ALS TDI speaks with Dr. Gupta to learn how new digital biomarkers could pave the way for faster clinical trials and more effective treatments, as detailed in a groundbreaking paper published in Nature.

In the quest to end ALS, finding more accurate ways to measure disease progression stands out as a pivotal challenge. Digital measures of ALS progression, or digital biomarkers, are gaining traction as highly sensitive and objective tools that could not only accelerate clinical trials, but also make them more effective.

This discovery of new digital biomarkers is a primary goal of the ALS Research Collaborative (ARC), an initiative spearheaded by  the ALS Therapy Development Institute (ALS TDI). Through the ARC study, ALS TDI seeks to learn more about the disease by gathering comprehensive data from people with ALS  and then sharing it with researchers all over the world.

One avenue pursued by ARC to achieve this involves equipping  people with ALS with wearable accelerometers – smartwatch-like devices that meticulously track movement. By wearing these devices on each wrist and ankle, ARC study participants can generate data about how their disease is affecting their movement over time.

Recently, researchers from Massachusetts General Hospital (MGH) partnered with ALS TDI researchers to analyze these data. Their analysis of ARC accelerometer data, demonstrating  that these devices can be used as reliable measures of ALS progression, was published in the esteemed journal Nature in 2023.

Dr. Anoopum Gupta, M.D., Ph.D., a neurologist at MGH and the paper’s lead author, joined us on the Endpoints podcast last year to discuss the paper and what it means for people with ALS. Here, we present a condensed version of this conversation:

Q) What specific questions about ALS progression were you aiming to answer with this paper? How did the ARC data set help you to uncover these answers?

 A) [The ARC Study] is a really unique dataset that ALS TDI developed over many years. In this study, individuals with ALS wore a sensor on each limb for about a week at a time, every month, [sometimes] over the course of multiple years. 

We wanted to know, across this cohort, “how is movement of two arms and the two legs changing as the disease is progressing over time?” “Can we develop a method that can characterize those changes very sensitively, so that we can detect smooth changes over time, rather than the somewhat discrete changes over time that we often get with our existing rating scale?”

Q) How do these digital measurements differ from current methods useed to measure ALS progression?

A) So, the hope with using digital technologies to measure some of these motor functions is that you can have a more continuous measure and that you can detect very small changes over time. 

The gold standard in ALS is the ALS Functional Rating Scale Revised (ALSFRS-r). That’s composed of four main components that capture the ALS phenotype. Those include:

  • Fine motor function, or how an individual uses their hands.
  • Gross motor function, or how people walk and their balance.
  • Bulbar, which captures speech and swallowing.
  • And respiration, which looks at an individual's breathing. 

The ALSFRS-r has traditionally been performed by a neurologist or an ALS clinician who sees the patient and answers several questions in each of those different categories. There's also a self-report version where individuals can answer those questions on their own. That is a very useful way of characterizing the changes in ALS because it can be performed at home on your own, and more frequently over time.

The ALSFRS-r has advantages and disadvantages. One of the disadvantages is that there's an element of subjectivity in those questions, and there's also a lack of granularity. Each of those questions are answered from 0 to 4. But if somebody falls between those choices, you don't get that level of precision.

Q) What were some of your key findings after examining the ARC accelerometer data?

A) We found a lot of interesting things in this dataset. One was that if you look at each of the four limbs, the two arms show similar trajectories over time, for most individuals. The two legs are also very consistent over time. But, in many individuals, there's a divergence in the movement changes that you see in the arms and the legs. In some individuals, the movement changes that are occurring in the legs are happening faster than the movement changes in the arms.

The recordings from the four limbs really allowed us to look at the similarities between the left and the right side. We saw similarities between the left and right arm and between the left and right leg, but differences between the arms and the legs.

Another thing that we saw was that the movement characteristics that we were measuring – movement at home during natural behavior – very strongly corresponds with how people report on their fine motor function [on the ALSFRS]. We saw changes in the natural movements of the arms that very closely reflected how people were having difficulties with writing or eating. Similarly, the movements of the legs corresponded very closely with individuals’ reports about their walking and balance. So, there's additional information captured when you look at the upper limb versus the lower limb, and you characterize the movements.

Q) What implications do these findings have for the broader goal of discovering better methods for measuring ALS progression

A) One of the challenges in ALS and other neurodegenerative diseases is that any drug that's being developed has to show that it is slowing how the disease changes over time. To do that, you have to be able to measure that change over time.

Any tool that allows you to measure smaller changes in shorter intervals would allow you to shrink the duration of a  trial or the number of participants that are needed for that trial.

What we found in this paper was that when you combine the information across the four limbs, and you take the limb that is changing the fastest over time, that limb is changing faster than the changes we see on the ALS Functional Rating Scale. And since that's changing faster, if you use that as your measure of disease severity, you can shrink the size of your clinical trial from about 120 individuals per arm to 70 or so individuals per arm.

That reduction in the size of the trial means reduced cost of trial. It means less burden on individuals with ALS, and it can increase the incentives for companies to develop drugs for ALS. If the cost of doing those trials and the burden of doing those trials is reduced.

Q) What are the next steps in terms of getting accelerometers approved to be used as an endpoint in ALS clinical trials?

A) This study included longitudinal data for about 180 individuals. We had 376 individuals that had cross-sectional data – meaning data for at least one time point, but not data that spanned at least nine months of the year. With a dataset of that size, you can feel confident that it can generalize to the clinical trial population. So, I think this technology is now ready to be used as an exploratory or secondary endpoint in interventional studies.

One of the things we don't know is responsiveness of the measure to intervention. In a clinical trial, you want a measure that is likely to respond to whatever drug therapy is given. It will be very useful to compare how the sensor-based measure changes at each limb in response to the therapy, compared to how the ALSFRS-r, or other outcome measures. I think we're at that stage now where this can begin to be used as exploratory and secondary endpoints in interventional trials, and we'll learn a lot about the properties of this assessment approach through those clinical trials.

To hear the full episode of the Endpoints Podcast featuring Dr. Anoopum Gupta, click here.

What to Do Next: