CR Erasure

Computed radiography (CR) is a form of digital radiography that relies on reusable cassettes for imaging. Essentially, X-ray photons strike a plate in the cassette which can be read afterwards. Once image information has been read and stored by a computer, the plate is erased and the cassette can be used again.

This lab was performed to test whether erasure of the CR cassette is sufficient to be reusable. Any trace of the previous image visible is referred to as “ghosting.” In the clinical environment, ghosting is an undesirable artifact that requires a new image to be taken, causing increased patient dose and throughput time.

The thoroughness of erasure was evaluated qualitatively and quantitatively. After first acquiring an image of pennies stacked on one another, the cassette was erased and a “blank” image was taken. By then selecting a narrow window width and adjusting the level, any ghosting artifact should have been visible on the image. The above video shows the window level slowly being raised, and no ghosting of pennies (i.e, a round radiopaque shape) is visible, as also shown in image 1 below.

Image 1:

Since the CR system used was Kodak, the quantitative evaluation of cassette erasure used the EI value generated after each exposure, as well as the mean pixel value of the image. Theoretically, if the plate is completely erased then the EI value after a blank image should be 0. The acceptable range of EI values is between 0 and 80, but the EI value obtained from this lab was 100.

These results seem to suggest that the erasure efficiency of the CR system only passed qualitatively, and consequently, corrective measures should be taken. These corrective measures would include ensuring proper calibration of the CR reader, the X-ray equipment, and the cassette itself. However I have reservations about the usefulness of a qualitative evaluation. Yes, a uniformly erased plate is important, and the EI value is a good indication of such erasure. However there are many potentially conflicting factors for this evaluation alone to be fair: variations in kVp, mAs, collimation, or SID could all change EI values. This drawback is further exacerbated if these factors do not fall within acceptable limits.

I think the visual inspection of the blank image is most important. If the technologist (or radiologist) cannot detect any ghosting then why should it matter if the EI reading is slightly outside the “accepted” limits?