Takeda Elaborates on the Future of CNS

Takeda’s Atul Mahableshwarkar offers insight into central nervous system drug development and how biomarkers can improve therapy development of this discipline.

In a previous interview, we discussed some of the challenges with CNS, and how biomarkers can improve CNS therapy development. Atul Mahableshwarkar, Senior Medical Director at Takeda, offered a different perspective with CNS drug development at ExL’s 2nd CNS Clinical Trial Forum. In this interview, Atul will elaborate on his perspective. 

Moe Alsumidaie: What are the biggest challenges facing CNS trials?

Atul Mahableshwarkar: A major challenge is properly understanding the illnesses, as we all look at illnesses based on the different categorical diagnoses. We do not know the pathophysiology of psychiatric illnesses and so a single diagnosis could be made up of a number of different conditions. There is a lack of complete knowledge of the brain and brain illnesses, and this leads to challenges of designing and running trials. There is also the problem of methodology. Since we do not have clearly defined assessments, we need to rely on asking patients questions or patient reported outcomes (PROs). This results in great data variability and difficulty in translating the data into clearly defined endpoints. Additionally, you may include people who may not have the illness itself or their illness may not be as severe, to take part in the trial. That makes it much more difficult to qualify patients and detect signals of efficacy. An obstacle also arises from where the trials are conducted. There are issues of translating scales and interviews to culturally sensitive norms so that the information is consistent. Healthcare systems in different countries tend to be different. The kinds of patients coming in add to the variability and data heterogeneity. Thus, CNS studies involve a combination of challenges arising from research design, methodology, practicality, and the complexities in the science itself.

MA: What are the biggest factors that contribute towards CNS study failure?AM: When looking at issues leading to trial failure, there are some common themes that emerge. Trial designs may be too ambitious and try to get too many answers to too many questions. This increases burden on sites and patients participating in trials. There is the classic quandary of cost, quality and speed and the practical observation that one can only achieve any two of these three. Attempting to get things done with speed, while trying to achieve quality and cost leads to problems in recruitment and study completion. One way to manage some of these issues is through computerized approaches. There are solution providers who have created computer generated approaches to asking questions and getting answers from patients. These answers can then be used to rate symptom severity on rating scales which could be the end point in a trial. While much work needs to be done for this approach to be acceptable for assessing efficacy in registration trials, nonetheless if successful this would be a big step in overcoming some of these challenges.

This approach would reduce data variability, but that would still be an intermediary step to get to a point, ultimately, where will be collecting data directly from patients; the next generation is not of PROs, but of DROs: Device Reported Outcomes. With DROs, no one asks the patients any questions. For example, in a depression trial, the activity levels (number of steps taken), the number of conversations, the type of words used, would vary in people suffering from depression. Data collection is not based on answering a question when asked, but how you go about your daily routine. These will be captured by the devices that they carry which will be translated into assessments and endpoints. There is a lot of work that needs to be done before DROs will be ready for regulatory approval. Nonetheless, the DRO approach will help us improve study efficiency and data quality.

MA: The use of biomarkers has changed the way we develop oncology therapeutics. Do you think that incorporating biomarker research will enable us to develop targeted therapies in CNS?

AM: Absolutely, however, how long it might take is still to be determined. I assume it would be a circuit-based approach. The brain is interconnected and there are multiple circuits so we need to be able to identify the issues with the circuit. The treatment response is, to some extent, based on genomic expression, whether it is genetics, the epigenome or our microbiome. A number of things need to be understood and integrated but that work is ongoing. With biomarkers, we may be able to for instance identify different subsets of populations in a disorder such as Major Depression Disorder, or conversely be able to identify the source of a certain symptom let’s say attention; patients may have symptoms because of deficits in attention, hypomania, cyclothymia or major depression. As our understanding of CNS biomarkers advances, we may be able to identify appropriate sub-populations to develop more effective targeted therapies.

MA: CNS has been stagnant for quite some time, especially in psychiatric indications. Do you think that we are on the verge of a breakthrough?AM: We are still a few years away from such a breakthrough. Once we are able to better sort and understand the source of symptoms, we may be able to identify what may not work in a broad population, but may work for a particular sub-population or some form of a genetic expression and get better treatments for patients. With such approaches not only may we be able to develop new and effective treatments but we may also be able go back and look at different drugs that may have failed in the past but could get new efficacy data to regulatory standards. There are a large number of mechanisms which have been tried and failed and we may be able to revive many of those therapies to match them with the target population. Many of failed therapies have been sitting on the shelf and their patents have expired; reviving those therapies may be much less costly to develop, and may bring the prices of such new medicines down as well. It is not exactly a personalized medicine approach, but looking at discretely identifiable subpopulations who would respond to targeted treatments. Not only will these therapies be more effective, but because they are specific and discrete, they may be even better tolerable.

Moe Alsumidaie, MBA, MSF is Chief Data Scientist at Annex Clinical, and Editorial Advisory Board member for and regular contributor to Applied Clinical Trials.