Ovarian cancer (stage I-IV) - Chemotherapy

A range of regimens have been investigated, most of which were compared with the combination of carboplatin and paclitaxel (considered as standard first-line chemotherapy). The type of comparison was usually a substitution of different agents or the addition of a third agent. Various platinum/taxane combinations were compared. Additional agents investigated included anthracyclines (doxorubicin, epirubicin), antimetabolites (gemcitabine), and topoisomerase inhibitors (topotecan, irinotecan.)14-43, 72

Almost all the regimens failed to demonstrate an overall or progression-free survival benefit compared with standard chemotherapy (most often platinum/taxane combination). Trials which did show survival differences were either in specific patient populations or compared older chemotherapy regimens no longer considered standard.

Overall survival

A multiple-treatment modelling meta-analysis reported hazard ratios for death for first-line treatment, for each type of regimen compared with monotherapy with a non-platinum, non-taxane agent, not administered intraperitoneally.  Modelling estimated a 92% probability that combinations of platinum and taxane with intraperitoneal administration were the most effective regimens.72

Two phase III trials reported differences in overall survival. The trial by Reed et al (2006), which included patients unfit to receive cisplatin, reported improved survival in the carboplatin arm (median 15 months) compared with treosulfan (median 12 months) (p<0.026).25  Long-term follow-up of the OV10 trial of older chemotherapy regimens, reported that paclitaxel and cisplatin combination improved survival compared with cyclophosphamide and cisplatin HR 0.75 (95% CI 0.63 to 0.90), p=0.001.35

Progression-free survival

Twenty-three of the primary randomised controlled trials reported on progression-free survival. One trial reported on disease-free survival rather than progression-free survival.24  Most reported no statistically significant progression-free/disease-free survival differences between treatment groups however, the phase II trials were not designed/powered to detect survival differences.

Three phase III trials reported differences in progression-free survival: Reed et al (2006), OV1035 and OVAR9.17   In two trials, those in the standard chemotherapy arm had longer progression-free survival than those in the intervention arm. The trial by Reed et al (2006) reported longer time to progression in the carboplatin arm (10 months) compared with treosulfan (5 months) (p<0.001).25  The OVAR9 trial reported median progression-free survival in the standard paclitaxel/carboplatin arm as 19.3 months compared with 17.8 months for the paclitaxel/carboplatin/gemcitabine arm (p<0.01).17

In the other trial, improved progression-free survival, was reported in the intervention arm. OV10 reported better PFS in the paclitaxel/cisplatin arm compared with the older chemotherapy regimen cyclophosphamide/cisplatin (p<0.001).35

Adverse events

The adverse events reported varied between each trial. Overall, the addition of agents to standard chemotherapy tended to increase toxicity, particularly haematological toxicity such as anaemia and neutropenia. Adverse effect profiles reflected the various agents used. However, there were often limited differences in toxicities between treatment arms.

Quality of life

Quality of life (QoL) was assessed in 13 trials, with detailed data reported in 11 trials. Most trials reported no significant differences in quality of life between treatment arms investigated.28, 29, 31 OVAR 3 found that those in the carboplatin/paclitaxel arm showed better overall QoL, physical functioning, role functioning, and cognitive functioning compared with those in the cisplatin/paclitaxel arm after treatment.30 However, an additional paper which reported a retrospective analysis of data from OVAR 3, OVAR 5 and OVAR 7 found that correlations between toxicity grading and quality of life functioning scales were weak and symptom level agreement between clinician and patient reporting could differ.30