group) three cycles of bevacizumab 15 mg/kg (Fig. 1). Stage IV disease, necessity of extensive/multiple bowel resection, high Fagotti score, extensive miliary peritoneal carcinomatosis,… usually classified patients as deemed unresectable. The extent of carcinomatosis was evaluated laparoscopically before neoadjuvant therapy and before IDS. IDS was scheduled 28 ± 7 d after the last neoadjuvant course. Adjuvant chemotherapy began 4–5 weeks after surgery for four cycles with bevacizumab reintroduced at cycle six (for a maximum of 26 cycles). All patients consented to serial blood sampling for further analyses. The trial was conducted at study sites with teams experienced in the multidisciplinary treatment of ovarian cancer.
ANTHALYA study design. AUC, area under the curve; C, cycle; CP, carboplatin-paclitaxel chemotherapy; BCP, bevacizumab plus CP; FIGO, International Federation of Gynecology and Obstetrics; IDS, interval debulking surgery.
The study was approved by the independent ethics committee of each participating centre and conducted in accordance with Good Clinical Practice guidelines and the Declaration of Helsinki. Written informed consent was obtained from all patients before screening. The trial is registered (clinicaltrials.gov: NCT01739218).
Eligible patients were women aged ≥18 years with histologically confirmed, chemotherapy-naïve, high-risk FIGO stage IIIC/IV epithelial ovarian carcinoma, fallopian tube carcinoma or primary peritoneal carcinoma (deemed ineligible for primary complete debulking surgery by a surgeon experienced in the management of ovarian cancer during a laparoscopic procedure); an Eastern Cooperative Oncology Group performance status (ECOG-PS) 0–2; life expectancy ≥3 months; and eligible for CP chemotherapy. To ensure somehow a degree of inter-tumour homogeneity, women with ovarian tumour with low malignant potential, mucinous and clear cell ovarian cancer or carcinosarcoma, and those who had undergone previous systemic therapy for ovarian cancer, prior radiotherapy or major surgery were excluded (Appendix A details other exclusion criteria).
The treatment period for each patient was approximately 21 months. Each cycle lasted 3 weeks. In the neoadjuvant period, all patients received area under the curve 5 mg/mL/min of carboplatin and 175 mg/m2 of paclitaxel intravenously on Day 1 (cycles 1–4); patients in the BCP group received 15 mg/kg of bevacizumab administered intravenously every 3 weeks (q3w) on Day 1 (cycles 1–3) in addition to CP. In the adjuvant period, all patients received area under the curve 5 mg/mL/min of carboplatin and 175 mg/m2 of paclitaxel intravenously q3w on Day 1 (cycles 5–8), and all patients received 15 mg/kg of bevacizumab intravenously q3w on Day 1 (cycles 6–26). Bevacizumab, carboplatin, and paclitaxel were administered according to the protocol or until disease progression, the occurrence of an unacceptable toxicity, or at the patient's or treating physician's request (Fig. 1).
2.4. Study objectives
The primary objective was to evaluate the benefit of neoadjuvant bevacizumab and chemotherapy assessed by the CRR at IDS (primary endpoint). The secondary objective was to evaluate the safety profile of bevacizumab added to neoadjuvant and adjuvant chemotherapy. The final analysis of the primary endpoint was conducted at the interim analysis of the trial after all patients had completed the post-IDS follow-up assessment.
2.5. Study assessments
Physical examination, vital signs, laboratory safety assessments, and recording of AEs according to the National Cancer Institute's cancer toxicity criteria for AEs (NCI-CTCAE; v4.03), were performed by the Investigator. Additional safety assessments (in line with local standard of care or those that were symptom-directed) were undertaken at the discretion of the treating physician. Verbatim reported AE terms were coded into standardised terms according to the industry standard Medical Dictionary for Regulatory Activities. Events were graded according to the National Cancer Institute Common Terminology Criteria for Adverse Events (NCI-CTCAE) version 4.03. Multiple occurrences of each event per patient were counted once at the maximum severity. Left ventricular ejection fraction was assessed at baseline and in case of symptomatic cardiac failure in which case it was repeated every 3 weeks until the resolution or stabilisation of the event. An independent safety data monitoring board regularly reviewed all safety data and a stopping rule was implemented during the neoadjuvant period, at IDS, and during 30 d following IDS in the BCP group if ≥20% of the patients experienced severe surgical or general complications related to bevacizumab (described in Appendix B).
