Pixel correcting EPID images

Comments Off on Pixel correcting EPID images

Pixel-gain correcting without flood-field correcting

Peter Greer’s idea
“Use experimentally-measured EPID response vs phantom thickness mapping to work out a compensator that would give uniform EPID response.”

Clinical EPID images are usually automatically flood-field corrected: the signal in each pixel of each image is divided by the signal in the corresponding pixel in an image of a large, open field. This correction removes the effects of variations in both pixel sensitivity and beam intensity from the resulting images. Removal of the lateral variation of the beam intensity causes difficulties when the EPID images are processed for dosimetry because it results in the removal of the record of a real and possibly strong variation in delivered dose. However, if the flood field correction is avoided then variations in pixel sensitivity will have a strong and unwanted effect on resulting EPID images. There is a need, therefore, to be able to correct for pixel sensitivity variation (and, importantly in the case of Varian EPIDs, to be able to correct for scatter into the EPID from downstream hardware such as the support arm) without removing the beam profile from the images.

Obvious questions
1. Why not use the flood-field un-correcting method? Clinical use of the flood-field un-correcting method requires some initial measurements and code writing, followed by post-processing of every EPID image after it is acquired. The method described here requires some initial measurements, a bit of phantom design and fabrication and some further imaging, but after that all of your EPID images will be automatically pixel (gain) corrected, without you needing to do anything else.
2. Why go to all the trouble of designing a thin and wonky compensator when a big, thick, flat block of plastic does the same job? This is a good question, and you can go and use big, thick, flat block of plastic if you want to. But if you are interested in using your EPID images for dosimetry, then a compensator that is as thin as possible might be your best bet. (Big, thick, flat blocks of plastic do attenuate/scatter/harden the beam rather a lot.) And the thinner your compensator, the wonkier it will be.
3. Why are you doing this? Doesn’t your radiological opacity (thickness) calibration method render the flood-field correction irrelevant? I mean, haven’t you already solved this problem? (I really was asked this, once. *smugface*. Anyway…) People still want to be able to obtain pixel corrected EPID images for radiotherapy dosimetry, and people might also want to be able to obtain pixel corrected EPID images for other reasons. This method is (maybe) a little bit of help for them.

1. Obtain a series of EPID images with the beam attenuated (and hardened and scattered) by a series of isocentrically-positioned rectilinear phantoms, of various thicknesses.
2. Using the flood-field un-correcting method,  remove the flood field correction from the images.
3. Use these un-flood-field-corrected images to calculate a radiological opacity (thickness) calibration relationship, including maps of α(x,y) and β(x,y). (The method for doing this is reported in the literature and can be accessed from here.)
4. Use the radiological opacity calibration relationship to design a compensator by calculating the physical thickness profile of your compensator material (remember to correct for any differences in electron density between your compensator material and your calibration material) that will attenuate your beam from

by solving this:

5. Build the resulting compensator, which might look something like this:

6. Use the compensator (and an otherwise open field) to obtain a new “flood field” image. This image, in which the beam appears flat, can be used to flood-field correct all of your subsequent images. In your resulting images will have the shape of the beam will be preserved,  while your pixel gain is corrected out.

Anyway, that’s the idea.

Comments are closed.