Why do PET scanners need patient-specific absorption calibration?
February 24, 2006 12:58 PM Subscribe
In positron emission tomography (PET) scans, why does there need to be an absorption calibration for each patient?
A PET scan works by injecting a tracer that is attached to a metabolically active molecule (like glucose) and is therefore subsequently absorbed by metabolically active sites in the body. This tracer emits positrons which very quickly encounter electrons. They annihilate each other and the result is the emission of two gamma ray protons. The scanner must detect the arrival of these two protons almost simultaneously for them to be counted.
If the energy of these protons is 511 keV, which should be powerful enough to pass through any tissue, why would there need to be an absorption calibration for each patient?
I am asking on behalf of my father, who is a brilliant electrical engineer and builds cardiac catheterization equipment for a living. He doesn't understand the physics of a PET machine and it's apparently driving him crazy. He seems to think that I might know since I have a Ph.D. in biology. I defer to those of you who may know more physics than I do.
A PET scan works by injecting a tracer that is attached to a metabolically active molecule (like glucose) and is therefore subsequently absorbed by metabolically active sites in the body. This tracer emits positrons which very quickly encounter electrons. They annihilate each other and the result is the emission of two gamma ray protons. The scanner must detect the arrival of these two protons almost simultaneously for them to be counted.
If the energy of these protons is 511 keV, which should be powerful enough to pass through any tissue, why would there need to be an absorption calibration for each patient?
I am asking on behalf of my father, who is a brilliant electrical engineer and builds cardiac catheterization equipment for a living. He doesn't understand the physics of a PET machine and it's apparently driving him crazy. He seems to think that I might know since I have a Ph.D. in biology. I defer to those of you who may know more physics than I do.
Nitpicking: They are photons, not protons. That is, particles of light like those you see with your eyes, only with much higher energy.
posted by springload at 1:15 PM on February 24, 2006
posted by springload at 1:15 PM on February 24, 2006
Response by poster: Good call. I was thinking photons and writing protons.
They would have to be photons because they are picked up by a scintillator.
posted by nekton at 1:18 PM on February 24, 2006
They would have to be photons because they are picked up by a scintillator.
posted by nekton at 1:18 PM on February 24, 2006
Just a shot in the dark from an ex-X-ray tech who quit before PETs became so prevalent and who got lost about halfway through zouhair's post.
Two guesses here. One, people's bodies absorb and expel things at different rates, so they need to make sure they can catch the isotope at the right moment to see what they're wanting to see. Two, maybe it's a question of burning out the areas of interest too? If I have a big tumor, I would think it would absorb more than a small one, so if it were calibrated for a small tumor, the big tumor would just wash everything out?
Now I feel really dumb. Incidentally, from zouhair's description, does that mean they're drawing blood again at the end of the scan? I've been having a lot of PET scans lately (hellooooo cancer!) and I haven't had that done yet. They just inject the isotope, give me the oral contrast to chug down, then bring me in a bit later to do the scan. I spend a lot of time on the table, so they could be doing various calibration things, but they don't draw any blood from me. I'm having my next one done at Yale, so maybe that one will be different.
posted by Moondoggie at 2:37 PM on February 24, 2006
Two guesses here. One, people's bodies absorb and expel things at different rates, so they need to make sure they can catch the isotope at the right moment to see what they're wanting to see. Two, maybe it's a question of burning out the areas of interest too? If I have a big tumor, I would think it would absorb more than a small one, so if it were calibrated for a small tumor, the big tumor would just wash everything out?
Now I feel really dumb. Incidentally, from zouhair's description, does that mean they're drawing blood again at the end of the scan? I've been having a lot of PET scans lately (hellooooo cancer!) and I haven't had that done yet. They just inject the isotope, give me the oral contrast to chug down, then bring me in a bit later to do the scan. I spend a lot of time on the table, so they could be doing various calibration things, but they don't draw any blood from me. I'm having my next one done at Yale, so maybe that one will be different.
posted by Moondoggie at 2:37 PM on February 24, 2006
maybe he's being misled by the terminology - or maybe people are simply using the wrong terminology - but absorption could refer to uptake of the tracer in the body/organ, rather than absorption of the photons.
the basic problem is converting from photon counts to whatever they are trying to measure (which is going to be something medically interesting like mass of some tissue). to do that they need to calibrate. just like you'd calibrate any other measurement that involves things too complex to predict accurately (like bodies).
moondoggie - i just read a site suggesting there were different methods for calibrating, and some people avoided blood samples because patients found it unpleasant.
posted by andrew cooke at 3:12 PM on February 24, 2006
the basic problem is converting from photon counts to whatever they are trying to measure (which is going to be something medically interesting like mass of some tissue). to do that they need to calibrate. just like you'd calibrate any other measurement that involves things too complex to predict accurately (like bodies).
moondoggie - i just read a site suggesting there were different methods for calibrating, and some people avoided blood samples because patients found it unpleasant.
posted by andrew cooke at 3:12 PM on February 24, 2006
just wanted to add that it could well be that the operator is the one misunderstanding. it's quite possible that it's called "absorption correction" and the operator has interpreted that as being a correction for photon absorption. in my experience people who operate complex machinery often don't understand the physics involved and so give answers (to questions from physicists/engineers) that are completely wrong, but which would assure most patients (which is what is important - rather than physical accuracy - as far as the operator is concerned). thankfully, that doesn't mean they're useless at their job, just that they work by following man-made rules, rather than those handed down by nature.
posted by andrew cooke at 3:16 PM on February 24, 2006
posted by andrew cooke at 3:16 PM on February 24, 2006
I've no experience with PET technology but I did a little work with biouptake of radio-isotopes, so I'm going to guess that the absorption correction has something to do with subtracting the background radioacivity present in the blood at any given time.