2.6. Statistical method
The rate of complete resection at IDS with 95% one-sided CI was assessed using Pearson-Clopper's exact method . Complete resection was defined as the removal of all macroscopic residual tumour (completeness of cytoreduction [CC] score of zero). Patients not having IDS performed were considered as failure.
In the BCP group, the sample size calculation was based on a one-step phase II study using the Fleming A'Hern method . Vergote et al. had shown a CRR of 46.3% (51.2% in patients with IDS) for patients with FIGO stage IIIC/IV ovarian cancer after neoadjuvant platinum-based chemotherapy. Therefore a CRR of 45% (null hypothesis: P ≤ 45%) was considered insufficient evidence for the benefit of adding bevacizumab to chemotherapy (BCP group) in the neoadjuvant setting and a CRR of 65% (alternative hypothesis: P ≥ 65%) was considered as the minimal target of clinical efficacy. Accepting a type 1 error α = 0.05 one-sided and type 2 error β = 0.10, 54 evaluable patients were required in the BCP group. Proof of potential efficacy was considered if ≥31 patients had a CC score of 0. As the CP group was considered the ‘calibration group’ for the CRR and no comparison between treatments was expected, we used a 2:1 randomisation design. Therefore, the calibration group was planned to include 27 patients treated without bevacizumab in the neoadjuvant setting. Assuming a 10% drop out rate, a total of 90 patients were required: 60 in the BCP group and 30 in CP group. As 58 patients were ultimately included in BCP group, 33 patients with a CC score of 0 were required to reject the null hypothesis (Fig. 2A), with an actual type 1 error α = 0.046 one-sided and type 2 error β = 0.08 (power of 92%) .
(A) Statistical hypothesis and sample size for the primary criterion (Fleming A'Hern design). Four different situation are illustrated, two which indicate potential proof of efficacy (yellow diamonds and bars) and two which show insufficient evidence for efficacy (blue diamonds and bars; B) Results of the primary endpoint and sensitivity analyses in the mITT population who received at least two cycles of bevacizumab and the population of patients for whom IDS was performed. CP, carboplatin-paclitaxel chemotherapy; CR(R), complete resection (rate); BCP, bevacizumab plus CP; H0, null hypothesis; H1, alternative hypothesis; IDS, interval debulking surgery; mITT, modified intent-to-treat.
Efficacy and safety analyzes were performed in patients having received at least one administration of treatment. Efficacy analyses were performed in the modified intention-to-treat (mITT) population according to the assigned group at randomisation. Safety analyses were performed according to the actual treatment received (safety population). The primary endpoint was also assessed in two sensitivity analyses in the mITT population for whom IDS was performed, and the mITT population who received at least two cycles of bevacizumab.
Between January 2013 and June 2014, 205 patients were screened at 15 French study sites, of which 95 were included in the mITT population (37 [39%] in the CP group and 58 [61%] in the BCP group; Fig. 3). Fifty-five (58%) patients received at least one treatment with bevacizumab and chemotherapy and 40 (42%) patients received only chemotherapy (safety populations). Baseline patient and disease characteristics were balanced between groups (Table 1). Median age was 63 years (range 33–87 years), 94% of the patients had ECOG-PS 0/1, and 70% and 30% had FIGO stage IIIC and IV tumours, respectively. More than 90% of the patients had high grade and serous histological subtype tumours.
