The appraisal committee reviewed the evidence from a number of sources (see sources of evidence).
Four randomised controlled trials (RCTs) provide the main evidence base for the consideration of paclitaxel as first-line therapy in ovarian cancer. Full results from the ICON3 trial and updated results from 2 others (GOG111, OV10) have become available since NICE issued its last guidance on the use of paclitaxel in the treatment of ovarian cancer.
The GOG111 trial compared combination treatments of paclitaxel (135 mg/m2)/cisplatin (75 mg/m2) and cisplatin (75 mg/m2)/cyclophosphamide (750 mg/m2) in 410 women. All had severe disease (as defined by the International Federation of Gynaecology staging system, FIGO stage III or IV) and sub-optimal tumour reduction following surgery. No statistically significant difference in overall tumour response (that is, complete and partial response) was found (relative risk = 1.19, 95% confidence interval [CI] = 0.95 to 1.5). However, median progression-free survival was statistically significantly longer for patients receiving the paclitaxel/cisplatin combination (18 months vs 13 months, relative risk = 0.7, 95% CI = 0.5 to 0.8, p<0.001). Overall survival was also statistically significantly longer in these patients (38 months vs 24 months, relative risk = 0.6, 95% CI = 0.5 to 0.8, p<0.001). Estimates from updated longer-term study results suggest that the death rate is 30% less among those treated with the paclitaxel-containing regimen (relative hazard: 0.7, 95% CI = 0.57 to 0.87). No statistically significant difference in performance scores was found between the 2 groups.
The OV10 trial also compared the combinations of paclitaxel (175 mg/m2)/cisplatin (75 mg/m2) and cisplatin (75 mg/m2)/cyclophosphamide (750 mg/m2). The 680 women had optimal or sub-optimal tumour reduction following surgery, and 93% had FIGO stage III or IV disease. A statistically significant difference in overall tumour response (that is, complete and partial response) in favour of the paclitaxel combination was found (relative risk = 1.92, 95% CI = 1.52 to 2.42). Like GOG111, the study also found statistically significantly longer median progression-free survival for the paclitaxel combination (15.3 months vs 11.5 months, hazard ratio = 0.74, 95% CI = 0.63 to 0.88, p=0.0005). Overall survival was also statistically significantly higher in this group (35.6 months vs 25.8 months, hazard ratio = 0.73, 95% CI = 0.60 to 0.89, p=0.0016).
The GOG132 trial included comparison of combination paclitaxel (135 mg/m2)/cisplatin (75 mg/m2) with cisplatin (100 mg/m2) alone. All 424 women had FIGO stage III or IV disease and sub-optimal tumour reduction following surgery. No statistically significant difference in overall tumour response (that is, complete and partial response) was found between the group receiving cisplatin alone and those receiving the paclitaxel/cisplatin combination (relative risk = 0.97, 95% CI = 0.86 to 1.09). However, unlike GOG111 and OV10, no statistically significant differences were found in progression-free survival (14.1 months vs 16.4 months, hazard ratio = 1.06, 95% CI = 0.86 to 1.30), and overall survival (26.6 months vs 30.2 months, hazard ratio = 0.99, 95% CI = 0.80 to 1.23). The difference between the findings of the trial and those reported for the GOG111 and OV10 studies may be explained by the extent of patient cross-over between treatments before the disease progressed. However, it is unlikely that this is sufficient to explain such markedly different findings.
The most recent trial, ICON3, compared a different combination of paclitaxel (175 mg/m2)/carboplatin (5 AUC) with either carboplatin (5 AUC) alone or a combination of cyclophosphamide (750 mg/m2)/doxorubicin (75 mg/m2)/cisplatin (75 mg/m2) (CAP). The trial differs from the others, in that patients had a wider range of residual tumour following surgery (54% had optimally reduced tumours), and a smaller proportion (80%) had FIGO stage III and IV disease. Of the total 2,074 women recruited, 1,421 were randomised to receive the paclitaxel/carboplatin combination or carboplatin alone. The findings of the ICON3 trial after more than 3 years' follow-up also differ from those of the GOG111 and OV10 studies. No statistically significant difference was found between the groups receiving the paclitaxel/platinum combination or carboplatin alone, in terms of progression-free survival (17.1 months vs 16.1 months, hazard ratio = 0.94, 95% CI = 0.84 to 1.05, p=0.24) or overall survival (37.6 months vs 36.1 months, hazard ratio = 0.96, 95% CI = 0.84 to 1.09, p=0.53). Also, no statistically significant differences were found in anxiety and depression scores. It is possible that the recruitment of more patients with less severe disease could have diluted the effect of paclitaxel treatment, but sub-group analyses by FIGO stage and extent of residual tumour did not show any trend supporting this. The trial design allowed choice of the control arm before randomisation, and although some suggest that this could also have diluted any treatment effect, it may be that this may better reflect clinical practice in some respects.
The 4 trials showed consistently that treatment with paclitaxel in combination with platinum leads to more side effects. Over the 4 trials statistically significantly higher rates of neutropenia, allergic reactions, cardiovascular problems, hypersensitivity, neuromotor and neurosensory problems, fever and alopecia were reported in patients receiving the paclitaxel/carboplatin combination compared with the control treatments.
