QurAlis teaming with Unlearn to use AI to improve ALS clinical trials
Company is currently testing therapeutic candidates QRL-201 and QRL-101
The aim of the collaboration is to minimize variability and increase the statistical strength of QurAlis’ clinical trials that are evaluating their primary ALS therapeutic candidates, QRL-201 and QRL-101.
“We are excited to partner with Unlearn to help advance our clinical program with AI and other innovative technologies to generate evidence suitable for supporting regulatory decisions and help speed new, lifesaving precision medicines to patients with ALS and other neurodegenerative diseases,” Kasper Roet, PhD, QurAlis’ founder and CEO, said in a company press release.
Unlearn uses an AI technique called machine learning to create “digital twins” of patients enrolled in clinical trials. Its patented machine learning models are trained on large volumes of completed clinical data and can predict, at the individual level, the outcomes of these digital twins if they were given the control treatment.
Data collected from each patient is fed to the AI model to generate a digital twin, which will be utilized in randomized controlled trials (RCTs) called TwinRCTs.
TwinRCTs are designed to provide more statistical power with smaller control groups than traditional trial designs. Many patients are wary of enrolling in clinical trials because they might receive a placebo, but a smaller control group boosts the chance of getting the investigational treatment.
Benefits of using AI in ALS clinical trials
The technology will also speed up a trial’s enrollment process, accelerating a therapy’s time to market.
“By using machine learning to leverage the wealth of existing patient data from completed clinical trials, our technology significantly shortens typical timelines by months while generating evidence suitable for supporting regulatory decisions,” said Charles Fisher, PhD, Unlearn’s founder and CEO.
At the end of the trial, outcome scores derived from the digital twins will be integrated into the primary trial analysis. This will allow for an accurate estimate of treatment effects and minimize false positives (Type-1 errors), a result that indicates a treatment’s effectiveness.
“Advances in machine learning and AI make it possible to enhance trial power to detect a positive result when one truly exists while controlling for Type-1 error and significantly shorten timelines without introducing bias into the study,” Roet said.
The technology will be used to accelerate the development of QRL-201, a treatment designed to raise stathmin-2 levels and lower the toxic TDP-43 clumps that build up and damage nerve cells in ALS.
It will also boost clinical trials of QRL-101, which is designed to reduce the excessive firing of electrical signals, known as hyperexcitability, seen in about half of ALS cases.
QRL-201 is being investigated in a Phase 1 clinical trial called ANQUR (NCT05633459), which recently dosed its first patient. The study is testing the tolerability and safety of ascending doses against a placebo in up to 64 adults with ALS without SOD1 or FUS gene mutations.
The trial is ongoing in Canada and U.S. sites are expected to open soon. QurAlis anticipates sites in the European Union by the end of the year.
QRL-101 is also undergoing clinical testing in a Phase 1 trial (NCT05667779). The study, which began dosing this year, will assess the therapy’s safety, tolerability, and pharmacological properties in up to 40 healthy adults at a single center in the Netherlands.
“The AI technology we’re developing today will revolutionize the future of clinical research,” Fisher said.