The Panagora 7th Annual Digital Biomarkers in Clinical Trials Summit, held in Basel, Switzerland, was a hub of innovation, particularly highlighted by Roche’s presentation on remote cognitive monitoring for Alzheimer’s disease (AD). Thanneer Malai Perumal and Kirsten Taylor from Roche shared their pioneering work, showcasing the potential of digital biomarkers in transforming clinical trials and early detection of cognitive decline.
Remote Cognitive Monitoring: A New Frontier
Kirsten Taylor, a leader in neurocognitive, digital, and clinical computational biomarkers at Roche, emphasized the critical role of cognition in understanding brain pathologies. She explained that different brain disorders affect various brain circuits, leading to distinct cognitive dysfunctions. This specificity allows for the use of cognitive profiles in diagnostic processes, tracking disease progression, and estimating concomitant pathologies, such as the burden from vascular dementia in Alzheimer’s patients. Taylor elaborated on how clinical neuropsychology leverages these cognitive profiles to generate detailed assessments of cognitive impairment. She illustrated this with a graph showing cognitive domains on the x-axis and the degree of impairment on the y-axis. These profiles are disease-specific, enabling clinicians to differentiate between conditions like Alzheimer’s disease and vascular dementia, which have unique cognitive impairment patterns.
In the context of clinical trials, Taylor noted that cognitive profiling aids in several critical areas. It helps in screening participants to ensure they meet the study criteria, testing for treatment effects to evaluate the efficacy of interventions, and conducting subgroup analyses to understand how different populations respond to treatments. She referenced longitudinal studies that track cognitive changes in individuals over decades, highlighting that early cognitive declines often manifest in long-term memory, learning new information, and executive functioning. These insights are crucial for developing targeted interventions and understanding the progression of Alzheimer’s disease from its earliest stages.
The Roche Digital Assessment Suite for Alzheimer’s Disease (Ad DAS)
Taylor introduced the Roche Digital Assessment Suite for Alzheimer’s Disease (Ad DAS), a tool designed to measure cognitive functions that decline in AD’s preclinical and early stages. The suite includes tasks adapted for remote smartphone cognition settings, ensuring ecological validity and minimizing variability in outcome measures. The Ad DAS was developed in collaboration with the University of Oxford and Eli Lilly and includes tasks targeting episodic memory, executive functioning, semantic memory, visual-spatial working memory, psychomotor speed, and attention. Taylor explained that these tasks were designed to be administered in short bursts over several days rather than in a single, lengthy session. This approach not only reduces participant burden but also increases the ecological validity of the assessments.
For example, the long-term memory test in Ad DAS asks participants to recall objects they learned days ago, mimicking real-life memory challenges more closely than traditional clinical tests, which typically ask for recall of objects learned 20 minutes prior. The suite also offers passive monitoring options, allowing participants to use their private smartphones to collect data unobtrusively. This innovative approach ensures that the data collected reflects real-world cognitive functioning, providing a more accurate picture of cognitive health. The suite’s design also allows for the collection of longitudinal data, essential for tracking disease progression and the impact of interventions over time.
Pilot Study Insights
The pilot study for Ad DAS involved 120 participants across five sites in two countries, including healthy individuals, those with subjective cognitive decline (SCD), and early AD patients. The study aimed to assess digital tasks’ feasibility, adherence, and reliability. Taylor reported that the digital tasks correlated with specific brain regions as expected, confirming their validity. For instance, memory tasks correlated with hippocampal structures, while language processing metrics were associated with classic neurological areas. She provided detailed examples, such as visual learning and episodic memory correlating with atrophy in the occipital areas of the brain and visual working memory and response inhibition correlating with the insula, hippocampus, and inferior orbital frontal cortex.
The study also included comprehensive screening procedures, such as amyloid PET scans and magnetic resonance imaging (MRI), to estimate brain atrophy. Participants underwent 30 days of remote cognitive testing at home, with daily tasks designed to take only a few minutes. Taylor emphasized that the demographic characteristics of the groups were comparable, ensuring the study’s robustness. The pilot study’s design and execution provided a solid foundation for evaluating the feasibility and effectiveness of the Ad DAS in a real-world setting. The results demonstrated that the digital tasks could reliably measure cognitive functions and correlate with known biomarkers of Alzheimer’s disease.
Feasibility and Adherence
Thanneer Malai Perumal, Roche’s digital biomarkers technology and science lead, addressed concerns about the feasibility and adherence of remote cognitive assessments. He reported excellent feasibility, with 97.5% of participants completing the evaluations and 96.4% adherence over the 28-day monitoring period. Interestingly, participants with dementia showed higher adherence compared to healthy subjects, indicating strong engagement with the digital tasks. Perumal explained that the study’s success in feasibility and adherence was partly due to the human contact provided by clinical sites, which helped set expectations and integrate the assessments into participants’ daily lives.
He noted that more than 85% of participants rated their experience with the app as good or very good, and over 86% found the time commitment acceptable. On average, participants spent around 10.9 minutes per day on the assessments, slightly above the expected ten minutes for early AD participants and below for healthy controls. This high adherence and positive participant feedback underscores the potential of digital cognitive assessments to be integrated into clinical trials and real-world settings. The study’s findings suggest that with proper support and clear communication, participants are willing and able to engage with digital tools for cognitive monitoring.
Establishing Validity and Reliability
Perumal discussed the statistical methods used to establish the digital metrics’ construct validity, reliability, and group validity. The results showed good to excellent test-retest reliability for most metrics and significant associations with in-clinic metrics, confirming the construct validity of the digital tasks. He elaborated on the statistical analyses, including intra-class correlation for reliability, regularized ridge regression for convergent validity, and proportional logistic regression for group validity. These analyses accounted for clinical covariates such as age, sex, education, and site, ensuring robust results.
Perumal highlighted that the digital metrics could differentiate between known groups, such as early AD participants and healthy controls, with significant effect sizes. Notably, the episodic memory construct showed sensitivity in detecting differences between SCD participants who were amyloid-positive and those who were amyloid-negative, a distinction not captured by traditional clinical scales. This finding underscored the potential of digital metrics to detect early cognitive changes that conventional methods might miss. Establishing such detailed and reliable metrics is crucial for the future of digital biomarkers in clinical trials, providing a more nuanced understanding of cognitive health and disease progression.
Future Directions
While the initial results are promising, Perumal acknowledged that the digital metrics are not yet ready for primetime. The next step involves longitudinal studies to assess the sensitivity of the digital metrics to changes over time, aiming to perform as well as or better than existing clinical metrics. Perumal emphasized the importance of longitudinal sensitivity to change, particularly in detecting early cognitive declines. He mentioned that Roche is currently running a study to collect longitudinal data in similar populations, intending to present these findings at future conferences.
He also highlighted the collaborative effort required to achieve these results, thanking the patients, their families, and the principal investigators who believed in the feasibility of self-administered cognitive assessments in dementia populations. The ongoing research and development efforts aim to refine the digital tools and ensure they can provide reliable, actionable insights in clinical trials and real-world settings. As the field of digital biomarkers continues to evolve, the work being done by Roche and its collaborators represents a significant step forward in the early detection and monitoring of Alzheimer’s disease.
If you want to access the complete Digital Biomarker Summit event recordings, click here.
Moe Alsumidaie is Chief Editor of The Clinical Trial Vanguard. Moe holds decades of experience in the clinical trials industry. Moe also serves as Head of Research at CliniBiz and Chief Data Scientist at Annex Clinical Corporation.