While design differences between the 4 trials, in terms of severity of disease of included patients, differences in treatment and control drugs and doses, length of follow-up, and the extent of cross-over (before and after disease progression), may hamper statistical pooling of results, meta-analyses have been undertaken by the Medical Research Council (MRC) and Bristol-Myers Squibb (BMS). These take account of statistical heterogeneity as far as possible, and their results appear consistent, reporting that the findings for progression-free survival (hazard ratios = 0.84, 95% CI = 0.70 to 1.02 [MRC] and 0.87, 95% CI 0.72 to 1.05 [BMS]) and overall survival (hazard ratios = 0.82, 95% CI 0.66 to 1.01 [MRC] and 0.82, 95% CI 0.68 to 1.00 [BMS]) across the trials do not show statistically significant differences between paclitaxel/platinum and the alternatives.
Four published RCTs on the second-line use of paclitaxel (monotherapy) in the treatment of ovarian cancer were identified. However, 2 of these studies compared paclitaxel with unlicensed treatments, and 1 compared different dosing schedules of paclitaxel itself.
In the remaining RCT, paclitaxel was compared with topotecan in 235 women who had been previously treated with a platinum-based compound (they had not been previously treated with paclitaxel). The trial found no statistically significant differences in overall tumour response, progression-free survival or overall survival. The incidence of neutropenia, anaemia, thrombocytopenia, leukopenia, nausea and vomiting was significantly lower among patients receiving paclitaxel than among those receiving topotecan. However, there was a significantly higher incidence of alopecia among the paclitaxel-treated group.
Eleven cost-effectiveness analyses and 3 cost–utility analyses were available as evidence on the first-line use of paclitaxel. All were based on trials favouring paclitaxel (that is, GOG111 or OV10), and therefore found the paclitaxel/platinum combination to be more costly and more effective than control treatments. Three of the analyses could be directly applied to the UK.
Two published UK cost-effectiveness analyses found that the incremental cost per life-year gained for paclitaxel/platinum ranged between £7,173 and £12,417, depending on the effectiveness trial results and drug doses applied. One of the studies reported the incremental cost per progression-free life-year gained to be between £20,084 and £22,021, again depending on the trial results applied.
One published UK cost–utility analysis was available, but its methods were not well reported, and its results need to be interpreted with caution. An incremental cost–utility estimate based on this analysis, for paclitaxel/platinum compared with carboplatin alone, showed the incremental cost per quality-adjusted life year to be £5,273.
A cost-effectiveness analysis undertaken by the manufacturer of paclitaxel was also available. The analysis was based on resource use and outcomes from GOG111, though carboplatin was substituted as the control treatment, as this better reflects UK practice. Consequently, the analysis assumed equivalent efficacy between carboplatin and cisplatin in combination with paclitaxel. UK unit costs were incorporated from routine sources, and included: chemotherapy drugs, pre-medication, drug administration, management of febrile neutropenia, and other inpatient and outpatient care. For the paclitaxel/carboplatin combination versus carboplatin alone, the analysis reported an incremental cost of £7,074 per life-year gained and £10,808 per progression-free life-year gained.
Given that this analysis was based on the survival in the most favourable survival findings available (that is, a hazard ratio of 0.61 in favour of paclitaxel/platinum combination for overall survival), sensitivity analyses were undertaken by NICE to indicate the likely magnitude of effect on the cost-effectiveness ratio of changing the survival gains attributed to paclitaxel/platinum. Simply adjusting the manufacturer's analysis to the survival difference reported by ICON3 (hazard ratio of 0.96) suggests an incremental cost per life-year gained in the region of £45,000. However, other analyses undertaken by NICE suggest that the cost per life-year gained could be much higher.
Having carefully considered the design and full findings of ICON3 in conjunction with the 3 other published (updated) RCTs, the committee concluded that all of the trials contribute to the understanding of the clinical effectiveness of paclitaxel in the first-line treatment of ovarian cancer.
The committee noted that the availability of the full ICON3 evidence meant that 2 of the 4 published trials favoured paclitaxel in combination with a platinum-based compound, whereas 2 trials failed to show a significant difference in survival between the combination and a platinum-based compound alone. The combination of these findings in meta-analyses suggested that there was no statistically significant survival advantage for one of these therapeutic approaches over the other. In addition, cost-effectiveness estimates varied considerably with the assumed magnitude of the survival difference.
The committee took account of this range of trial evidence as well as other factors that would differentiate between the 2 regimens including the side-effect profiles of the treatments, and the broad range of cost-effectiveness estimates presented. On this basis the committee considered that paclitaxel/platinum combination treatment should no longer be recommended exclusively as standard therapy for women receiving first-line chemotherapy for ovarian cancer. As a consequence, the committee considered that both platinum therapy alone and a combination of paclitaxel and a platinum compound were appropriate first-line treatments for women with ovarian cancer.
The committee discussed pathways of care for women with ovarian cancer. It was recognised that women with a good initial response to first-line therapy will be offered additional courses of the chosen treatment, and will be offered second-line treatment options once the tumour fails to respond to the chosen first-line regimen.
In view of the limited evidence available on the clinical effectiveness of paclitaxel in second-line treatment, the committee concluded that paclitaxel should be considered as an option for second-line treatment only for women who do not receive it as part of their first-line therapy. For such women, it should be offered as one option alongside other drugs that are licensed for second-line treatment of ovarian cancer.