Cancer Drugs Fund ( CDF ) Panel has assessed the application for Eribulin ( Halaven ) to remain in the CDF list as 3rd line chemotherapy for patients with locally advanced or metastatic breast cancer who have already had at least two previous lines of chemotherapy for advanced disease.
The CDF Panel was aware that this indication for Eribulin has been appraised but not recommended by NICE.
There was one randomised trial ( EMBRACE trial ) submitted by the manufacturer as the main evidence base related to this indication. The trial was published in The Lancet in 2011.
This randomised 762 patients with locally advanced or metastatic breast cancer previously treated with at least two lines of chemotherapy for advanced/metastatic disease to receive Eribulin vs treatment of physician’s choice.
The comparator group of treatment of physician’s choice ( TPC ) represented a mix of agents to mirror clinical practice at the time in this setting, the authors of the publication confirming that treatments in this comparator arm were real-life treatment choices.
The primary endpoint was overall survival.
The trial was stratified for geographical region, previous chemotherapy with Capecitabine and HER-2 status. 68% of patients had previously received treatment with Capecitabine. 96% of the TPC arm received chemotherapy, 25% were treated with Vinorelbine, 19% with Gemcitabine, 18% with Capecitabine, 15% with taxanes, 10% with anthracyclines and 10% with other chemotherapies.
Median progression free survival ( PFS ) was independently assessed and was not significantly greater with Eribulin ( 3.7 vs 2.2 months, difference [ delta ] 1.5 months, hazard ratio [ HR ] 0.87, 95% confidence interval [ CI ] 0.71-1.05, p=0.137 ).
A sensitivity analysis was done on progression-free survival as assessed by investigators and the median figures were 3.6 vs 2.2 respectively ( delta 1.4 months, HR=0.76, 95% CI 0.64-0.90, p=0.002 ).
Updated overall survival ( OS ) was significantly superior in the Eribulin arm ( 13.2 vs 10.5 months, delta 2.7 months, HR=0.81, 95% CI 0.67-0.96, p=0.014 ).
There was no evidence of differential plateauing between the Eribulin and TPC survival curves for overall survival.
The CDF Panel noted that the trial was stratified for regions of the world and that subgroup analysis indicated that overall survival was statistically significantly greater in the Eribulin arm in the 64% of the trial population that came from North America, Western Europe and Australia whereas overall survival was similar in the two arms in the 36% of the trial population recruited in Eastern Europe, Latin America and South Africa.
The CDF Panel did not regard these subgroup analyses as being robust and chose to use the trial outcomes on an intention to treat basis.
Safety was assessed according to National Cancer Institute Common Terminology Criteria for Adverse Events. Serious adverse events occurred in 25% of patients on Eribulin and 26% of those on TPC, adverse events leading to therapy discontinuation in 13% and 15%, respectively.
Treatment-related fatal adverse events occurred in 1% in both arms.
Grade 3 and 4 neutropenia was greater in the Eribulin arm ( 45% vs 21% ). When the TPC arm was separated by treatment and comparisons made of Eribulin versus individual TPC agents, the trial publication stated that grade 3 or 4 neutropenia was seen in 45% of 503 Eribulin patients vs 40% of 61 Vinorelbine patients, 29% of 38 taxane patients and 27% of 46 Gemcitabine patients.
Granulocyte colony stimulating factors were given to 18% of Eribulin patients and 8% in the TPC group. Febrile neutropenia occurred in 5% vs 2%, respectively.
All grade toxicities were observed as follows for Rribulin vs TPC: fatigue ( 54% vs 40% ), alopecia ( 45% vs 10% ), peripheral neuropathy ( 35% vs 16% ), arthralgia/myalgia ( 22% vs 12% ), weight loss ( 21% vs 14% ), pyrexia ( 21% vs 13% ) and palmar-plantar erythrodysaesthesia ( 1% vs 14% ).
The manufacturer also presented selected all grade toxicity data of Eribulin versus individual drugs within the TPC arm which showed similar levels of toxicities for some side-effects ( eg fatigue: 54% for Eribulin and 51% for Vinorelbine; hair loss: 45% for alopecia and 34% for taxanes ).
In assessing all of these toxicities, the CDF Panel concluded that Eribulin resulted in greater toxicity than TPC as a whole. It regarded the comparator of TPC to be clinically relevant and also considered that comparison of toxicity is the most robust when done on an intention to treat basis.
It was also concerned that the comparisons of the full toxicity data set for all the Eribulin patients with much smaller numbers of the individual treatments that made up TPC introduced uncertainty and bias as it was not necessarily comparing like with like, an example being the compared populations having different degrees of pre-treatment.
In addition, the CDF Panel noted that the authors of the Eribulin publication in the Lancet had stated that the TPC arm included several different chemotherapies, each with their own characteristic toxic effects, precluding detailed comparison with Eribulin.
