Slashdot Mirror
New Algorithm Could Substantially Speed Up MRI Scans
An anonymous reader writes "In a paper to be published in the journal Magnetic Resonance in Medicine, researchers detail an algorithm they have developed to dramatically speed up the process of producing MRI scans. The algorithm uses information gained from the first contrast scan to help it produce the subsequent images. In this way, the scanner does not have to start from scratch each time it produces a different image from the raw data, but already has a basic outline to work from, considerably shortening the time it takes to acquire each later scan."
Not to say the patent system isn't unfair/broken, but it's often been said that we shouldn't patent algorithms, and they should be OS etc. But how about if it's someone's livelihood, and years of research went into the algorithm, like I bet this did? Should programmers be given less due just because they're not working with real materials or making physical inventions?
Why OpalCalc is the best Windows calc
in an electromagnetically charged TUBE is still unfun. Where's my tricorder already?!
Forget about the staff's time, the facility, the equipment - they'll raise the price of an MRI to pay off whichever patent troll gets the rights.
They are boosting the stargate? MRI always reminds me im crossing a stargate....
TFA seems to imply that a patient spends more time in the tube because of the slow processing of the images.... Is there a reason they need to be in the tube while the images are developed? Maybe they need to retake images if they didn't come out? Barring that... wtf? Why not just do the scan and process the images offline on another piece of equipment? Storage is cheap nowadays, and caching the received RF data temporarily should be feasible.
http://www.masturbateforpeace.com/
There are too many people completely misunderstanding that innovation and inventions and new technology drives costs down.
MRI shortens the time it takes to understand what is wrong with a patient, the precision of understanding goes very high, there is very little need (if any) for exploratory surgery with good imaging. Yet people (and on THIS site!) argue that technology is almost supposed to push medical costs higher. This is absolute nonsense.
When this new algorithm makes it into the MRI machines, do you suppose the costs of using the machines will go or down.
Think about it: if there is more throughput with the same equipment, so there are more people being scanned faster, do the costs go up or down? It's the same salaried stuff. It's the same electricity! It's the same machine that's being amortized (crazy government not letting businesses write cost of capital purchases against taxes in the first year, forcing even loans to be taken for that!)
So think about this new algorithm and every new technology that has been developed in the last 60 years. Now think about the actual costs of health care going up. Does this compute?
Now realize that the other factors include government created inflation, government money - Medicare, government subsidies and protections via FDA (you think you can build a better MRI machine in your garage and sell it more competitively to make a profit? Do you HAVE half a billion dollars to push it through FDA?)
This technology is great, but it only really brings costs down, increases efficiency and actually helps people by not participating in destruction of economy in free markets, not in any type of economies that are government controlled.
Also - there is much to do in medical field. There are all sorts of needs there, but with all the regulations you won't be able to just build a company making this stuff, coming up with new innovations and inventions. You can't do it in USA or the rest of the Western world because of government.
If you have these ideas of profiting from providing people with better technology - learn Mandarin.
You can't handle the truth.
... when someone can make an MRI scan quieter. Going through an MRI scan to diagnose the causes of my migraines is almost enough to give me a low-grade headache on its own. Most MRI scans sound like an Atari arcade game cranked up loud to try to overcome the noise of the active quarry it was installed in.
Granted, it beats being irradiated, but if I could change anything about it I would make it quieter. Hell I'd tolerate it taking longer if it was quieter; they are kinda cozy and I could take a nap in the scanner if it wasn't so damned loud.
Damn_registrars has no butt-hole. Damn_registrars has no use for a butt-hole.
+5 for having knowledge of the matter being discussed
I'd be more interested to know whether this will speed up acquisition of BOLD (blood oxygen level dependent) signals during fMRI (functional magnetic resonance imaging). The temporal resolution (time between frames) of fMRI has been a huge limiting factor in research. Increasing the rate even by a modest factor of 3 would go a long way to towards making fMRI competitive with EEG (electroencephalogram) (which can collect data in real time but with very little spatial resolution). While you're wikiing, check out DTI (diffuse tensor imaging).
Since when is caching a "new" algorithm? From the basic info in the summary it appears this is just a cache of some initial processing that is then reused for subsequent image generation passes...
In the end, Dr. House will still barge in, pop a few Vicodin and flip the patient over, revealing the infected tattoo on his ass* (the one his buddies got for him after a night of heavy drinking) that is the cause of all the peculiar symptoms.
