In some ways digital radiography is similar to film radiography. In other ways, it is very different. One way it differs is it’s response to exposure. When user changes the exposure of traditional film, the image gets dark or light and there is a loss of contrast. With digital radiography, the contrast is maintained and (for the most part) images don’t get dark or light. Rather, they get more or less noisy. With digital radiography, there is a trade off between exposure and noise.
Noise Defined: Noise is defined as a “random variation in image brightness.” Practically speaking, noise is the graininess, mottling, or textured appearance that is present in all medical images. Depending on the imaging detector there are many causes of noise. Although noise can give an image an objectionable appearance, the most important problem with noise is that noise can reduce the visibility of low contrast objects because the visibility threshold of low-contrast objects, is very noise dependent.
Exposure and noise with traditional film-screen radiography: With traditional film-screen radiography overexposing an image makes the images too dark with a loss of contrast and underexposing an image makes the image too light with a loss of contrast. Changing technique does nothing to alter the amount of noise or mottle in an image. The amount of noise is set by the film screen combination you are using. Since the individual film grains and phosphor crystals in the screens are very small we don’t generally appreciate grainy film images unless a very fast (sensitive) film screen system is used.
Exposure and noise in digital radiography: If a digital radiographic detector (CR, Flat Panel or CCD) is improperly exposed the image does not get too light or too dark. The wide dynamic range of digital radiography maintains contrast and image appearance over a wide range of exposures. Rather, if you underexpose a digital radiographic system, a principal change to the images is that the image will get noisy. The following is an image of an unacceptably noisy and underexposed DR image:
As you can see in the preceding image, the image does not get underexposed in the way that film gets underexposed. The primary change is that the image is more grainy. Just so you can see that the amount of noise will change with exposure, the following is an image taken with the exact same digital unit but the mAs was increased.
As you can see, there image is now properly exposed and diagnostic but there is still some noise in the image. This amount of noise is often seen with DR systems such as flat panel detectors and CCD units. CR images have some noise but it is usually not this noticeable.
The presence of noise in an image is a trade off between exposure and the amount of noise in the image. With most digital detectors, if the technician cranks the exposure up enough the noise go away***. This, however, is not a prudent practice because technicians will be unnecessarily exposing patients and themselves. Unfortunately, well meaning technicians don’t appreciate this point and they routinely overexpose patients. This practice has been termed “dose creep” in digital radiography. Exposure is one area that we are comparing at this years SHOWDOWN.
Users must understand that digital images look different than film images and purchasing a digital system simply because the images “look like film images” can be limiting limiting. Digital images look different than film and that takes some getting used to.
Digital images at proper exposures will generally have some degree of noise in the image. As a general rule, CR has less noisy images than DR systems. Within the DR category, the CCD systems currently on the veterinary market appear to be more noisy than flat panel detectors but there is a huge variability between vendors as other factors such as image processing, scintillator type, number of pixels, CCD camera technology, electronic noise, crystal size (CR), CR phosphor packing density etc. impact the final image.
Veterinarians researching digital systems are often unaware of image noise and react (often violently) to the presence of noise in an image. Some vendors smooth out their images so they don’t appear noisy. This helps make sales as veterinarians (in my experience) who are new to digital radiography invariably will select a blurry image over a sharp grainy images. Unfortunately, this type of processing not only blurs the noise, but it blurs the image as well. In our teleradiology practice , we consider blurry images much more objectionable than slightly noisy images.
Ways to decrease noise and prevent dose creep: Technicians quickly learn that they get less complaints about their images if the images are not noisy. As stated previously, increasing technique just to decrease noise is not safe for them or for the patients. Technicians should be advised to try to prevent “dose creep.”
- A technique chart is recommended. This technique chart can be programmed into the anatomic programming of newer x-ray machines. There is only one digital system that I know of that does not need a technique chart. All of the other digital systems absolutely need a technique chart. Do not accept a "generic" technique chart from your vendor. A technique chart must be developed for your system on the day of the install.
- Consideration should be given to purchasing a digital radiography system that provides technicians with an EXPOSURE CONTROL INDEX. The exposure control index tells the technicians if the image is properly exposed. Some digital radiography systems on the veterinary market now offer an exposure control index. This is one area that we will be evaluating in this years SHOWDOWN.
In conclusion, image noise is a factor in digital radiography and the amount of noise present in an image should be a consideration when purchasing a system. Some systems have more noise than others. Eliminating noise in digital radiographic images is possible (for many systems) but (for many systems) would require an unnecessary increase in dose to the patient and personnel.
***Unfortunately, there are some detectors on the veterinary market that are so insensitive to radiation that the images are objectionably noisy regardless of technique. In my experience, these detectors are usually low end or older model CCD units.
