Direct Digital and Computed Radiography

The technology that led to the development of detectors for CT found another application. If film could be replaced in conventional radiography, it would eliminate a number of problems. The same type of postprocessing procedures applied in CT imaging could be employed to eliminate repeat films. Once film was no longer used as the image receptor, wet processing would be a thing of the past. Imaging departments would no longer need to pay for maintenance of automatic processing units. It would no longer be necessary to recover silver from used fixer. In fact, with the processor gone, the space it occupied could certainly find a more profitable use.

Two approaches were employed to solve the problem. The first, more commonly referred to today as direct digital radiography or DR, was created employing an early, second-generation CT unit design. General Electric Medical Systems termed the unit Scan Projection Radiography (SPR), and used a high collimated fan beam x-ray source that moved in conjunction with a linear detector array moving across the patient (Bushong 2008). The principal disadvantage of SPR was the several seconds required for the "translate" motion across the patient to acquire the image. The long acquisition time frequently failed to eliminate patient motion. To eliminate the problem, another approach was developed. The fan beam was replaced by an area beam that would cover a 14 x 17 in. (35 x 43 cm) area, and the linear detector array was substituted by one of two technical approaches to image acquisition. The first, known as a charge-coupled device or CCD, was composed of either amorphous silicon or a selenium-based material and directly generated electrical impulses proportional to the x-rays incident on the detector. Many CCDs are electrically linked together to form a matrix. The second approach, promoted by Canon, employs "complementary metal oxide semiconductor" (CMOS) microprocessors that convert light to an electrical signal. A matrix of CMOS devices requires less power consumption than the CCD approach. Both CCD and CMOS systems are directly linked to a computer that processes the data and within several seconds of the exposure, the image appears on the monitor.

The other approach taken by Kodak, Konica, and Fuji is known as computed radiography or CR, and employs a photostimulable phosphor plate. However, instead of being directly connected to the computer that will process the data, the plate is placed into what appears to be a regular cassette. Following exposure to radiation, the cassette containing the plate is put into a "CR reader." In the reader, a laser scans the surface of the plate, releasing the x-ray energy captured by converting it to flashes of light. The flashes are converted to an electrical signal that is then processed by the computer within the reader to produce the image that appears on the monitor.

Of course, there is debate over which system is better. Each has its advantages, but the principal gain of the DR system is speed. Since it eliminated the middle step, the reader, it can produce an image in a couple of seconds where the CR system may requires a few minutes. If we are working with patients and concerned about throughput, time is a real consideration. When dealing with mummified or skeletal remains, the difference of a minute or so is not very significant. Therefore, the discussion of advantages and disadvantages will basically disregard the time factor. Instead the focus will be on portability, suitability in a wide range of applications, image quality, and cost.

Since the CR plate is loaded into a "plate holder" or cassette, it can be taken wherever there is an x-ray source. The CR reader can be located thousands of miles away. This was demonstrated during a study of the "Amazonian Princess," a fake mummy, at a Las Vegas, Nevada museum owned by the magician David Copperfield. A more complete description of the Princess is given in Chapter 9. The Fuji NDT research and development facility in Stamford, Connecticut, provided five plates for the study. To minimize risk of the plates being exposed during airline security screening procedures during transportation, they were shipped directly to the museum. Since the CR reader was back in Connecticut and we didn't know if the plates had been exposed to radiation during a preflight inspection, the plates were exposed to fluorescent light for 30 min to "clear" them. The study was conducted with both Fuji CR plates and Polaroid film. The latter was employed to ensure that acceptable

Figure 3.43A An AP projection with the Fuji CR system of a portion of the right side of the chest and arm of the Amazonian Princess. Note the cow ribs (A), the tacks (B) holding the arm together, the wires (C) forming the fingers, and the material (D) covering the arm and hand.

images would be obtained prior to leaving the museum. Following the study, the plates were shipped back to Stamford and placed into the reader. Even though approximately a week had elapsed between the time the plates were exposed and processed, there appeared to be no loss of image quality with the time delay (Figures 3.43A and 3.43B).

DR can also be portable. The Canon CMOS system can be coupled with a portable radiographic unit, in the United States with a MinXray® and in Great Britain with Xograph Healthcare, Ltd. The 17 x 17 in. (43.2 x 43.2 cm), the flat plate is connected to the computer by a cord that can be up to 21 ft (7 m) in length. Although we have not had the opportunity to test the system on mummified remains, we were able to use a Canon system on a cadaver at University College Dublin at Dublin, Ireland, in July 2008 (Figure 3.44). The system performed outstandingly. It was particularly advantageous to see the images within seconds of the exposure.

Which system, CR or DR, would be more suitable in a variety of situations? There were two limitations noted with the DR approach. Although most imaging studies will have a source-to-image receptor distance (SID) less than 21 ft (7 m), the maximum cord length could present a problem. In the study of plaster casts in the Slater Museum, discussed more thoroughly in Chapter 9, a 50 ft (15.25 m) SID was necessary. Although wireless DR systems are now entering the market, a long SID would require multiple exposures to produce sufficient radiation to expose the plate without overheating the unit. With any DR system, multiple exposures are not possible. As soon as an exposure is terminated, the computer begins processing the image.

Probably all major medical centers and most hospitals have moved or are in the process of moving away from film to a filmless approach with either DR or CR. Manufacturers

Figure 3.43B a composite created with five Polaroid images covering the same region as seen in the Fuji cR image. the ribs, tacks, and wires seen in the Fuji image were also clearly visible in the Polaroid image. However, since Fuji cR had a greater latitude than the Polaroid film, the material covering the arm and hand was seen clearly only on the former.

