Diagnostic utility of threshold tracking TMS in ALS – STARD study
One of the running themes of the symposium was that we need more tools for early and reliable diagnosis of ALS. This presentation talked about assessing transcranial magnetic stimulation (TMS) as a diagnostic test.
Clinical examination has been the gold standard for identifying signs of upper motor neurone (UMN) involvement in ALS (e.g. the Awaji criteria). Dr Geevasinga suggested that we should now be moving towards transcranial magnetic stimulation (TMS) instead.
The conclusion from this study was that TMS testing confirmed cortical hyperexcitability as a feature of ALS, and that the short interval intracortical inhibition (SICI) parameter is a reliable diagnostic test in differentiating ALS from other mimic neuromuscular disorders. Because of this, it was suggested that the threshold tracking TMS technique could be used to complement the current diagnostic criteria, which could i) aid in earlier diagnosis and ii) aid in earlier recruitment into clinical trials.
Structural connectome analysis in ALS at multicentre level: a controlled study in 200 patients
Out of everything I saw in the whole symposium, this was the talk that I found the most exciting, as it directly relates to what I’m working on. I came away from this talk with many ideas, questions and a few interesting answers as well.
This study was, in some shape or form, representative of most of the difficulties experienced in MRI studies:
- It is expensive in time and money to scan patients, so most studies have small numbers
- Small numbers make it harder to generalise results (have you done your power calculations for your study?)
- One solution is to combine scans sets from multiple centres, which creates its own problems:
- For starters, different centres physically have different MRI machines
- As well as the major differences between 1.5T/3T/5T/7T scanners, there are potentially even differences between two scanners of the same make and model
- Never mind that the centres could have used different sequences/acquisition parameters etc
One thing that everyone agrees on: MRI is a very powerful tool for use in ALS research. So, finding a way to successfully combine data from multiple centres would be extremely beneficial for research. However, “successfully” is the key word – any process has to be accurate and reproducible.
One of the many techniques used here was to have the data from some of the centres act as a “reference cluster”, and data from the remaining centres were then warped to this reference cluster.
After correction for covariates, data were combined from a total of 8 centres. The correction process is one aspect that I want to look at closely (it’s on my very long list of things to do), as I suspect I’m about to run into these exact problems in my research.
As a first result, a characteristic pattern was shown: a Fractional Anisotropy (FA) decrease along the corticospinal tract (CST). This pattern is not unexpected – an FA decrease in the CST in ALS patients compared with healthy controls (at the group level) is known to be a feature of ALS – but that is exactly what we want to see in a study investigating the feasibility of pooling MRI data from multiple centres. Finding exciting, novel and frankly weird results could well indicate that the pooling process had not worked properly.
Having said that, if a pooling process only ever produces results that are already established, it’s not much use for research into ALS – to jump into a metaphor, it’s a fine line between getting rid of the noise, and completely losing the signal.
To see so many centres collaborating on a single study, pooling their resources, and managing to produce results that tie in with known results and existing staging systems is both encouraging and exciting, and I sincerely hope we see many more of these multicentre studies in the near future.