CONSORT flow diagram. CP, carboplatin-paclitaxel chemotherapy; BCP, bevacizumab plus CP; mITT, modified intent-to-treat; SAF, safety population.
Baseline and disease characteristics (mITT).
|Carboplatin + paclitaxel (CP, n = 37)||Bevacizumab + carboplatin + paclitaxel (BCP, n = 58)||Total (n = 95)|
|Mean ± SD||61 ± 10||62 ± 10||62 ± 10|
|Median (range)||63 (39–79)||63 (33–87)||63 (33–87)|
|ECOG performance status, n (%)||1 missing|
|0 – Fully active||15 (42)||21 (36)||36 (38)|
|1 – Ambulatory||18 (50)||35 (60)||53 (56)|
|2 – Restricted||3 (8)||2 (4)||5 (6)|
|FIGO stage, n (%)|
|IIIC||24 (65)||43 (74)||67 (70)|
|IV||13 (35)||15 (26)||28 (30)|
|Origin of cancer, n (%)||4 missing||5 missing||9 missing|
|Ovary (epithelial)||30 (91)||52 (98)||82 (95)|
|Primary peritoneal||3 (9)||1 (2)||4 (5)|
|Histological subtype, n (%)|
|Serous||36 (97)||54 (93)||90 (95)a|
|Endometrioid||1 (2)||1 (1)|
|Serous/endometrioid||1 (2)||1 (1)|
|Otherb||1 (3)||2 (3)||3 (3)|
|Histological grade, n (%)||1 missing||5 missing||6 missing|
|Well differentiated (Low grade)||5 (14)||5 (6)|
|Poorly differentiated (High grade)c||31 (86)||53 (100)||84 (94)|
|≥1 target lesion ≥50 mm, n (%)||26 (70)||34 (59)||60 (63)|
|≥1 target lesion ≥100 mm, n (%)||11 (30)||16 (28)||27 (28)|
|Mean ± SD, U/mL||3838 ± 9331||3942 ± 7003||3902 ± 7943|
|Median (range), U/mL||1281 (44–54,152)||1008 (128–37,537)||1045 (44–54,152)|
|≤500 U/mL, n (%)||12 (32)||16 (28)||28 (30)|
|>500 U/mL, n (%)||25 (68)||42 (72)||67 (70)|
a Eighty-one of 90 patients (90%) had high grade serous tumour (one patient had serous/endometrioid tumour).
b Two patients with ‘adenocarcinoma not specified’ and one patient with mucinous tumour (initially unclassified).
c Nine patients were grade 2, 75 patients were grade 3.
Abbreviations: CA-125, Cancer Antigen 125; ECOG, Eastern Cooperative Oncology Group; FIGO, International Federation of Gynecology and Obstetrics; mITT, modified intent-to-treat; SD, standard deviation.
Missing values are excluded from the calculation of percentages.
Mean neoadjuvant treatment exposure was 2.7 months (standard deviation [SD]: 0.5 months) and was balanced between groups. Mean bevacizumab exposure in the BCP group was 2.0 months (SD 0.3 months). Overall, 90% of the patients received all four planned neoadjuvant CP treatment cycles and 87% of the patients in the BCP group received all three cycles of bevacizumab. Eight (22%) patients from the CP group and 10 (17%) patients from the BCP group withdrew prematurely from the study (Fig. 3).
Interval debulking surgery was performed in 62 patients (40 [69%] in the BCP group and 22 [60%] in the CP group; Table 2). The mean times from first study treatment administration to IDS laparoscopy and surgery were comparable between the two groups. Unresectable tumour assessed by laparoscopy before IDS was the most common reason for surgery not being performed in both groups (Table 2).
