Slashdot Mirror

← Back to Stories (view on slashdot.org)

GPUs Helping To Lower CT Scan Radiation

Posted by kdawson on from the healthy-green-glow dept.

Gwmaw writes with news out of the University of California, San Diego, on the use of GPUs to process CT scan data. Faster processing of noisy data allows doctors to lower the total radiation dose needed for a scan. "A new approach to processing X-ray data could lower by a factor of ten or more the amount of radiation patients receive during cone beam CT scans... With only 20 to 40 total number of X-ray projections and 0.1 mAs per projection, the team achieved images clear enough for image-guided radiation therapy. The reconstruction time ranged from 77 to 130 seconds on an NVIDIA Tesla C1060 GPU card, depending on the number of projections — an estimated 100 times faster than similar iterative reconstruction approaches... Compared to the currently widely used scanning protocol of about 360 projections with 0.4 mAs per projection, [the researcher] says the new processing method resulted in 36 to 72 times less radiation exposure for patients."

17 of 77 comments (clear)

  1. Voodoo lied to us by jgtg32a · · Score: 2, Funny

    They said we'd use these processors for video games not medical technology

    http://www.youtube.com/watch?v=o66twmBEMs0

  2. Funny what drives the HPC market... by onionman · · Score: 3, Insightful

    It's remarkable that high performance computing is driven by video games. So, legions of PC enthusiasts and uber-gamers, I salute you for your contributions to technology! P0wn on.

    1. Re:Funny what drives the HPC market... by toastar · · Score: 2, Insightful

      Games are about the only thing that low-spec system can't do

      I take you never met someone who's job it is to solve the wave equation on very large datasets.

    2. Re:Funny what drives the HPC market... by Jeng · · Score: 4, Insightful

      Neither is Email or internet usage.

      I'm pretty sure the comment was for general usage, which is normally just Email and internet usage with some office apps thrown in. That is what a celeron with 2gigs of ram would be sufficient for.

      Yes, there are many many programs that are used in many fields that would not fit into the celeron with 2 gigs comment. I work in an office environment, we don't need massive processors, we don't need massive video cards, all we need is a low end processor with a good amount of ram.

      That is what I got from reading his comment, but apparently I am in the minority.

      --
      Don't know something? Look it up. Still don't know? Then ask.
    3. Re:Funny what drives the HPC market... by localman57 · · Score: 4, Funny

      I met one once. I pulled his underware up over his head, then took his lunch money.

    4. Re:Funny what drives the HPC market... by ljw1004 · · Score: 2, Insightful

      That sounds fun. Is it available on Steam?

    5. Re:Funny what drives the HPC market... by Score+Whore · · Score: 2, Interesting

      You and a couple of others in this sub-thread are defining the problem backwards. As near as I can tell you're approach is to look at computer A and computer B and then to say "B is five times faster than A, therefore I need B." The correct way is to lay out your requirements: technical, financial and SLAs for delivery of your "product." Then to identify the system you need.

      While it's nice to be able to cache gigabytes of data, the reality is is that 2 GB is a fuckload of memory. Say you have a 21 MP camera (a 5D Mark II for example) and want to do some imaging work. Give up 1 GB of your RAM to your OS and apps. The remaining 1 GB can hold more than six complete copies of your images at 16 bits per channel + 16 bits of alpha.If you've got 8 bits per channel then you can have twelve copies. A 10 megapixel/8 bits per channel image (sufficient for most commercial work), in that case 1 GB is enough space for twenty-five images in RAM simultaneously. For the vast majority of users that's enough. Yes, it's possible to have that not be enough, but that says more about the user than the system.

  3. CPU speed determines req. radiation amount? by iPhr0stByt3 · · Score: 4, Insightful

    So, they pump in all that radiation because the processor is too slow? Doesn't seem right to me. I would think if they could have simply put another $10000 into the machine (adding CPU cycles) to lower the required radiation they would have done that a long time ago. So is the use of a GPU just a side effect of some new technology that allows the machine to estimate or predict the image with a lower radiation dose? That GPUs are more effecient for some operations is nothing new - what's the real breakthrough here?

    1. Re:CPU speed determines req. radiation amount? by Barny · · Score: 3, Interesting

      Pretty much.

      The reconstruction time ranged from 77 to 130 seconds on an NVIDIA Tesla C1060 GPU card, depending on the number of projections –-- an estimated 100 times faster than similar iterative reconstruction approaches, says Jia.

      So in essence they have built a parallel optimised calculation system rather than an iterative one, and we all know the one thing CUDA and OpenCL do VERY well is parallel processing.

      It seems the real win here is the new code, it could run on a TI-82 calculator and still require that level of radiation, its just that its very well suited to GPU to crunch.

