Biomarkers Flashcards
What is quantitative MRI?
Commonly used in clinical practice
The most common quantitative measurements are size/volume of an object or region and the distance between structures of interest
What are we looking for in conventional quantitative MRI?
Is there inflammation or not and how big is it?
What is an issue with MRI?
There are differences e.g. in contrast between vendors #
-We are trying to minimise these differences
What colour is commonly used to represent normal brain matter?
Purple
Why dont we always use quantiative MRI in routine MRI?
Because to quantify you have to repeat the scans many times and modulate the imaging parameters each time
this is costly in time and expense
What do we do with the repeated scans?
Model the changes and apply them to the data to apply more complicated analyses after
- this filters out information that is relavant to the vendor or machines
What is a limitation of this type of MRI for the patient?
Spend a long time in the scanner
But if we reduce the number of scans to reduce time patient spends in the scanner then the images become less accurate
Its a trade off between time and accuracy
When can we compromise on accuracy?
Depends on purpose of the scan
e.g. to assess brain mestastises in cancer- we dont need a high quality scan so we can compromise
When might detail be important?
When we assess inflammation- is there inflammation, how much/how big is the inflammation
Make comparisons between previous visits
Does it require dedicated interventions
What is normal appearing white matter?
Visually on a quantitative scan it appears normal but actually it isnt
What is important to know clinically?
Whether patients are on a fast path of decline e.g. need aggressive inteventions or on a slow path of design e.g. less intense intervention
What is a biomarker ?
A quantitative measurement plus something that allows you to interpret it
A quantitative imaging biomarker can be defined as an imaged characteristic that is objectively measured and evaluated as an indicator of normal biological processes, pathogenic processes or a response to a therapeutic intervention
What can perfusion scans tell us?
Measure how long it takes for contrast agent to get somewhere in the brain
- could be an indicator of stroke e.g. shows the area in danager of becoming infarct because the arrival time will be greater than 6 seconds
this can be effective at improving the outcomes of patients
What area can quantitative MRI be useful for looking at treatment outcomes?
Radiotherapy
What is the aim of quantiative MRI in radiology?
Find out whether treatment killed everything or whether there is reoccurance that could spread again after treatment
distinguish areas with damage caused by the cancer to areas still with cancer
What are qualitative biomarkers?
Different scanners produce different images of the same anatomy
Only image contrast can be meaningfully interpreted
The vast majority of all MRIs taken in clinical practice
What are quantitative biomarkers?
Produces the same results on any scanner anywhere in the world
Image intensity is physically meaningful (has units)
Requires additional scans (scan time) and more complex analysis
Trade off between scan time and accuracy
What are MRI biomarkers used for?
Can be used to study biology & disease
Can be used to inform patient management.
Must be accurate & cost effective
Why are biomarkers useful?
Typically we take histories, blood, urine samples, biopsies and ultrasounds but there is a gap in thne diagnostic spectrum
- none of these look in detail at the actual organs and this is where MRI fits in as a technique- none of them inform about the level of scale that you can get from an MRI scan
What can we compare MRI to, to get an understanding of disease in practice?
Biofluids
What is an issue with trying to track diseases with biofluids (blood and urine)
Traces of disease will be within biofluids but this is convoluted by other bodily activity and is also diluted so will be in small quantities
How is MRI better than biofluid assessment?
Detect disease or treatment effects in situ
Not confounded by changes outside of target organ
What are the limitations of biopsies
Biopsy- goes to the source of the problem, extracts a piece of tisses
- in the brain this is very invasive and dangerous- high risk for only a small piece of tissues
e.g. in cancer there is hetrogeinity- some areas of cancer may be dormant and some other areas may be very aggressive- taking the wrong part could lead to over or under- estimation of the cancer severity - none of this same sampling error in MRI
Don’t observe tissue in its natural state- often frozen, squashed between plates, not the same as seeing the tissue in the organ, MRI doesnt disturb tissue state
What is a strength of MRI?
Versatility
Morphology, size, thicker/thinner than normal, can use relaxation times and diffusion to look at microstructure, levels of oxygen, perfusion of tissue, blood flow, inflammation, and we can also see function
All non-invasively
What was a recent finding?
Gut-brain interactions that can directly affect diseases
No other techniques allow us to look at several organs in such depth and specificity
What are the unique features of MRI?
Comprehensive multi-scale assessment
(anatomy/structure/function)
Enables study of organ interactions
What what the first MRI biomarker than was proposed by Raymond Damadion?
Tumor detection by nuclear magnetic resonance in 1972
Led to the development of the MRI scanner
What did Aerts et al. (2003) find?
Found an automatic system to find better and more accurate predictions of biological state
What gave tise to the emergence of quantitative radiology?
How are imaging biomarkers linked to AI?
E.g. Validation of AI Cardiac MRI Measurements
Moving MRI’s
Detection of cardiac contours
Extraction of MRI biomarkers
Prediction of clinical outcomes
Autogenerate reports
Can automatically extract the moving heart from scans of a moving heart - this has important information for patient outcomes and generate reports that can be generated by a physician, important for medical imaging
How does this translate to research and clinics?
Imaging biomarker suppliers
- e.g. neurologist treating a patient with MS, use a form of things you would like to assess and you get a report sent back providing numbers and scales of normality or abnormality which you can base future decisions on
Are MRI biomarkers useful for clinical practice?
Increasing body of evidence for clinical utility
Now computational technologies & big data enabling
growth
Synergies with AI – both relies on biomarkers and
simplifies extraction
What is the practical deployment of MRI?
Increasing number of companies providing biomarker
services
Mostly centralised service model: accreditation of
local labs + central QA
Clinician’s perspective – very similar to blood/urine
test
What are the two translational gaps to cross for MRI biomarker development?
Still a time-consuming and costly process
Requires technical, biological and clinical validation + cost-effectiveness