Interval debulking surgery results (mITT).
|Carboplatin + paclitaxel (CP, n = 37)||Bevacizumab + carboplatin + paclitaxel (BCP, n = 58)||Total (n = 95)|
|IDS performed, n (%)|
|Yes||22 (60)||40 (69)||62 (65)|
|No||15 (41)||18 (31)||33 (35)|
|Time from first study treatment administration to IDS laparoscopy|
|Mean ± SD, days||95.0 ± 7.5||92.6 ± 7.6||93.5 ± 7.6|
|Time from date of IDS laparoscopy to surgery|
|Mean ± SD, days||1.0 ± 3.4||1.2 ± 3.4||1.1 ± 3.3|
|Reason IDS not performed, n|
|IDS procedures performed, n (%)a|
|Omentectomy||22 (100)||40 (100)||62 (100)|
|Total hysterectomy with bilateral salpingo oophorectomy||22 (100)||38 (95)||60 (97)|
|Pelvic lymphadenectomy||18 (82)||37 (93)||55 (89)|
|Para aortic lymphadenectomy||17 (77)||37 (93)||54 (87)|
|Pelvic peritoneum stripping||14 (64)||27 (68)||41 (66)|
|Diaphragm stripping/resection||12 (55)||23 (58)||35 (56)|
|Abdominal peritoneum stripping||9 (41)||13 (33)||22 (35)|
|Rectosigmoidectomy anastomosis||8 (36)||14 (35)||22 (35)|
|Large bowel resection||6 (27)||5 (13)||11 (18)|
|Splenectomy||4 (18)||6 (15)||10 (16)|
|Liver resection/s||1 (5)||2 (5)||3 (5)|
|Small bowel resection/s||2 (9)||2 (5)||4 (6)|
|Diverting stomia||3 (14)||2 (5)||5 (8)|
|Operative time (h)a|
|Mean ± SD||5.1 ± 1.8||5.9 ± 1.8||5.6 ± 1.8|
|Median (range)||5.0 (1.6–8.0)||5.5 (2.0–10.0)||5.3 (1.6–10.0)|
|Missing data, n||3||2||5|
|PCI total score|
|Mean ± SD||9 ± 6||10 ± 7||10 ± 7|
|Median (range)||10 (0–20)||11 (0–27)||10 (0–27)|
|Missing data, n||1||1|
|CC score, n (%)|
|No disease (CC = 0)||19 (51.4)||34 (58.6)||53 (55.8)|
|Present, <0.25 cm (CC = 1)||2 (5.4)||3 (5.2)||5 (5.2)|
|0.25 cm–2.5 cm (CC = 2)||1 (2.7)||1 (1.7)||2 (2.1)|
|>2.5 cm (CC = 3)||2 (3.4)||2 (2.1)|
|Missing data (IDS not performed)||15 (40.5)||18 (31.0)||33 (34.7)|
a Percentage are out of the number of patients in whom interval debulking surgery (IDS) was performed (carboplatin and paclitaxel, n = 22; bevacizumab, paclitaxel and carboplatin, n = 40; total, n = 62).
Abbreviations: CC, completeness of cytoreduction score; PCI, peritoneal cancer index; SD, standard deviation; mITT, modified intent-to-treat; IDS, interval debulking surgery.
The CRR at IDS in BCP-treated patients was 58.6% (34 patients) with a one-sided lower 95% confidence interval of 47.0% (Fig. 2B). Consequently, the CRR was significantly higher than the reference rate. In a sensitivity analysis limited to patients who underwent IDS (n = 40), the CRR at IDS in BCP-treated patients was 85.5%. In patients from the mITT population who had at least two cycles of bevacizumab (n = 52), the CRR was 63.5% and in the safety population the CRR was 61.8%. The CRR in CP-treated patients in the mITT population was 51.4% and 86.4% in CP-treated patients who underwent IDS (n = 22).
