h a l f b a k e r ycarpe demi
add, search, annotate, link, view, overview, recent, by name, random
news, help, about, links, report a problem
browse anonymously,
or get an account
and write.
register,
|
|
|
Science magazine had a blurb, as well as a paper about a researcher who estimated that 1 of every 1000 persons with a computed tomography body scan got cancer. thus of about 2.4 million scans, 24,000 lethal cancers were being caused each year (an amount comparable to the 30,000 annual road deaths).
This is a very simple approach to reducing that risk 20 or 100 times. place a big chunk of bismuth with numerous very tiny microchannels in front of the radiation source. The radographic photions only travel linear paths through the microchannels to pass through the body to reach the imaging surface.
think of each square millimeter of a radiograph as a grid "#" if you just created a microchanel at the center, then 1/9th the radiation absorbed dose is transmitted, yet the intensity of activation (amplitude) on "film", or more likely a semiconductor imaging surface is preserved at that one place. Then software just draws lines between all the known data points to make a millimeter accurate "interpretive" CT scan Now thats just 1/9th the radiatuion dose. It is pretty simple to go with the idea the a 5 on a side square would also function, giving 1/25th the usual radiation absorbed dose. Also the physician can always specify different bismuth microchannel chunks if they prefer anything from a 9th, a quarter or even a full dosing image.
a slightly fancier approach uses the multiplier effect to create CT scans with just 1/100th the previous amount of radaition. (1/25th * 1/4th = 1/100th) That is to have the software guess what it is looking at with each pixel of a scanning radiographic beam. if the last 3 pixels, as well as mapping derived shape (sort of like OCR or nudity autodetection) suggest a soft tissue type then the energy to the radiation source is reduced, to provide the more minimal amount of energy required to image that kind of tissue). Its possible that permits a quarter of the usual radiation dose. alternatively, I think (but do not actually know) that actual cancer risk from radiation is related to highest energy, thus its possible that say ten 1/10th strength radioactive photons reaching the CCD imager as 100 microsecond pulses has half or a quarter the risk of one dose of radiation ten times larger.
Update: I think this could work with X-rays, just put a plate with lots of holes in it in front of the x ray source, then use the lower resolution that is capable of doing effective imaging.
as always my ideas are public domain
Science Magazine link
http://www.sciencem...f-81fc-861836d61860
[beanangel, Oct 26 2012]
This link actually has some non-login text
http://www.aboutcan...sk_brenner_2007.htm
[beanangel, Oct 26 2012]
[link]
|
| |
Hang on. I got the impression that, in all CT
systems, all the X-rays going through the patient are
registered. In other words, all the X-ray energy
delivered into the patient contributes to the final
image. |
|
| |
If this is true, then channelling the X-rays will just
create a narrower beam, requiring a longer scan and
the same total exposure. No? |
|
| |
Its interpretive. kind of like replacing a watercolor with a pointillist painting, where the software draws the lines between the |
|
| |
"pointillisms" to recreate the look of an orginal watercolor. the tiny areas of radiation are the result of the microchanneled bismuth chunk |
|
| |
It is true that the resolution is 1mm rather than say 1/25th of a mm, its just that I think think a human clinician is looking at larger detail than the 1/10 of a millimeter size |
|
| |
Right, for those giant hairline fractures. |
|
| |
Actually, my understanding is that one of the
advantages of CT (and MRI) is the high level of
detail
they provide, revealing things such as stress
fractures that don't show up on normal X-rays. |
|
| |
I'm not sure how this idea is any different from
taking a low-res picture and scaling it up. You're
just using the computer to interpolate detail that
doesn't exist, the same as if you just doubled the
resolution of the image in Photoshop. In fact, it's
worse than that, because it might mask details that
are too small to be resolved in the image, giving
the impression that there's nothing wrong with an
area that may have a problem. Big [-] on this one. |
|
| |
CT also doesn't use point results (I think). It use the distortion/diffraction to produce it's image. This means that narrowing of the input beam will produce an impressionist image, not pointillist, since there will be less total data still covering the entire scanned area. |
|
| |
//CT also doesn't use point results (I think).//
True, but it probably could. |
|
| |
However, [ytk] is right, I think. You might as well
use a lower intensity and get a noisier, less
contrasty, image. |
|
| |
In the end, you put a certain amount of X-rays
into a person and get a certain amount of data
out. Whether you reduce the overall intensity or
image fewer points with higher intensity and then
interpolate, the end result won't be much
different. |
|
| |
Actually there might be one difference. The
relationship between carcinogenicity and
radiation dose is non-linear and (I think - not sure;
[EDIT - the jury is out])
it flattens a little at high doses. Therefore,
imaging a few points at high (normal) intensity
may actually be worse than making a continuous
image at low intensity, if the total delivered dose
is the same. |
|
| |
Ultimately, the solution is either to use a
different imaging technique, or to improve the
sensitivity of the detectors (which, I imagine,
people are working hard to do). |
|
| |
// improve the sensitivity of the detectors // |
|
| |
Umm .. there are quantum noise limits on the sensetivity of the detectors. And CT (to a certain extent) relies on scattering. Yes, ionising radiation dose is stochastic.
Much like a flash photo taken in darkness, the quality of the image is dependant on both the duration and the intensity of the flash. It all depends on how the detector integrates the incident energy. |
|
| |
// a researcher who estimated that 1 of every 1000
persons with a computed tomography body scan got
cancer. thus of about 2.4 million scans, 24,000 lethal
cancers were being caused each year // |
|
| |
Let's take a look at some of the assumptions you're making
here: |
|
| |
1) a single researcher making a fairly general statement
does not make something true. There's a 'researcher' who's
spent most of his life 'proving' that the Great Pyramid at
Giza is actually an ancient landing pad for giant alien
spaceships. His work is taken so seriously that it has led to
a feature-length documentary called 'Stargate' and a
ground-breaking investigative journalism series called
'Stargate SG-1'. |
|
| |
2) 'cancer' does not necessarily mean 'deadly cancer'. My
aunt had cancer; some doctors cut it out of her and
replaced it with a rather attractive pouch of saline (I say
this objectively, mind you). Now my aunt no longer has
cancer, and she is not dead. |
|
| |
3) 'estimate' does not mean 'exact figure'. By scooping up a
bucketful of gravel and examining the contents I can
estimate that for every ten rocks in my driveway, one of
them will be black, but there's no way to come up with a
specific ratio unless I count all of the rocks in my
driveway. |
|
| |
Science, [beany], science! |
|
| |
What [Alterother] said, except that the bit
about // the Great Pyramid at Giza is actually
an ancient landing pad for giant alien
spaceships // which is actually correct, but
only by pure coincidence. |
|
| |
// replaced it with a rather attractive pouch
of saline // |
|
| |
Was it expensive? Is it socially acceptably to
have the entire person replaced with a more
attractive bag of saline? Just curious … no-
one specific in mind, you understand … |
|
| |
they don't use CCDs or at least the scanners in any of
the hospitals I've come across don't. There's work
going on to implement electron miltiplier CCDs.. EM
CCDs, which with a quantum efficiency of >90%, gives
you a 20% drop in the amount of radiation necessary. |
|
| |