* Rabbit tracks down one cheek into the gully and Elmer Fudd with shotgun on the other saying, "Come out of there you Qwazy Wabbit!"
Have gnu, will travel.
Link to paper: http://www.rle.mit.edu/stir/documents/BilgicGA_MagResMed2011.pdf I guess the reason this shows up on slashdot is that it was on MITnews (and of course the work was done at MIT). It is nothing really groundbreaking (or novel for that matter). They use Compressive Sensing where they assume that the different scan types (T1, T2,etc) have a similar structure (same sparsity profile, enforced through shared precision hyperparameters in a Bayesian formulation).
The reason these algorithms can actually give a significant gain, is because most MRI scanners and other medical scanners use computers that we ourselves abandoned 10-20 years ago. I've seen SGI machines not being sold new anymore for over 15 still in active use on digital Rontgen scanners. The MRI they took of me was processed on a P3 PC.
I was promised a flying car. Where is my flying car?
Why was this modded down? Parent is referring to the shortage of helium, which is needed for cooling the superconducting magnets.
While I can not comment on the content or merit of this work since I have not read the paper I can say that there is a lot of research going on in the area of MRI and a number of exciting things are happening. I work in a research group that works on impressive MRI techniques combined with algorithms that allow taking entire MRI movies (not just photographs). This is extremely interesting when looking at the beating human heart, speaking humans or moving joints. We are really just beginning to explore the potential this technology has for clinical applications. More information including impressive videos can be found here: http://en.wikipedia.org/wiki/Real-time_MRI.
The idea behind the extremely fast acquisition times is extreme under-sampling combined with iterative algorithms that are able to produce high quality images despite the incomplete measurements. This algorithms also exploit the fact that between two frames there is not a lot of change in the images since they are taken at time intervals of only dozens of ms. They are of course more complicated than simple fourier transforms as they model the acquisition process in more detail.
Hey dude,
Want to reduce scan times? Check out compressed sensing MRI [1]. You should be able to take way fewer scans than thought possible with 20th century math. Regularized reconstructions are the new hotness, but don't take the word of a Slashdot user who says "dude" and "new hotness"; read these fricking things.
[1] M. Lustig, D. Donoho, J. Santos, and J. Pauly, “Compressed sensing MRI,” IEEE Signal Processing Magazine, vol. 25, no. 2, pp. 72–82, 2008.
Expected time to finish is 1 hour and 60 minutes.
Does anyone have experiences with (brain) MRI's for kids (10 and under)? In order to 'keep them quiet' enough for the scan (and scary noise!), is a full-blown anaesthesia necessary? What options do exist?
MRIs do have false positive rates. An item may appear in an image that does not exist in the patient, is misinterpreted by the reviewing physician or does not reappear on subsequent MRI imaging. False positives create unnecessary patient anxiety, increase medical costs, result in unneeded follow up testing, unneeded medical procedures and misdiagnosis. Increasing the speed of an MRI image must be balanced against any increase in the false positive rates. It does not appear that any data is given on the new false positive rates and whether they remained static, increased or decreased. In many ways, patients would gain more form a decrease in an initial false positive rate than they would from an increase in imaging speed. An increase in false positives counteracts the patient gain from an increase in speed.
Still, although they are old, they are adequate because the speed at which the data is coming out of the scanner is "pain-in-the-ass" slow.
Scanning is very slow. Emit the radio impulse, wait a little bit, read the signal. Repeat this (with varying magnetic fields) for the number of data point you need. So the current bottle-neck is not the ageing computer, but the data acquisition it self.
The current trend is to find a compromise between the number of data point.that has to be acquired and the definition and quality of the image.
To few data points, and the output is completely blurry or full of artefacts.
To many data points and everybody (patient and staff) has plenty of time to die of old age until the scanner has finished repeating the "emit-wait-read" cycles.
This new algorithm tries to reuse some of the data already acquired from the previous picture, so less data is needed for the new picture. Less data point, less slow waiting cycles, less time spent waiting in the tube, happier patient.
It has nothing to do with the fact that the computer cannot run Crysis 2, but only Quake3.
"Sufficiently advanced satire is indistinguishable from reality." - [Tips: 1DrYakQDKCQ6y52z6QbnkxHXAocMZJE61o ]