Overall, 94 (99%) patients experienced at least one AE before or during the IDS period (Table 3). Serious AEs (SAEs) were more frequent in the CP group versus the BCP group. The most common treatment-emergent SAEs in the CP and BCP groups were gastrointestinal disorders (13% versus 7%, respectively), infections and infestations (8% versus 9%, respectively), and respiratory disorders (10% versus 2%, respectively). Other SAEs seen in BCP-treated patients were: blood and lymphatic disorders (5%), vascular disorders (4%), general disorders (2%) and injury/procedural complications (2%). In the overall population, 59 (62%) patients experienced one AE of grade ≥3; these were blood and lymphatic (28%), gastrointestinal (13%), vascular (11%), and general disorders (10%), and infections and infestations (6%). One AE lead to death in the CP group; a bilateral pneumopathy after the third cycle.
Summary of adverse events (neoadjuvant and IDS periods combined; safety population).
|n, (%)||Carboplatin + paclitaxel (n = 40)*||Bevacizumab + carboplatin + paclitaxel (n = 55)a||Total (n = 95)|
|At least one AE||39 (98)||55 (100)||94 (99)|
|At least one serious AE (TEAEs)||15 (38)||14 (25)||29 (31)|
|Gastrointestinal disorders||5 (13)||4 (7)||9 (10)|
|Infections and infestations||3 (8)||5 (9)||8 (8)|
|Respiratory disorders||4 (10)||1 (2)||5 (5)|
|Blood and lymphatic disorders||1 (3)||3 (5)||4 (4)|
|General disorders||2 (5)||1 (2)||3 (3)|
|Metabolism disorders||2 (5)||–||2 (2)|
|Vascular disorders||–||2 (4)||2 (2)|
|Injury/procedural complications||–||1 (2)||1 (1)|
|Neoplasms benign, malignant||1 (3)||–||1 (1)|
|Psychiatric disorders||1 (3)||–||1 (1)|
|Renal and urinary disorders||1 (3)||–||1 (1)|
|Reproductive system disorders||1 (3)||–||1 (1)|
|At least one AE grade ≥3||25 (63)||34 (62)||59 (62)|
|Main AEs grade ≥3b|
|Blood and lymphatic disorders||11 (28)||16 (29)||27 (28)|
|Gastrointestinal disorders||7 (18)||5 (9)||12 (13)|
|Vascular disorders||1 (3)||9 (16)||10 (11)|
|General disorders||5 (13)||4 (7)||9 (10)|
|Infections and infestations||3 (8)||3 (5)||6 (6)|
|AE leading to death||1 (3)||–||1 (1%)|
|≥1 Bevacizumab-related AE||–||46 (84)||–|
|≥1 Paclitaxel-related AE||38 (95)||54 (98)||92 (97)|
|≥1 Carboplatin-related AE||36 (90)||50 (91)||86 (91)|
a N numbers in this table relate to actual treatment received and not study groups (i.e. if a patient was randomised to bevacizumab and carboplatin, but only actually received chemotherapy alone they would fall into the carboplatin group in the safety population).
b With incidence ≥5%.
Abbreviations: AE, adverse event; SAE, serious adverse event; TEAE, treatment-emergent adverse event; IDS, interval debulking surgery.
Of the 62 patients who had IDS, 32 (52%) patients had at least one abnormal examination at 30 d post-IDS (both arms combined): 42% were abdominal, 18% gynaecological, and 8% gastrointestinal (Table 4). The overall mean (±SD) duration of hospitalisation was 13 ± 7 d and the mean (±SD) duration of hospitalisation in the intensive care unit was 3.5 ± 3.2 d. In total, 27 (44%) patients needed blood transfusions following surgery. Postoperative complications occurred in 19 (31%) patients. Three (5.5%) patients in the BCP group experienced four surgical complications (all grade 4 haemorrhages) and one general complication (grade 3 thromboembolic event) of special interest to bevacizumab.
