2023: When AI & Clinical Trial Innovations Converge
At BIO 2023, Alto founder and CEO Amit Etkin, MD, PhD, had an opportunity to connect with fellow industry leaders to discuss how data-driven approaches and patient-centric endpoints can promote successful clinical trials in the space, and which areas AI and machine learning show promise for accelerating psychiatric drug development.
Can new clinical trial approaches support success rates for CNS drug development?
The panel discussion “Beyond Drug Approvals: Clinical Trials that Actually Advance CNS Innovation” featured a leading policy expert, researcher, and psychiatrist alongside Dr. Etkin to discuss the promises and pitfalls of stratifying patients, machine learning approaches, decentralized studies, real world data, as well as sharing results to advance CNS innovation.
Until now, drug approval rates in CNS have been dismal, with mental health proving to be a particularly challenging area for development. Current trials lack a feedback loop and are often underpowered, hampering the comprehension of drug mechanisms. Early in the discussion, Dr. Etkin set the stage: “A lot of our treatments within psychiatry have been discovered initially through serendipity and reiterating things that we’ve found and made observations about, but not truly understood.”
Countless failures have resulted from trials leveraging outdated models and recycled approaches, with little diversity in drug mechanisms of action. This panel attempted to shift the industry standard for measuring clinical trial progress through the notion that real innovation requires understanding why – or more often, why not – studies did or did not hit their endpoints.
Through the course of the conversation, new approaches to increasing speed, efficiency, success and outcomes of trials were uncovered. Panelists shared how to build knowledge iteratively, using earlier stage trials to de-risk drug development and reduce late stage failures, as well as how to build trials to maximize clinical impact. Watch the full discussion:
How can psychiatric drug developers unlock AI's potential in mental health?
A second panel, “A New Framework for Applying AI to Mental Health Drug Development,” captured the input of several experts in mental health drug development, each of whom are leveraging AI and big data applications in different ways.
In the past decade, revolutionary breakthroughs in therapeutic areas including oncology and gene therapy have been enabled by precision medicine. Although tailoring treatments to a patient’s unique biology has proved beneficial and life-saving, psychiatry has failed to innovate in this way. Speakers shared how they have harnessed big data and machine learning to develop digital tools, diagnostics, and therapeutics that show potential to reveal and treat sub-populations of patients across CNS disease states.
To begin the discussion, Dr. Etkin provided his vision for the future of psychiatry: “20 years from now, my expectation is that, beyond a dictionary diagnosis, there will be scalable tests, just like in many other areas of medicine, to guide mental health treatments for our population. These could involve a test of cognition, brain function using EEG, or a measure in the blood – different biomarkers that guide the selection of treatment. While our diagnoses are not meant to be anchored in biology, our interventions will be biologically-driven.”
Panelists agreed that in our quest for scientific advancement, we can finally move beyond trial-and-error for psychiatric treatments. However, they cautioned that while abundant data exists, standardizing it for analysis is essential.
Widening research participation, through approaches like decentralized trials, can help reduce stigma and improve data quality and diversity. In parallel, shifting from labeling individuals by their symptoms to understanding underlying biological issues further promotes the de-stigmatization of mental health conditions.
The emergence of precision psychiatry underscores the need to recognize different patient responses and treatment efficacies. Identifying successful biomarkers involves deciphering recurrent patterns to enhance outcomes beyond existing standards of care. Integrating AI into this process introduces complexities tied to data quality.
AI's potential hinges on robust data sets, necessitating expanded patient numbers and thorough database exploration. Deriving meaning from machine learning at scale, making measurements as bias-free as possible, is vital to drive clinical value. This comprehensive approach, combined with patient-centric recruitment and study designs, holds the key to a more effective and compassionate era in psychiatric care.
In addition to sharing key learnings from other CNS disease states and recent data readouts, panelists shared how viable brain biomarkers are selected for and how stratifying patients will forever change psychiatry. Watch the full discussion:
