Exposure Issues

The wide exposure latitude of digital radiography devices can result in a wide range of patient doses, from extremely low to extremely high. An “appropriate” patient dose is that required to provide a resultant image of “acceptable” image quality necessary to confidently make an accurate differential diagnosis. If the detector is underexposed due to inadequate radiographic technique factors, even though the image can be amplified and rescaled to present a good grayscale rendition, the quantum mottle in the image is likewise amplified, resulting in a noisy and grainy image. This causes the low contrast resolution sensitivity to be compromised, and often necessitates a retake. At least in this situation the underexposure is easy to recognize based upon the appearance of the image.

A more problematic situation occurs with detector overexposure caused by inappropriately high radiographic technique factors, resulting in needless patient dose. Except for extreme overexposures, images that are produced are usually of excellent radiographic quality with high contrast resolution sensitivity and low quantum mottle, due to the ability of the digital detector system to rescale the high signals to a grayscale range optimized for viewing on a soft copy monitor or hard copy film. Unfortunately, the patient in this situation has received needless radiation exposure, often without the knowledge of anyone involved in the acquisition or reading of the case. In some cases, a three to five times overexposure or more can happen, without any complaints from anyone. A phenomenon known as “dose creep” can occur based on the visible negative impact that underexposure can have on image appearance, and lack of perceived negative impact when the patient is overexposed but with beautiful electronic images. In the analog screen-film detector paradigm, the fixed speed of the detector requires that the exposure be correct, otherwise the response of the film optical density in the processed image is either too light (underexposure) or too dark (overexposure). Since there is no direct correlation with image appearance and grayscale rendition (brightness/contrast) in the digital image acquisition, the immediate feedback is lost. Fortunately, most digital detector systems have an “exposure indicator” that provides some feedback as to the relative exposure that was incident on the detector based upon the analysis of the raw image data intensity and subsequent scaling necessary to produce an image with appropriate brightness and contrast settings. Unfortunately, each manufacturer has a unique way of indicating this exposure indicator feedback signal. Until a formal exposure index standard is adopted by all manufacturers, it is imperative that technologists and radiologists become familiar with the specific way a given digital detector device indicates and reports the relative exposure intensity at which the image was acquired. This allows identification of under and overexposed examinations (and patients) and assists the technologist in performing adjustments in radiographic techniques to achieve consistency in radiation exposure and to optimize image quality simultaneously with safety to the patient.  

Figure 1 shows the comparison of the classic characteristic curve response of a variety of screen-film detector “speeds” as a function of incident exposure, and comparison to a generic digital radiography detector response. Clearly, the latitude of the digital detector spans a large range of “equivalent speed class” screen-film detectors. Of note is the extremely large range of very high exposures (red ellipsoid) that fall on the linear response curve of the digital detector, which is a cause for concern when digital feedback signals (exposure indices) are not tracked.

Figure 1. Characteristic curve response of screen film detectors of various radiographic speeds and digital radiography detectors.

The outcomes of wide latitude response of digital radiography devices are illustrated in Figure 2, demonstrating a set of images of a chest phantom at various exposure levels (an exposure level of 1 X is comparable to a 200 speed screen-film detector response). Screen-film image response in terms of optical density is strongly affected by the variation in incident exposure levels. Digital radiography images are scaled uniformly, despite the incident exposure variation; however, as the contrast resolution phantom depicts in the lower row, larger statistical variations in the underexposed images have a larger impact on the ability to resolve small, low contrast signals, whereas at very high exposures (compare 2.5X to 5X images) the image contrast resolution / sensitivity responses do not benefit significantly from increasing the dose to the patient. In fact, at even higher exposures, a loss of contrast resolution occurs from inclusion of other non-stochastic noise sources (e.g., detector imperfections) and saturation of the signals.

Figure 2. Digital radiography phantom images acquired with screen-film (top row), computed radiography (middle row), and an extracted and magnified insert from the digital images (bottom row). The variation in incident exposure in each column corresponds to a range from one-half up to five times the exposure of a typical “200 speed” screen-film detector.

A clinical example of underexposure is illustrated in Figure 3, demonstrating the lack of detail in the image and preponderance of a grainy, mottled appearance.  This underexposure is likely due to improper radiographic technique (mAs too low) or Automatic Exposure Control phototimer malfunction.  A repeat exposure of the same patient is shown in Figure 4, clearly demonstrating improved image quality and diagnostic information not shown in the underexposed image.

Figure 3. Underexposed computed radiography image of the abdomen (click on image for full sized version).

Same patient, proper exposure is shown in Figure 4.

Figure 4.Properly exposed computed radiography image of the abdomen. (click on image for full sized version).

In some cases, particularly in areas of the image with little or no attenuation, overexposure of the patient and the digital detector can result in saturation and a loss of image information beyond the linear operating range of the detector, as shown in Figure 5 for lung areas and un-collimated areas adjacent to the patient anatomy.

Figure 5. Overexposure and saturation of areas of the digital image, in which digital data is lost and unrecoverable (click on image for full sized version).
rsna