Safety summary for patients with surgery performed (neoadjuvant and IDS periods combined).
|Surgery performed||Carboplatin + paclitaxel (n = 22)||Bevacizumab + carboplatin + paclitaxel (n = 40)||Total (n = 62)|
|Examination at 30 d post-IDS, n (%)|
|At least one abnormal examination||13 (59)||19 (48)||32 (52)|
|At least one abnormal abdominal examination||11 (50)||15 (38)||26 (42)|
|At least one abnormal gynaecological examination||4 (18)||7 (18)||11 (18)|
|At least one abnormal gastrointestinal examination||2 (9)||3 (8)||5 (8)|
|Duration of hospitalisation in the ICU|
|Mean ± SD, days||4.0 ± 4.4||3.2 ± 2.3||3.5 ± 3.2|
|Median (range), days||3.0 (0–13)||3.0 (0–8)||3.0 (0–13)|
|No. with missing data, n||1||–||1|
|Total duration of hospitalisation|
|Mean ± SD, days||15 ± 10||12 ± 4||13 ± 7|
|Median (range), days||11 (7–51)||11 (7–24)||11 (7–51)|
|No. with missing data, n||1||–||1|
|Blood transfusion, n (%)||9 (41)||18 (45)||27 (44)|
|At least one postoperative complication, n (%)||8 (36)||11 (28)||19 (31)|
|Wound complications||2 (9)||6 (15)||8 (13)|
|Infectious complications||3 (14)||4 (10)||7 (11)|
|Gastrointestinal complications||1 (5)||4 (10)||5 (8)|
|Urinary complications||1 (5)||2 (5)||3 (5)|
|Lymphatic complications||2 (5)||2 (3)|
|Pulmonary complications||1 (5)||1 (3)||2 (3)|
|Bleeding complications||1 (3)||1 (2)|
|Other complications||2 (9)||1 (3)||3 (5)|
Abbreviations: ICU, intensive care unit; IDS, interval debulking surgery; SD, standard deviation.
This randomised study investigated neoadjuvant CP chemotherapy with and without bevacizumab followed by IDS in women with advanced ovarian or primary peritoneal cancer deemed unresectable in primary debulking surgery. We included a calibration group with conventional CP as we believed that the surgical effort would be the main driver of complete resection.
Neoadjuvant chemotherapy with bevacizumab achieved a CRR at IDS of 58.6% with a lower confidence limit of 47.0%, significantly higher than the threshold of 45% established based on the CRR rate reported in the EORTC study of neoadjuvant platinum-based chemotherapy followed by IDS (46.3%) . Therefore, the primary objective of our study was met indicating proof of potential efficacy of the BCP combination. The CRR was high in patients who underwent IDS after neoadjuvant therapy (85.5%). The complication rate was acceptable, demonstrating the feasibility of administering bevacizumab in the neoadjuvant setting.
All patients had FIGO stage IIIC/IV ovarian cancer. At baseline, metastatic lesions were >10 and >5 cm in diameter in 28% and 63% of patients, respectively and were in line with the EORTC and CHORUS study populations  and , although the frequency of serous tumours (95%) was greater than in the EORTC study (58%). These broad similarities in patient characteristics justify the use of the CRR from the EORTC study as the reference rate in our statistical hypothesis.
Patients in this study were enrolled in highly specialised centres. This may explain the relatively high CRR (51.4%) also observed in the CP group. While surgical effort was heterogeneous in the EORTC and CHORUS studies and was considered a limitation  and , surgical interventions in our study were typically more aggressive. We performed twice the number of bowel resections and four times the number of supra-colic surgeries and fewer patients (65%) were considered eligible for IDS after neoadjuvant chemotherapy than in the EORTC study (88%) . This suggests the decision to perform IDS is more adequately taken by trained surgeons working in experienced centres. Our good results may also partly be due to a highly selected, homogeneous population presenting mainly with high grade, chemosensitive ovarian carcinoma. This study is a phase II study and warrants additional studies to answer specific points as the centre effect.