      --
      ...
      /me sighs
    2. Re:CPU speed determines req. radiation amount? by FTWinston · · Score: 4, Informative

      The TFA says that this tech is usually used prior to treatment, while the patient is in the treatment position.
      Because processing a limited number of scans into a useful model previously took several hours, they were forced to perform many more scans to get a more accurate picture with which to build their model - because they don't want to leave the patient lying in the scanner for 6 hours prior to treatment.
      With this improvement in processing power, they can produce the model from limited data in a feasable time.

      So the summary does actually describe the breakthrough quite well: It's not a new image processing technique for working with limited data, it's just new hardware allowing that process to be run in a quicker way. Yes they're using a slightly new algorithm, but I doubt that is a massive breakthrough in itself.

    3. Re:CPU speed determines req. radiation amount? by Anonymous Coward · · Score: 2, Interesting

      The real breakthrough is the development of Compressed Sensing/Compressive Sampling algorithms; this is just an application.

    4. Re:CPU speed determines req. radiation amount? by jandrese · · Score: 2, Interesting

      My guess is that each scan requires a considerable amount of processing to render into something we can read on the screen. Probably billions of FFTs or something. You can make a tradeoff between more radiation (cleaner signal) and more math, but previously you would have needed a million dollar supercomputer to do what you can do with $10k worth of GPUs these days, which is how they're saving on radiation.

      --

      I read the internet for the articles.
    5. Re:CPU speed determines req. radiation amount? by Zironic · · Score: 2, Interesting

      What's going on is that instead of taking a clear picture they take a crappy picture and have the ludicrously fast GPU clean it up for them. While you could have done that by just putting 50 CPU's in parallel the GPU makes it quite simple.

      The speed is important because their imaging is iterative, with the GPU they're apparently waiting 1-2 minutes, without the GPU it takes them 2-3 hours which is a rather long time to wait between scans.

    6. Re:CPU speed determines req. radiation amount? by Anonymous Coward · · Score: 2, Informative

      The technique is called iterative backprojection. The reconstruction process assumes an array of pixels which, at the beginning, are of some uniform value. It then looks at a ray of attenuation data from the CT projection (along this ray, the tissues in the target result in this degree of attenuation of the xray beam), and asks "how must the pixels along this ray be adjusted, so that their attenuation along the ray matches the data from the CT beam?". It does this for every measured ray taken during the acquisition, over many different angles. The more sparse the acquired data, the more iterations (and thus, longer) it takes to get a reliable (approximated) image.

  4. lower rad dose by SemperUbi · · Score: 4, Informative

    CT scanning is associated with an increased risk of cancer in children. This development will significantly lower that risk.

  5. Re:context by budgenator · · Score: 2, Informative

    "Our work, when extended from cancer radiotherapy to general diagnostic imaging, may provide a unique solution to solve this problem by reducing the CT dose per scan by a factor of 10 or more," says Jiang.
    It's probably applicable to diagnostic cone beam scans, which are the hot item in implant dentistry. The reason it's first applied to therapy scans is because the tissue surrounding the tumor suffers radiation from scattering of the therapeutic beam, making dosage reduction highly desirable.

    --
    Apocalypse Cancelled, Sorry, No Ticket Refunds
  6. The real hero here is compressed sensing by HuguesT · · Score: 2, Informative

    This has been said elsewhere in this thread, the real breakthrough here is due to compressed sensing, but here are some extra information:

    1- Compressed sensing basically used the idea that it is not necessary to sample an image (or a projection in this case) everywhere because natural data is fairly redundant. This is why you can capture a 10 Mpixel image in a digital camera and have it compressed to a 2 Mbyte JPEG file without losing much visible information. Compressed sensing basically does the compression *before* the sampling and not after. Researchers at Rice University for instance built a working, one-pixel camera using this brilliant principle.

    2- Compressed (or compressive) sensing was proposed by Emmanuel Candes and Terence Tao respectively at Stanford and UCLA. Tao is a recent Fields medalist. I recommend reading his blog if you like mathematics.

    3- This field is really less than 10 years old, it has completely turned on its head classical ideas about sampling-limited signal processing (Nyquist, Shannon, etc). It is a brilliant combination of signal, image processing and recent advances in combinatorial and convex optimization.

    4- However this is only the beginning. Because compression happens before sampling, you need to make so-called sparsity assumptions about the signal ; in other words you need to know a great deal about what you are going to try to image. In interventional therapy, precise imaging of the patient is made beforehand in a classical way (CT or MRI), and this kind of technique is only used to make fine adjustments as therapy is ongoing. This is extremely useful and safe because of lower radiation output and because the physicians know what to expect.

    5- Here the GPU is useful because it makes the processing fast enough to actually be used. It is an essential brick in the application, but of course not in the theory.

    Best.