Analysts see a new hope in three-year ExteNET data
If stock prices are anything to go by, investors have found some positive aspects on closer examination of the three-year data from the ExteNET phase III trial of Puma's neratinib, presented at the San Antonio Breast Cancer Symposium last week.
Those positive findings would seem to centre on the pan-Her inhibtor's greater activity in hormone receptor-positive patients, which could tip a delicate balance in the project's favour given its otherwise modest efficacy in the extended adjuvant breast cancer setting and well-known diarrhoea side-effect.
This new-found optimism was reflected in yesterday’s 11% share price rise, which more than reversed a decline seen on Friday caused by the fact that neratinib's benefit on invasive disease-free survival (iDFS) had narrowed relative to that seen at two years.
Neratinib strikes back
It is evident from the full San Antonio data that the difference in iDFS grows when the analysis is restricted to those women with centrally confirmed Her2-positive status, hormone receptor-positive status (~60% of the ITT group) or both, as well as those who had completed adjuvant treatment with Herceptin within a year of starting on the trial.
This is consistent with the fact that prior experience with Herceptin suggests that women are at a higher risk of disease progression the closer they are to the completion of their treatment with adjuvant therapy.
|Three-year ExteNET data|
|Group analysed||Neratinib||Placebo||Hazard ratio, p value|
|ITT group (n= 2,840)||92.0%||89.9%||0.74, p=0.023|
|Treated <1yr from adjuvant completion (n=2,297)||91.5%||88.9%||0.72, p=0.020|
|Her2+ centrally confirmed (n=1,709)||91.8%||89.6%||0.70, p=0.037|
|Her2+ centrally confirmed, treated <1yr from adjuvant completion (n=1,392)||91.7%||88.2%||0.63, p=0.009|
|HR+ (n=1,631)||93.6%||89.3%||0.57, p=0.003|
|HR+, treated <1yr from adjuvant completion (n=1,334)||93.3%||88.6%||0.57, p=0.004|
|HR+, Her2+ centrally confirmed (n=903)||94.4%||88.0%||0.43, p<0.001|
Investors initially seized on the 2.1% absolute difference reported in the rate of iDFS between neratinib and placebo at three years, finding it concerning as this was lower than the 2.3% difference reported at two years. This suggests a lower, 26% reduction in risk of invasive disease or death over the longer time period.
This would still ostensibly be statistically significant, having a p value of 0.023, but for the fact that all of the “alpha” – the statistical penalty for multiple analyses – has been spent. Thus it would be wrong to claim this as being significant, and Puma is careful not to do so. Under the ExteNET study’s statistical analysis plan, prospectively defined analyses can only occur at years two and five.
Interestingly, this two-year iDFS figure has also now been revised up to 2.5%, with the inclusion of new data on some patients who were previously censored at the original two-year assessment. This change hints at one of the major issues with the ExteNET data: its extensive censoring, caused by the original sponsor Wyeth's decision to discontinue follow-up when it was thought unlikely to render a positive result.
The loss of some 1,400 patients to follow-up means that Puma’s claims for neratinib rest on small, though still statistically significant, differences in relatively low numbers of patients even though these were derived from a large trial.
Return of the data
In the presentation at SABCS the Kaplan-Meier curves that extended to year four suggested that the iDFS benefit fell to 1.9 points (90.5% vs. 88.6%).
In terms of subset analyses, the 4.3-point absolute difference in iDFS in the hormone receptor positive subgroup at three years was effectively matched, with 4.4 points at four years. Puma intends to file neratinib early in the EU and US in early 2016, based on the two-year data.
The group has indicated that the two key regulators diverge somewhat on their likely interest in subsequent data cuts, with the EU dismissing these for being potentially subject to statistical bias. Both the EU and US regulators are understood to be interested in the findings in hormone receptor-positive and the recent adjuvant completer subgroups, however.
Puma does not know if the FDA will place any value on the updated year-two data, but notes that the hazard ratio and p values are roughly the same.
Many analysts expect the regulators to narrow the indication, and have removed hormone receptor-negative patients from their models. Such a decision would be unusual, if not quite precedent-setting. More likely, however, is the same practical effect occurring as a result of physicians reserving neratinib's use for those women who are believed to be at elevated risk of progression.
Investors are also concerned about Roche’s ongoing Aphinity trial, which combines chemotherapy with Herceptin and Perjeta, and could improve iDFS in the adjuvant setting and leave little room for neratinib to show a subsequent benefit.
Neverless, neratinib still remains one of the most valuable unpartnered phase III projects in the biotech space, and pharma companies will no doubt continue to watch its progress with interest.