Anti-angiogenic therapy with bevacizumab in the neoadjuvant setting could potentially increase the rate of operative and postoperative complications. The rate of pre-specified complications in bevacizumab-treated patients (5.5% overall) did not exceed the pre-defined 20% probability of toxicity threshold at any point during the study (thus the stopping rule for toxicity was not implemented) and was similar to that observed in the EORTC study, where 4.1% of the patients experienced grade 3/4 haemorrhage after three courses of neoadjuvant chemotherapy and IDS . Importantly, the BCP regimen also did not increase the risk of perioperative complications in patients who ultimately did not undergo IDS.
Notably, patients in our study received four courses of neoadjuvant chemotherapy versus three in the EORTC and the CHORUS studies. Almost all patients in the CP group completed the four courses with similar complication rates as the BCP group. The addition of bevacizumab did not increase the rate of AEs or SAEs after neoadjuvant therapy compared with chemotherapy alone. Indeed, the serious adverse event rate in BCP-treated patients (25%) was comparable to the EORTC study (29%) . The open-label design of this study had the potential to influence perioperative decisions, but, the expertise of the surgeons and objective evaluation of carcinomatosis using pre-specified photographs may have limited any potential bias.
In conclusion, adding bevacizumab to neoadjuvant chemotherapy achieved an encouraging CRR at IDS in patients with initially unresectable FIGO stage IIIC/IV ovarian, tubal, or peritoneal adenocarcinoma. Our data suggest that bevacizumab may be safely added to a preoperative program in patients deemed non-optimally resectable, whatever the final surgical decision. The role of bevacizumab in this setting should be further investigated.
This work was supported by F. Hoffmann-La Roche Ltd.
Conception and design: R.R., F.S., E.L., V.F., Y.G., S.Z., P.C., F.J.; collection and assembly of data: R.R., S.G., F.S., P.-E.C., Y.G., A.F., E.K., S.M.-F., R.F., P.F., A.L., F.L., E.C., P.C., F.J.; data analysis and interpretation: R.R., F.S., C.P., Y.G., J.D., S.Z., P.C., F.J.; manuscript writing: all authors; final approval of the manuscript: all authors; provision of study material or patients: C.P., P.C.
Conflict of interest statement
R.R. declares honoraria and consulting or advisory roles from Roche and NanoString Technologies; S.G. declares a consulting or advisory role from Roche; Frédéric Selle declares consulting or advisory roles, expert testimony support, and travel and accommodation expenses from Roche, MSD, and PharmaMar; A.F. declares a consulting or advisory role from Roche and AstraZeneca; P.-E.C. declares a leadership role from Roche; E.K. declares, travel and accommodation expenses from Roche, PharmaMar, Amgen, and Novartis; P.F. declares travel and accommodation expenses from Roche; Y.G. is an employee of Roche; J.D. declares a consulting or advisory role and travel and accommodation expenses from Roche; S.Z. declares a consulting or advisory role from Roche and Servier; P.C. declares honoraria from Roche, Novartis, Pfizer, a consulting or advisory role from Novartis and Pfizer, and research funding from Pfizer and NanoString Technologies; F.J. declares a consulting or advisory role from Roche, Pfizer, Novartis, and Sanofi, research funding from Astellas Pharma, expert testimony support from Roche, Sanofi, Pfizer and Novartis, and travel and accommodation expenses from Roche, Novartis and Janssen. All other authors have no conflicts to report.
This study and editorial support for the preparation of this manuscript were funded by F. Hoffmann-La Roche Ltd.
The authors thank the patients and their families for their participation in this study and the staff at the study sites. Support for third-party writing assistance for this article, furnished by Jamie Ashman and Bilal Bham, was provided by Prism Ideas.
