This week, Roche released detailed results of a closely watched and, ultimately, unsuccessful study testing an experimental treatment for Huntington's disease that it's been developing alongside partner Ionis Pharmaceuticals.

The study, which enrolled hundreds of participants, was previously stopped early, on the recommendation of outside experts who concluded Roche's drug was not likely to be any better than a placebo at improving patients' brain and motor function. Though only high-level results were available at the time, the outcome contrasted with an earlier, much smaller trial that showed the drug, known scientifically as tominersen, was able to lower the mutated protein responsible for the neurodegeneration associated with Huntington's.

The larger study's failure was a surprise to many, and raised several questions about the tominersen program, not least of which was whether Roche and Ionis had miscalculated in their approach.

Tominersen is what's known as an antisense medicine, meaning it's designed to block the instructions cells use to make certain proteins. Antisense therapies have gained more traction in recent years as treatments for rare, genetic disorders  — Ionis, notably, has three on the market  — and the thinking was they could be valuable in Huntington's, too.

Despite the negative results last year, Roche believes there could be a path forward for its drug. On Tuesday, the company announced plans to run another mid-stage clinical trial that will evaluate tominersen specifically in Huntington's patients who are younger and whose disease is considered less severe.

That decision was informed by follow-on analyses of the late-stage clinical trial. In a Thursday webinar hosted by the patient advocacy group Huntington's Disease Society of America, Roche researchers detailed how certain patients who received tominersen less frequently seemed to do slightly better than those given placebo about 17 months later.

"These findings suggest that lower exposure to tominersen may benefit younger adult patients with lower disease burden," said Peter McClogan, clinical director of the Huntington's disease program at Roche, in the webinar.

"These findings are not definitive," he added, "and that's why we need to further investigate and evaluate them in a prospectively designed, randomized and controlled study."

Roche's cautious optimism isn't shared, however, by analysts who follow the company and partner Ionis. Several were skeptical of Roche's after-the-fact findings and its plans to move forward with another trial. One, Mani Foroohar of the investment bank SVB Leerink, described the new data as "weak evidence" and noted that Roche has a "challenging path ahead" to prove its hypothesis.

That hypothesis rests on statistically shaky ground. After Roche stopped its Phase 3 study of tominersen last March, the company continued to collect data on the nearly 800 Huntington's patients who were enrolled in the trial. The conclusions Roche reached and disclosed Thursday were based on a review of that data, a practice known as "post hoc" analysis that's generally considered to be exploratory and speculative.

In their presentation, McClogan and his colleague Lauren Boak, who leads neuroscience product development at Roche, acknowledged the uncertainty, emphasizing that the seemingly positive signs in younger patients with less advanced disease may be a "chance finding."

Because of this, McClogan and Boak said Roche won't make tominersen available through compassionate use, a framework that's used to allow patients access to drugs outside of clinical trials.

"We realize this is hard to hear," Boak said, "But safety is our number one priority and, to be able to offer compassionate use, we really need to have confidence in a positive benefit-risk [balance]."

In addition to being post-hoc, Roche's conclusions are based on dividing the trial participants into smaller subgroups based on age and a score used to assess Huntington's disease severity. Only 50 patients were in the group singled out as having improved outcomes over placebo.

Further, in the overall study, patients who received tominersen more frequently — every 8 weeks rather than every 16 — actually did worse than those who were given placebo, raising questions about whether the drug could be harmful.

"One hypothesis," said McClogan, "is that the amount of drug that patients have had, having treatment every eight weeks, was maybe too high and not maybe related to the drug itself."

The drug was also associated with changes in the size of ventricles in the brain among those who received the medicine. Though the majority of those instances didn't lead to any health problems, increasing ventricular volume over time "could come with serious safety events" or reduce the drug's effectiveness, wrote SVB Leerink's Foroohar in a Jan. 21 note to clients. 

In general, reports of side effects were comparable across the subgroups used by Roche and the placebo arm, although there were slightly fewer severe adverse events among younger, healthier patients taking tominersen.

Tominersen wasn't the only drug of its type to fall short in testing last year. Along with Roche and Ionis, a small, Massachusetts-based biotechnology company called Wave Life Sciences had attracted attention for its own antisense drug targeting Huntington's.

But disappointing data last year led Wave to scrap two of its most advanced candidates, a turn of events that, coupled with Roche's setback, led to questions over the companies' treatment approach. Particularly alarming was that Roche's treatment had lowered levels of the mutant protein — called huntingtin — behind Huntington's, but didn't lead to patient benefit.

"It doesn't mean that [antisense medicines] or huntingtin lowering doesn't work for Huntington's disease," said Vicki Wheelock, director of the Huntington's Disease Society of America Center of Excellence at University of California, Davis, on the Thursday webinar. "It means that in this study, in this population of patients, with this dose, at this interval, it didn't work."

"I think we have to take the time to sit back," she added, "and take the time for [Roche's new study] to happen and see what we can learn."