The manufacturer in its CDF application stated that limited inference can be drawn from direct comparison between patients treated with Eribulin and those treated with TPC in study 305 ( EMBRACE ), as the TPC group comprises patients treated with a wide range of therapies and dosing regimens.
The CDF Panel considered the issue of selective interpretation of toxicity comparisons in the context of the various treatments that were chosen as part of TPC. It noted that the manufacturer had stated in its application that it is apparent that for patients with LABC/MBC [ locally advanced breast cancer / metastatic breast cancer ], numerous types of treatments may be used.
The CDF Panel agreed with this position and noted too that the authors of the publication considered that the benefits of Eribulin are, therefore, arguably more likely to be generalizable to clinical practice than if the control treatment had been artificially constrained.
The CDF Panel thus concluded that all the outcomes of the trial should be considered on an intention to treat basis as this was the design of the trial and the plurality of the comparator with Eribulin in this trial ( TPC ) reflected clinical practice.
In addition, it did not consider as statistically robust the comparison of the toxicities of all 503 assessed patients who received Eribulin with much smaller subgroup populations of individual treatments within the TPC arm.
The CDF panel noted too that the overall survival results reflected the comparison of Eribulin versus TPC as a whole and thus the toxicity comparison should do likewise.
Quality of life measures were not collected in this study, the authors of the Eribulin publication stating that the differing TPC agents would have complicated the interpretation of quality of life.
The median duration of treatment in the Eribulin arm was 3.9 months.
The manufacturer also submitted data from another randomised trial ( study 301 ) which randomised 1102 patients with locally advanced or metastatic breast cancer previously treated with up to 2 lines of chemotherapy to receive Eribulin vs Capecitabine. Median progression free survival and overall survival were not significantly different in the intention to treat analysis, the latter determined by predefined criteria for statistical significance.
Toxicity was increased in the Eribulin arm for some side-effects but reduced for others. Quality of life ( QOL ) was collected in this study and the main result has been published in a peer-reviewed paper. There were no significant differences in global QOL between the two arms.
The manufacturer presented the QOL results from the subgroup constituted by 27% of the total 301 study which was receiving 3rd line treatment for metastatic breast cancer and this comparison showed little difference in global QOL.
The main reason for the submission of these results from the Eribulin vs Capecitabine study was because the manufacturer chose to use the QOL results from this study and apply it to the scoring for Eribulin vs TPC.
The CDF Panel did not consider that the QOL results from study 301 could be applied to the assessment of clinical impact in study 305 ( EMBRACE ). This was because of major differences between these two studies: they had different comparators to Eribulin ( TPC in study 305, Capecitabine in study 301 ), different numbers treated with Capecitabine in the comparison with Eribulin ( 18% in 305, 100% in 301 ), different proportions previously treated with Capecitabine ( 68% in 305, 0% in 301 ) and they had different patients in different parts of the patients pathway ( 100% treated with 2 chemotherapy regimens for advanced disease in 305 whereas the figure was 27% in 301 ). For all these reasons, the CDF Panel concluded that there was no reliable QOL evidence to inform its assessment of the comparison of Eribulin versus TPC in the context of patients previously treated with at least two chemotherapy regimens for advanced/metastatic disease.
In terms of its assessment of the CDF scoring of unmet need, the CDF Panel was aware that the EMBRACE study was the first study to demonstrate a survival benefit in heavily pre-treated women with metastatic breast cancer.
It did not consider a survival benefit of 2.7 months to be a step change in the management of breast cancer when the drug was associated with significant and greater toxicity and did not have any QOL evidence to support its use in this place in the treatment pathway.
The CDF Panel noted that Eribulin did not fulfil the CDF definition of unmet need as there were clearly a number of other treatment options available as 3rd line chemotherapy for advanced/metastatic breast cancer.
The manufacturer also presented some real life data of treatment with Eribulin in breast cancer from 4 centres in England. These showed median overall survival durations of between 10 and 13 months and comments that the main observed toxicities with Eribulin were fatigue, alopecia, neutropenia, infection, neuropathy and nausea.
The Panel’s scores for Eribulin as 3rd line treatment for locally advanced or metastatic breast cancer were as follows: 0 for progression-free survival ( no significant difference in the secondary endpoint of independently assessed PFS; a benefit of 1.4 months in investigator-assessed PFS ); 2 for overall survival ( a difference of 2.7 months ); 0 for quality of life ( no direct or appropriate comparison ); minus 1 for toxicity; and 0 for unmet need.
The total clinical score was 1B.
The CDF Panel was aware that Eribulin was one of the commoner requested drugs in the Cancer Drugs Fund. It knew that a modified dosage vial may result in more efficient use of Eribulin by the NHS ( National Health Service )
In conclusion, the CDF Panel considered the clinical benefits of Eribulin as 3rd line treatment for locally advanced or metastatic breast cancer to be insufficient to justify retention within the Cancer Drugs Fund. ( Xagena )
Source: Cancer Drugs Fund ( UK ), 2015