-  R.J. Morgan Jr., R.D. Alvarez, D.K. Armstrong, B. Boston, R.A. Burger, L.M. Chen, et al. Epithelial ovarian cancer. J Natl Compr Canc Netw. 2011;9:82-113
-  I. Vergote, C.G. Trope, F. Amant, G.B. Kristensen, T. Ehlen, N. Johnson, et al. Neoadjuvant chemotherapy or primary surgery in stage IIIC or IV ovarian cancer. N Engl J Med. 2010;363:943-953 Crossref
-  S. Kehoe, J. Hook, M. Nankivell, G.C. Jayson, H. Kitchener, T. Lopes, et al. Primary chemotherapy versus primary surgery for newly diagnosed advanced ovarian cancer (CHORUS): an open-label, randomised, controlled, non-inferiority trial. Lancet. 2015;386:249-257 Crossref
-  R.A. Burger, M.F. Brady, M.A. Bookman, G.F. Fleming, B.J. Monk, H. Huang, et al. Incorporation of bevacizumab in the primary treatment of ovarian cancer. N Engl J Med. 2011;365:2473-2483 Crossref
-  T.J. Perren, A.M. Swart, J. Pfisterer, J.A. Ledermann, E. Pujade-Lauraine, G. Kristensen, et al. A phase 3 trial of bevacizumab in ovarian cancer. N Engl J Med. 2011;365:2484-2496 Crossref
-  C.J. Clopper, E.S. Pearson. The use of confidence or fiducial limits illustrated in the case of the binomial. Biometrika. 1934;26:404-413 Crossref
-  R.P. A'Hern. Sample size tables for exact single-stage phase II designs. Stat Med. 2001;20:859-866 Crossref
-  I. Khan, S.J. Sarker, A. Hackshaw. Smaller sample sizes for phase II trials based on exact tests with actual error rates by trading-off their nominal levels of significance and power. Br J Cancer. 2012;107:1801-1809 Crossref
a Institut Curie, Saint-Cloud-Paris, UNiversité Versailles-Saint-Quentin, France
b Institut Gustave Roussy, Villejuif, France
c GHU-Est Tenon, Paris, France
d Institut Paoli-Calmettes, Marseille, France
e Institut Bergonié, Bordeaux, France
f Centre Jean Perrin, Clermont-Ferrand, France
g Institut régional du Cancer de Montpellier, Val d'Aurelle, Montpellier, France
h CHU Besançon, Besançon, France
i Centre François Baclesse, Caen, France
j CHU Amiens, Amiens, France
k Centre Antoine Lacassagne, Nice, France
l Centre Oscar Lambret, Lille, France
m European Hospital George Pompidou, Paris, France
n Roche, Boulogne-Billancourt, France
o ITM stat for Roche, Boulogne-Billancourt, France
p Centre de Recherche des Cordeliers, Université Paris 5, Université Paris 6, Paris, France
∗ Corresponding author: Département de Chirurgie oncologique, Institut Curie, rue Dailly, 92210 Saint-Cloud 35, France.
∗∗ Corresponding author: Département d'oncologie médicale, Centre François Baclesse, Université Basse Normandie, Avenue du Général Harris, 14000, Caen, France.
☆ This work was presented in part at the 2015 Annual Meeting of the European Society for Medical Oncology (Selle F, et al. ESMO 2015 Poster presentation (Abstract 2766)) and the 2015 meeting Annual Meeting of the European Society of Gynaecological Oncology (Gouy S, et al. ESGO poster presentation).
© 2016 Published by Elsevier B.V.
© 2016 The Authors. Published by Elsevier Ltd under the CC BY-NC-ND license.
No responsibility is assumed by Elsevier, its licensors or associates for any
injury and/or damage to persons or property as a matter of products
liability, negligence or otherwise, or from any use or operation of any
methods, products, instructions, or ideas contained in the material herein.
Because of rapid advances in the medical sciences, in particular, independent
verification of diagnoses and drug dosages should be made.
This e-print is distributed with the support of Roche.
Elsevier España, S.L.U.
(A member of Elsevier)
Av. Josep Tarradellas, 20-30
Tel.: 932 000 711
Fax: 932 091 136