Figure 3.43B a composite created with five Polaroid images covering the same region as seen in the Fuji cR image. the ribs, tacks, and wires seen in the Fuji image were also clearly visible in the Polaroid image. However, since Fuji cR had a greater latitude than the Polaroid film, the material covering the arm and hand was seen clearly only on the former.

Figure 3.44 An image of a cadaver taken with a portable canon DR system. the oblong shapes (A) were created by the handholds in the board beneath the cadaver. Also seen on the image are a calcified femoral artery (B) and a fracture of the distal femur (c).

have applied the same philosophy in developing the equipment so little thought has to go into the imaging process. Units now have preprogrammed algorithms that are designed for living patients and render images with a balance of contrast and density. The optimal medical CR image has a sensitivity, or S value, of approximately 200. For an AP projection of George/Fred's chest at 70 kVp and 1.6 mAs, the S value was 243. Within a wide range of kVp (Figures 3.45A and 3.45B) and mAs (Figures 3.46A and 3.46B) values, the algorithm will automatically adjust the appearance of the image. Unlike film radiography, kV doesn't control contrast with a CR system. Remember the concept of a characteristic curve discussed in Chapter 2? It is a graphical representation of the intensity of radiation exposure and the resulting density on the film (see Chapter 2, Figure 2.3). The acceptable exposure was in the narrow region of the straight-line portion of the graph. The slope of that segment identified as the latitude or gradient determined the image contrast. The graphical representation for a CR system is quite different (Figure 3.47). There is no toe or shoulder. Also note that the slope, termed gradient, or G value, in Konica systems and latitude, or L value, in Fuji, is maintained across the entire graph. For the Konica system employed to produce the images in Figures 3.45 and 3.46, the algorithm maintained the G value at about 2.35.

Therefore, CR systems developed for the living are less than satisfactory for imaging the remains of the dead. However, the same technology has been modified for industrial applications. The industrial systems are less concerned with radiation dose to the object under examination than their medical counterparts. If small parts are the focus of the study, CR plates with smaller pixels requiring higher radiation doses are necessary.

Figure 3.45A The AP chest was taken at 100 kVp at 6.4 mAs. That is approximately 40 kVp higher than the optimal kV used with conventional radiographic or Polaroid film. The S value dropped to 17; however, the algorithm adjusted the appearance of the image to maintain a G value of 2.35.

Figure 3.45B In this example, the kV was reduced to 40, but the mAs remained constant at 6.4. The poor visualization of the spine and liver (a) suggested an underpenetrated image. The s value shot up to 1764, but the algorithm was able to maintain the G value of 2.35. However, the high s value indicated that insufficient radiation had reached the plate; this appearance is termed quantum mottle.

Figure 3.45B In this example, the kV was reduced to 40, but the mAs remained constant at 6.4. The poor visualization of the spine and liver (a) suggested an underpenetrated image. The s value shot up to 1764, but the algorithm was able to maintain the G value of 2.35. However, the high s value indicated that insufficient radiation had reached the plate; this appearance is termed quantum mottle.

Figure 3.46A For this image, 70 kV was used at 100 mAs. The mAs was 60 times the optimal value and resulted in an s value of 4. once again the algorithm was able to compensate and produced a G value of 2.39.
Figure 3.46B Once again the kV was set at 70, but the mAs was reduced to 0.8. Although the mAs was very low, the resulting S value was 489. Although there was a tremendous difference in mAs, because the G value was maintained at 2.35, the two images seem identical in appearance.

Similar to their medical counterparts, manufacturers have developed specific algorithms for each type of material that might be radiographed. Robert Lombardo (Lombardo 2008), an applications specialist from FUJIFILM NDT Systems, explained the principal differences between medical and industrial CR. Because of the tremendous

Figure 3.47 The graphical representation between the intensity of radiation exposure and the resulting density for film and CR system.

range of materials that might be examined in industry, algorithms have been designed for penetration settings from 50 to 350 kV. Since an x-ray tube is incapable of producing the high kV required for more dense materials, such as thick steel found in ships and submarines, a radioactive source, such as iridium or even cobalt, is necessary to provide the required penetration. In addition, to ensure that sufficient x-ray photons reach the CR plate, the exposure times may range from 20 s to 1 h or more when a radioactive source is employed to image very dense material. Because of the wide range of material densities, the Fuji industrial CR readers are calibrated for five times the radiation dose (10 mR) of their medical counterparts (2 mR). For the Fuji system, it was discovered that the rubber algorithm produced the greatest detail of mummified remains (Figures 3.48A and 3.48B).

As previously mentioned, the CR system has the capability of correcting for small errors in the technical factors that were set for a particular exposure. If too much radiation was used, an overexposure, the unit can more easily make corrections. However, if the plate was underexposed, the system would be unable to render an acceptable image. On a conventional radiograph, the underexposed film would appear "light," but on the underexposed CR image it has more of a "salt and pepper" or speckled appearance. This appearance is termed quantum mottle and is due to insufficient radiation reaching the plate (see Figure 3.45B).

Thus far, all the modalities discussed utilized radiation to produce images. In the mid-1980s, another modality was introduced that didn't employ an x-ray source, but rather used a high-intensity magnetic field. Initially termed nuclear magnetic resonance (NMR) the name was changed to magnetic resonance imaging (MRI) because of the bad connotations associated with the word "nuclear."

Figure 3.48A A lateral skull of George/Fred taken on a Konica medical CR system.
Figure 3.48B George/Fred's lateral skull taken on a Fuji industrial cR system and processed with a rubber algorithm.

Was this article helpful?

0 0

Post a comment