Skip to Main Content

A pivotal time for Alzheimer’s disease clinical development

The Alzheimer’s disease (AD) drug development landscape is exciting as of late. This month, two leading pharma companies expect to share data from their highly anticipated Phase 3 studies at the Clinical Trials on Alzheimer’s Disease conference. Eisai and Biogen’s lecanemab and Roche’s gantenerumab are monoclonal antibodies that target the reduction of amyloid beta plaques in the brain, which are widely considered to be a principal driver of AD. Eli Lilly’s donanemab is also making strides with Phase 3 TRAILBLAZER-ALZ 2 confirmatory study data expected in mid-2023.

In September, both Biogen and Eli Lilly released promising preliminary results for their hopefuls.1,2 Then, in early November, Roche announced that gantenerumab failed to significantly slow clinical decline in their Phase 3 GRADUATE program.3 Despite these recent disappointing results, Roche’s Phase 3 SKYLINE trial, initiated earlier this year, is testing gantenerumab’s efficacy in patients with the earliest biological signs of AD and may yield more positive outcomes in the future.4 Indeed, the ups and downs of AD clinical development necessitate continued innovation, as the Alzheimer’s Association has urged a redoubling of efforts to discover new targets and test new drugs for AD.5


Innovation is needed to overcome patient burden while generating evidence sooner

While the three drugs above are currently dominating headlines for being the furthest along and best-funded, an analysis by the Alzheimer’s Association found that there are currently 143 AD drugs in the development pipeline spread across 172 clinical trials, 141 of which are Phase 2 and 3 trials.6,7 Together, if fully enrolled, these trials would require 50,575 patients and a combined ~3.8 million patient weeks of participation, or over 150,000 years.

Of the 50,575 patients, 49,122 will be in Phase 2 and 3 trials, nearly 25,000 of whom will be assigned to the control arm, assuming the trials are placebo-controlled with a 1:1 randomization rate. Meeting the enrollment demands of these 141 trials will be difficult, especially given patients’ fears of receiving placebo. It’s well known that recruitment failure is one of the main reasons clinical trials are terminated or delayed.8 In fact, 80% of clinical trials fail to meet their enrollment timelines, and up to 50% of clinical sites enroll one or no patients.9 In a study of patients who declined clinical participation, 36% cited issues with the study protocol, including a disinterest in placebo-controlled studies.10

In most AD clinical trials, patient participation requires many years of commitment, regardless of whether that patient is assigned to the active or control arm, so it’s understandable why patients fear the possibility of receiving the placebo. For instance, the protocol for gantenerumab involved bi-weekly injections, monthly brain scans to check for bleeds, and yearly appointments to a facility in St. Louis, Missouri. The annual appointments required four days of brain scans, spinal taps, blood analyses, and cognitive tests.11

As leaders of clinical drug development, we owe it to trial patients to continuously innovate randomized control trial (RCT) designs that help to assuage fears of placebo by reducing the number of patients required to be enrolled in a control group — and to run faster trials that get answers sooner. With so many critical AD trials on the horizon, there couldn’t be a more urgent time to adopt innovative solutions that benefit patients and accelerate clinical development programs.

TwinRCTs: Faster trials with smaller control arms

Fortunately, current advancements in artificial intelligence (AI) are already paving a more patient-centric pathway for the future of clinical trials. Unlearn’s TwinRCTTM solution, for instance, leverages historical data and AI-generated prognostic digital twins to run RCTs with fewer control arm patients, allowing more patients to receive the experimental treatment. Since TwinRCTs require fewer patients, they also have significantly shorter timelines.

Unlearn’s recent reanalysis of the ABBY study, presented at the Alzheimer’s Association International Conference 2022, demonstrates how a TwinRCT reduces control arm sizes by up to 35%.12,13 To illustrate the value of TwinRCTs in future AD trials, we can use the average trial enrollment of the lecanemab, gantenerumab, and donanemab studies to estimate a typical Phase 3 AD trial of 1,890 patients. Running this trial as a TwinRCT could reduce the control arm size by up to 35% or 330 patients, with a 1:1 randomization ratio. Assuming it takes one month to enroll 50 patients, reducing the control arm by 330 patients could shorten enrollment timelines by six or more months, helping to determine the effects of a treatment sooner.

Since TwinRCTs use robust statistical methods to ensure control of type-1 error, sponsors can be confident that the evidence will meet the rigorous regulatory requirements for marketing authorization. In fact, the European Medicines Agency released its final favorable qualification opinion in September 2022 for using Unlearn’s regulatory framework for running Phase 2 and 3 TwinRCTs.14 This qualification opinion represents the first time a regulatory body has formally supported a machine learning-based method for reducing sample size in pivotal trials.

In addition to enabling faster, smaller, regulatory-suitable trials, TwinRCTs also honor the valuable contributions of past trial participants by leveraging existing patient data from previously completed clinical trials. A TwinRCT’s AI-driven approach takes advantage of high-quality, comprehensive datasets — even from ‘failed’ trials such as Roche’s GRADUATE program — to train machine-learning models that generate precise prognostic digital twins for AD trial patients.

More patient-centric trials will move research forward

As we eagerly await the leading Phase 3 readouts, our thoughts turn to the millions of people living with AD who urgently need these new drugs and the tens of thousands of trial patients without whom clinical research wouldn’t be possible. We hope for good news while remaining steadfastly committed to the effort of innovation in the clinical trial space. Hundreds of AD trials are being run today; who knows what life-changing treatments they will yield? But one thing we know for certain — their success will depend on meeting enrollment targets. By offering a trial design that mitigates fears of placebo and provides patients with a more likely chance of receiving the experimental treatment, these targets can be met more efficiently. Unlearn’s TwinRCTs help overcome participation challenges with smaller control arms while generating the regulatory-acceptable evidence needed to move research forward in less time.

Download the results of Unlearn’s reanalysis of the ABBY study, presented at AAIC 2022.