Click on image to enlarge 

Diagnosing and tracking the progression of neurodegenerative diseases – in life – is a challenge. There are many brain disease indicators, or ‘biomarkers’, validated by science. All are works-in-progress, in the quest to solve the complexity of these diseases.

A common tool used to track brain disease is the brain scan – typically magnetic resonance image or MRI. Structural MRIs provide a window into the brain and its complex anatomy and makeup.

‘Brain volume’ is a quantitative measure that is relevant both clinically and for research.

Brains typically shrink over time, especially in people living with neurodegenerative disease. However, some structures or parts of the brain, such as the ventricles*, actually grow in time. Tracking these volume measures over time is a crucial tool in managing neurodegenerative and other brain diseases.

“We judge brain volumes in our patients on daily basis, taking age into account. Shrinkage in different regions of the brain can be seen even in cognitively normal elderly people, while some patterns of shrinkage may suggest particular disorders. We examine clinical MRIs to assess for incidental findings which may require an action – for example silent strokes,” said Dr. Sandra Black, Cognitive and Stroke Neurologist at Sunnybrook Health Science Centre.

Dr. Black continues: “MRIs also allow us to get a general impression of shrinkage patterns and the degree of visible small vessel disease in the brain. We’ve always believed that it would be even better if we could verify our subjective impressions of brain health by knowing the actual total volume of the brain and its ventricles. If there was a quick way to do this, it would be an important tool for disease assessment and tracking. Happily, with recent advancements in our lab, we can now do this.”

Challenges with MRI processing

MR images are very large. Translating these large images into quantitative measures, which can be compared to others, is a challenging and time-consuming job.

Interestingly, there are no widely available tools that do this well in neurodegenerative populations. Lab technicians and researchers typically access open-source tools and adapt them for their lab. To that end, significant advances have recently been made in Dr. Black’s lab, which are being shared with the world.

As outlined in a recently published paper, Edward Ntiri, Dr. Maged Goubran and others set out to devise improved algorithms to accurately and efficiently delineate and measure:

  • Whole brain volume – a useful measure for overall brain health for anyone studying or tracking brain disease
  • Ventricular volume* – of relevance especially in neurodegeneration

Using ‘deep learning’ techniques, Ntiri et al developed algorithms, or computer programs, that are able to dramatically improve these measurements of brain volumes — vs. existing tools.

In addition, these new algorithms enable researchers to cut down the time to do this from up to six hours per image, to a matter of seconds.

Why is this Important?

Significant progress in science is achieved when sample sizes are large, allowing findings to be validated and replicated. Tools such as the new algorithms described above, help make such scientific innovation possible, while also helping to deliver accurate personalized assessments. These advances serve both a practical, clinical function, and a research function – the latter of which is designed to lead to big discoveries.

“The new tools we’ve devised at our lab allow us to move away from time consuming manual or semi-automated pipelines to accurate and quick, fully automated ones – when calculating structural brain metrics. We validated our tools on complex clinical populations with brain lesions and comorbidities, and we anticipate they will be very useful in other labs studying neurodegeneration. We have proudly shared our tools with the research community and welcome collaboration”, said Dr. Goubran.

Read abstract here.

For algorithms: icvmapp3r and ventmapp3r.

More on neuroimaging advances through ONDRI research to come.