Based on zouhair's information it looks like this fudge factor corrects for the amount of isotope present in blood, which would add background noise to the specific signal emitted by the organ which has absorbed the labelled compound. This background has to be subtracted in order to calculate the specific activity (ie uptake) in the organ. Since the isotopes used in PET scanning are extremely short-lived, you need to correct for the time elapsed between sampling the blood and measuring it in the gamma counter (or whatever that Tennelac NaI instrument is).
posted by Quietgal at 4:01 PM on February 24, 2006
Based on zouhair's information it looks like this fudge factor corrects for the amount of isotope present in blood, which would add background noise to the specific signal emitted by the organ which has absorbed the labelled compound. This background has to be subtracted in order to calculate the specific activity (ie uptake) in the organ. Since the isotopes used in PET scanning are extremely short-lived, you need to correct for the time elapsed between sampling the blood and measuring it in the gamma counter (or whatever that Tennelac NaI instrument is).
posted by Quietgal at 4:01 PM on February 24, 2006
Hmm... my answer is likely not scientific, but conceptually makes sense to me.
Each person will have a basal metabolic rate in various regions of their brain. This basal activity will be different for various people (especially those with neural pathologies). It would make sense to calibrate the machine to each indivdual so that when their brain activity changes, these can get measured.
posted by ruwan at 4:01 PM on February 24, 2006
Each person will have a basal metabolic rate in various regions of their brain. This basal activity will be different for various people (especially those with neural pathologies). It would make sense to calibrate the machine to each indivdual so that when their brain activity changes, these can get measured.
posted by ruwan at 4:01 PM on February 24, 2006
The calibration before your "real" pet scan is to correct for attenuation of the gamma rays by the tissues. People have different amounts & dispersal of tissue mass throughout the body (and between each other), hence differing levels of attenuation. To generate a good image PET scans must "subtract scatter and randoms (accidental coincidences) background and to perform quantitative attenuation correction (to correct for the gamma rays absorbed in the body) and correct for camera deadtime at high countrates." (source)
This page sums things up pretty well, actually:
...an absorption correction matrix has to be acquired to correct for tissue absorption. This is done by acquiring a transmission PET scan prior to or after the injection of the radiopharmaceutical. Typically, a germanium radiation source is rotated around the patient and the attenuation of this radiation by the patient recorded and reconstructed into a transmission scan, which is then used to correct the data of the emission scan. Alternatively, the correction is made by assuming that the attenuation values for the imaged pixels is homogeneous, an assumption which is only reasonable for brain scans.
for an example, a before-and-after-attenuation pic
and another pair (the bottom two of these four)
images from here
posted by neda at 4:39 PM on February 24, 2006
This page sums things up pretty well, actually:
...an absorption correction matrix has to be acquired to correct for tissue absorption. This is done by acquiring a transmission PET scan prior to or after the injection of the radiopharmaceutical. Typically, a germanium radiation source is rotated around the patient and the attenuation of this radiation by the patient recorded and reconstructed into a transmission scan, which is then used to correct the data of the emission scan. Alternatively, the correction is made by assuming that the attenuation values for the imaged pixels is homogeneous, an assumption which is only reasonable for brain scans.
for an example, a before-and-after-attenuation pic
and another pair (the bottom two of these four)
images from here
posted by neda at 4:39 PM on February 24, 2006
This thread is closed to new comments.
...
Once the relevant lines have been inserted and any (e.g. EEG or psychological) tests completed, the subject is positioned centrally in the PET Camera port so that, for example, their brain is centered in the 256 mm diameter FOV for a brain study or their heart is well centered axially in the (256 mm axial field) camera for a cardiac study. The subject is restrained from movement during the scan as much as possible. Blood aliquots of 0.25 ml are sampled during uptake and measured in the Tennelec NaI(Tl) well counter in the Bioassay Laboratory as whole blood or, after centrifuge, as plasma. This provides the arterial activity concentration in nCi/ml via the relation:
Blood: nCi/ml = (cpm in well counter × calibration factor [nCi/cpm] e L × Dt ) / ( sample volume [ml] )
where the calibration factor is the well counter calibration in nCi/cpm for that isotope e L × Dt corrects for decay between sampling and counting (delay of Dt minutes).
...
posted by zouhair at 1:06 PM on February 24, 2006