Biotech Values

[jq1234]

She had some good points, and some missed points, and some debatable points. I don't have time to write a full article like she did. Here are some points:

Her best points IMO are:

1) Sarepta conducted their phase IIb study at one single site.
2) The total number of patients in the trial are small.

Her missed points:

Muscle biopsies were repeated for all patients at 48 weeks. As reported by Sarepta, there was a statistically significant increase in dystrophin-positive fibers to 47% of normal in the six patients treated with eteplirsen since the start of the study. The placebo/delayed treatment patients also showed a significant increase of dystrophin-positive fibers to 38.3% of normal after just 24 weeks of therapy. This is surprising as it's nearly double the dystrophin increase observed after 24 weeks of treatment for the active treatment patients (which was only 22.5%.)

1) she is comparing combined 30mg/kg and 50mg/kg placebo/delayed treatment result 38.3% to 22.5% from 30mg/kg for 24 weeks double blind period. There is clear dose response here between 30mg/kg and 50mg/kg for placebo/delayed treatment arm, 34.2% and 42.9% respectively. So apple to apple comparison is 34.2% vs 22.5%, not big difference as she claimed for small N.

However, when we look at the relationship between percentage of dystrophin-positive fibers and mean change in the six-minute walk test (6MWT), there is no clear correlation between eteplirsen dose, percentage of dystrophin-positive fibers and 6MWT performance.

2) She misses the point here because she is asking one to one correlation between surrogate endpoint and final endpoint. The perfect example of this relationship between surrogate endpoint and final endpoint is ORR vs OS in oncology. Follow her logic, someone with PR definitely needs to show longer OS than someone with SD to make ORR as valid surrogate endpoint for OS. This is further from truth in reality. However unperfect correlationship between ORR and OS, it doesn't prevent FDA from giving accelerated approval by using ORR as surrogate endpoint for unmet medical needs in well defined patient population.

Her debatable points:

The "pivotal" eteplirsen phase II data are systematically biased.

It depends on how you look at the study. As she focused on the unblinded portion of data of 6MWT similar to many others who had done, I agree she has a point. However, if you look at what I think SRPT should present the data to FDA which is to focus on surrogate endpoint of dystrophin of overall study both blinded and unblinded extention, then I think they have a chance to convince FDA for accelerated approval filing - I am not a person who are so sure of things of one way or the other like so many others do.

First, the study is 24-week trial with primary endpoint change from baseline in the percent of dystrophin positive fibers in muscle biopsy tissue as measured by immunohistochemistry (IHC). The study met primary endpoint for 30mg/kg at 24 weeks, but missed for 50mg/kg at 12 weeks. Then the next 24-week period can be considered as extension trial where patients on 30mg/kg and 50mg/kg continued their treatment, while placebo patients randomized to 30mg/kg and 50mg/kg. The 24-week extension confirmed the following on surrogate endpoint:

Both 30mg/kg and 50mg/kg not only maintained their dystrophin at 48 weeks from 24 weeks (12 weeks for 50mg/kg), but also continued to improve. The placebo patients who didn't increase dystrophin at 24 weeks improved theirs after 24 weeks on eteplirsen. Thus, 24-week extension study further validated the surrogate endpoint from blinded 24-week period.

As of the 6MWT data, they should be presented as supplemental data to argue there is a reasonable chance the surrogate endpoint could predict final 6MWT endpoint in a confirmatory trial. It is no worse than ORR vs OS relationship. There is no need at this point to prove perfect correlationship between dystrophin surrogate endpoint and 6MWT endpoint, that is what confirmatory trial is for! This way, it can bypass the potential bias argument she made about unblinded 6MWT data and no 6MWT data from 2 patients on 30mg/kg.