Evidence strengths and limitations

Advice

It is not clear from the interim reporting on overall survival for the ICON7 trial which women with advanced ovarian cancer could benefit from bevacizumab.

The scope of this evidence summary is for advanced ovarian cancer, the definition of which is described in the NICE technology appraisal on bevacizumab in combination with paclitaxel and carboplatin for first-line treatment of advanced ovarian cancer (NICE technology appraisal guidance 284) as International Federation of Gynaecology and Obstetrics (FIGO) stages IIIB, IIIC and IV. ICON7 enrolled women with all stages of ovarian cancer, and separate analysis of women with only stages IIIB, IIIC and IV was not reported. Instead, a subgroup labelled those 'at high risk for progression', defined as FIGO stage IV disease or FIGO stage III disease and more than 1.0 cm of residual disease after debulking surgery, was reported. It is not clear if the results would be the same if women with FIGO stage III disease and less than 1.0 cm of residual disease were also included in this 'high risk subgroup'. Further subgroup analyses according to baseline characteristics and stratification factors were pre-planned but were not reported in full in the ICON7 trial. Additional data from the manufacturer's submission were presented in the NICE technology appraisal on bevacizumab. This reported that a subgroup of women with stage III suboptimally debulked ovarian cancer may have the most benefit from bevacizumab in terms of progression-free survival, with limited benefit for 2 other subgroups (that is, women with optimally debulked stage III disease and those with stage IV disease). However, the results for the subgroup of women with inoperable stage III and stage IV ovarian cancer were reported to a limited extent in the supplementary appendix of the ICON7 paper. In terms of overall survival, the only subgroup analysis reported so far is for women at 'high risk for progression'.

The current results of the Stark et al. (2013) quality of life study includes women with all stages of ovarian cancer before progression, and so this may not accurately represent quality of life for women with advanced ovarian cancer. There was a discrepancy in the number of quality of life questionnaires that Stark et al. (2013) was able to obtain compared with the number reported in the ICON7 trial. It is unclear why this occurred and what effect it may have had on the underlying results. Stark et al. (2013) reported that there was no difference in quality of life depending on stage of disease but no data were provided.

The ICON7 trial was of a reasonable size (n=1528) but there were some limitations to the open-label design that allowed clinicians and patients to know they were receiving bevacizumab. Therefore, there may have been some systematic differences in the care provided to the participants in the comparison groups other than the intervention under investigation (performance bias) in favour of bevacizumab.

Randomisation was appropriate and achieved comparable groups across disease stage, grade, histology and Eastern Cooperative Oncology Group (ECOG) performance status, although 96% of the women were Caucasian and it is unclear whether the results would be applicable to women of different ethnicity. Allocation to treatment was appropriately concealed by central randomisation using an interactive telephone or web-based system.

The vast majority of women recruited to ICON7 (94%) scored 0 or 1 for the ECOG performance status score meaning they were asymptomatic or symptomatic but completely ambulatory. They also had adequate coagulation values, and bone marrow, liver and renal function. It is not clear whether similar results would be found in women who have higher care needs or poorer organ function. It is also not clear whether those with higher care needs would be able to tolerate the raised frequency of adverse effects found to be associated with bevacizumab over standard chemotherapy.