Puma files, and the battle begins


The latest iteration of the now almost incomprehensible Extenet study of Puma’s neratinib seems to have yielded something positive: a recent statistical analysis tweak has flattered the results, no longer suggesting a vanishing benefit over time. Moreover, the company has delivered on its promise to file neratinib in the US.

But there the good news ends. Roche’s Aphinity trial of Perjeta is the elephant in the room, while a huge number of censored patients in a new five-year Extenet dataset raises questions over the reliability of the all-comers results. Investors who bid Puma up 20% this morning should bear these caveats in mind.

Of course, heavy patient censoring – either owing to loss to follow-up or because of missed scheduled doctor visits – has long been a major problem with Extenet. This has been caused by the original sponsor, Wyeth, discontinuing follow-up when a positive result seemed unlikely, and Puma then trying to pick up the pieces.

Stretching credibility

The upshot of Puma’s interim five-year data cut, revealed for the first time yesterday, is this: of an initial 2,840 breast cancer patients 1,651 have been censored, and the benefit in the remainder amounts to 136 neratinib patients progressing versus 172 on placebo.

It surely stretches credibility to rely on a difference of 36 patients at a time when 58% of the 2,840 all comers have been censored. Patient dropouts are especially relevant given neratinib’s notorious diarrhoea side effect; in March the three-year data cut showed an already high 46% of all comers being censored.

Extenet measures the effect of putting post-surgery patients on a year’s neratinib after a year of Herceptin in this “extended adjuvant” setting, against placebo. Prospective analyses were set at two and five years but, since all the “alpha” in Extenet has been spent by multiple data analyses, the two-year post-neratinib benefit is the only statistically robust measure, and this is the basis of yesterday’s US filing.

As before, this stands at a 93.9% invasive disease-free survival (iDFS) versus 91.6% for placebo – a statistically significant 2.3-point benefit. But it is longer follow-ups that had been problematic, previously suggesting that the already tiny numerical benefit got even smaller.

This seems now to have been allayed, thanks to implementation of a previously announced, FDA-mandated statistical change. While before the neratinib benefit shrunk to 1.9 points at four years, it now stands at 2.2 points at this time cut, and rises to 2.5 points at the new, immature, five-year measure.

A vanishing benefit on March 28
Time point Neratinib Placebo Relative benefit (pct points)
2 years 93.9% 91.6% 2.3
3 years* 92.0% 89.9% 2.1
4 years* 90.5% 88.6% 1.9
5 years Not available
*as presented at SABCS 2015.
…turns positive on July 21
Time point Neratinib Placebo Relative benefit (pct points)
2 years 93.9% 91.6% 2.3
3 years 92.5% 90.3% 2.2
4 years 91.4% 89.2% 2.2
5 years** 90.4% 87.9% 2.5
**interim data.

Still, questions around the reliability of Extenet are overshadowed by an even greater concern, namely that the FDA will have access to Roche’s Aphinity data before it concludes deliberations over neratinib (Minor delay puts neratinib on collision course with Aphinity, March 29, 2016).

In Aphinity Perjeta is added on top of Herceptin, not instead of it, for one year in adjuvant breast cancer treatment, and iDFS is measured. The risk of a positive Aphinity readout wiping out the case for neratinib has been known for some time; yesterday Roche confirmed that Aphinity should read out in late 2016, leading to a possible Perjeta line extension launch next year.

Puma has increasingly been playing up neratinib’s benefit in a post hoc subset of hormone receptor-positive patients, and yesterday showed a 3.9-point benefit favouring neratinib at three years, rising to 4.8 points at five years, in this subgroup.

The extent to which the FDA looks at this, or indeed the new analyses of Extenet out to five years, is anyone’s guess. All eyes now turn to the Aphinity readout.

To contact the writer of this story email Jacob Plieth in London at jacobp@epvantage.com or follow @JacobPlieth on Twitter

Share This Article