The majority of ALS patients are excluded from clinical trials. Experts say using biomarkers and stratifying trial populations can expand eligibility and provide additional scientific insights.
Pictured: A doctor speaks with a patient/iStock, ArLawKa AungTun
Approximately 60% of amyotrophic lateral sclerosis patients are excluded from clinical trials by the time they are diagnosed, according to a 2019 literature search. NeurologyLive reported that investigators believe this number is an understatement. Inherent in this statistic are both ethical and drug development implications.
Angela Genge, chief medical officer at QurAlis and top recruiter for the pivotal trial of Biogen’s Qalsody (tofersen)—which was recently approved to treat superoxide dismutase 1 (SOD1)-ALS—said the field is moving toward expanding eligibility for ALS trials in early-phase studies.
“It’s actually quite a practical question,” she told BioSpace, “and ultimately, we will need to expand our eligibility.”
First, such a change would allow more ALS patients to receive and potentially benefit from experimental treatments, Stacy Lindborg, Co-CEO at BrainStorm Cell Therapeutics, told BioSpace. “You think about the horrible nature of being diagnosed with a fatal illness and then looking to see the trials that may give you access to an investigational product, only to find out that you’ve just missed the standard criteria.”
In addition, expanding inclusion criteria could help drug developers get a truer sense of how well those treatments work, Lindborg noted.
The current inclusion criteria for trials selects for faster-progressing patients, who tend to receive a diagnosis more quickly, making them eligible for enrollment.
“If you have a molecule that takes time to work so that the fast progressors in the program never really get the full benefit, you could skew your results in the wrong direction,” Genge said.
Including fast progressors may have affected both Biogen’s Phase III VALOR trial and the Phase III trial of BrainStorm Cell Therapeutics’ NurOwn.
In a previous interview with BioSpace, Genge called Qalsody “a good drug with a poor statistical plan.”
The primary endpoint for the VALOR trial was changed from baseline to week 28 on the Revised Amyotrophic Lateral Sclerosis Functional Rating Scale (ALSFRS-R) among faster-progressing patients. In this subgroup, the change from baseline was -6.98 with Qalsody compared to -8.14 for the placebo group—a difference of 1.2 points that was not deemed statistically significant.
Genge said the definition of a fast progressor changed between the Phase I/II trials and VALOR, leading to the placebo arm performing better than expected.
Sabrina Paganoni, co-director of Massachusetts General Hospital’s Neurological Clinical Research Institute, said another limiting factor is a sponsor’s ability to afford a larger trial.
“The broader the criteria, the bigger the study will have to be to account for patient-to-patient variability,” Paganoni told BioSpace. “Some studies have actually restricted the inclusion/exclusion criteria to be able to have a relatively small study and still be powered to see a treatment effect.” However, she noted that with this approach, there may be questions about the generalizability of the results.
Limiting Factors
In addition to inclusion criteria, sponsors need to think about trial endpoints during patient selection, Lindborg said. Widely accepted as the gold standard for measuring disease progression, the ALSFRS-R “does not work perfectly and equally well across the spectrum,” she added.
She said this could lead to a floor effect, which occurs when a scale such as the ALSFRS-R cannot measure progression once patients’ health deteriorates beyond a certain point. BrainStorm ran up against this floor effect with the Phase III trial of NurOwn—a therapy composed of autologous mesenchymal stem cells (MSCs) and neurotrophic factors (NTFs).
NurOwn hit the Phase III primary endpoint, with 34.7% of participants seeing an improvement of 1.25 points monthly. Still, the placebo response of 27.7% of patients achieving that same improvement exceeded placebo responses observed in contemporary ALS trials. However, 34.6% of NurOwn recipients received a clinically significant response in a pre-specified subgroup of patients with early-stage disease versus only 15.6% in the placebo group.
Posthoc analyses accounting for the floor effect showed a statistical trend toward a clinically meaningful treatment effect, BrainStorm reported in November 2022. “That floor effect can result in misclassification of a clinical response,” Lindborg said.
Even so, there is value in broadening trial populations, she added.
“To generate evidence of a treatment, we need to use endpoints that measure disease progression, but we should also try to match real-world experience and broaden access to clinical trials,” Lindborg said. “As a community, we could benefit by including a broader group of people living with ALS beyond those meeting stringent inclusion criteria who can offer benefits like furthering ALS biomarker data, which could drive accelerated diagnoses and inform drug development.”
Getting to an Answer
Genge said ALS investigators are working toward a robust stratification model that is predictive of speed of progression. She pointed to the Personalized European Network for the Cure of ALS (ENCALS) survival prediction model as a possible option that has been widely tested.
Genge said that the strongest predictor of speed of progression is neurofilament in SOD1-ALS patients. She said it is “very likely” that a combination of clinical and biomarker results will help predict speed of progression.
With an accurate measure of disease progression, it would be possible to include everybody “because we can then . . . analyze the groups as a whole, but you’ll also plan to analyze the subgroups based on their speed of progression,” she explained. Then, if there is a relationship between speed of progression and efficacy of the molecule, “you can tease it out.”
Lindborg echoed this sentiment, saying one strategy for increasing patient access is expanding the inclusion criteria while defining a priority analysis population. She said this approach provided a “silver lining” in BrainStorm’s Phase III trial. “By definition of having some participants that wouldn’t have otherwise been included in many other trials, we actually now know that NurOwn is biologically active in all participants, including the participants the ALSFRS-R can’t effectively measure.”
Paganoni said that inclusion and exclusion criteria should be tailored to the goal of the specific trial and added that the mechanisms of action vary between drugs. “Some may have a faster onset of biological activity [and] others may take longer to have an impact,” she said.
While stratification based on biomarkers is more effective for drugs targeting the approximately 10% of ALS patients with a genetic mutation such as SOD1, she said, “As we learn more about biomarkers and develop more biomarkers, this could be a strategy for more programs depending on the mechanism of action.”
Paganoni said neurofilament levels could help stratify patients beyond genetic forms of ALS because “these are actually a measure of how rapid the disease is in a certain person.”
Genge said that, ideally, you would find a drug that works for everybody. “Failing that, you want to know if there is a category of patients that a particular drug will work for.” She said it might be possible to get to this answer without limiting trial inclusion. “I’d like to be able to bring everybody in and just use our outcome measures or biomarkers to stratify for efficacy.”
The experts who spoke with BioSpace agreed that there is no easy way to determine inclusion/exclusion criteria for ALS trials.
“I wish I could tell you the field now knows that you have to enroll patient X and not Y,” Paganoni said, “but I think it’s a much more nuanced conversation that depends on the specific project.”
Heather McKenzie is a senior editor at BioSpace, focusing on neuroscience, oncology and gene therapy. You can reach her at heather.mckenzie@biospace.com. Follow her on LinkedIn and Twitter @chicat08.
