Diaglostic Radiology

DigRal Radiography Using aComputed Tomographic Instrument 1 Chandra S. Katragadda, M.D., Stewart R. Fogel, M.D., Gerald Cohen, Ph.D., Louis K. Wagner, Ph.D., Charles Morgan, III, M.D., Stanley F. Handel, M.D., Sharad R. Amtey, Ph.D., Ll.B., and Richard G. Lester, M.D.

A prototype computed radiography (CR) system was evaluated for its efficacy as an independent diagnostic modality. Preliminary measurements of high contrast resolution, low contrast perceptibility, and dose were obtained. Clinical examinations including skull, abdomen, liver, gallbladder, biliary system, spine, and extremities were performed as an adjunct to either computed tomography or CR. The data suggest that CR can be an effective diagnostic imaging modality by Itself. Advantages over conventional radiography include high scatter rejection, low patient dose, wide dynamic range, and good low contrast sensitivity for large objects; disadvantages, its long exposure time and relatively poor high contrast spatial resolution. INDEX TERMS:

Digital radiography. Radiography, instrumentation. Radiography, technique

Radiology 133:83-87, October 1979

t-E PURPOSE of this report is to present preliminary

T

findings on the efficacy of a computed radiographic (CR) system as a diagnostic tool. The system to be discussed (Scoutview, General Electric) represents scatter free, low dose digital imaging with wide dynamic range. The basic hardware of a rotating detector whole-body scanner (CT/T 7800, General Electric) is used, with the x-ray tube and detectors kept fixed, while the patient is translated at a constant speed through the gantry. The reduction in scattered radiation is the result of the relatively narrow (1.5 mm) x-ray beam employed, the air gap between the patient and the detectors, and the inherent detector collimation. This is an important feature because of the well-known contrast reducing effect of scatter, and is especially marked in examinations of thick body parts such as the abdomen where the ratio of scatter to primary radiations is about 9 to 1 (1, 2). Even the hig,est ratio grids obtain only about 60 % of the primary beam contrast (3). The reduction in scatter, combined with the high signalnoise detectors and a viewing system with variable window width controls, should greatly increase the perceptibility of low contrast details. The low dose to the patient is a result of greatly reduced scatter and relatively high quantum efficiency (about 50 % ) of the high pressure xenon gas detectors used. The digital imaging system allows for manipulation of the image without fll1her exposure to the patient and involves various mathematical algorithms available to the computer such as smoothing, edge enhancement, and subtraction. This is important because it allows the radiologist to obtain the maximum amount of information from the image. Studies have indicated that the eye-brain system is suboptimal for

Fig. 1. a. CR image, AP view, shows a depressed fracture of the posterior portion of the right frontal bone. b. CT scan obtained following CR confirms the presence of the fracture.

viewing noisy scenes (5-7); therefore, multiple displays of image content should be available to the viewer, with a tunable contrast scale and a degree of sharpness/ smoothing. This can be readily achieved through the use of a digitized image in a computer. The dynamic range of a system generally refers to the range of x-ray exposures at the detector to which the system can respond without saturation and produce satisfactory gray-scale images. Xenon detectors are reported to be linear over a range of radiation intensities of at least 10 4 (4). The relatively wide dynamic range of the present system should greatly reduce the need for repeat radiographs and allow for the simultaneous display of a wide range of object contrasts in the same image.

1 From the Department of Diagnostic Radiology, University of Texas Medical School, Houston, Texas. Received Jan. 18, 1979; accepted and revision requested Mar. 30; revision received May 25. as

83

84

CHANDRA

S.

KATRAGADDA AND OTHERS

Fig. 2. a. CR image obtained prior to abdominal CT shows an unexpected circular lucency in the right upper quadrant (arrows). b. Abdominal CT scan shows a circular mass with fluid in the subhepatic space, corresponding to the lucency on CR, which subsequently proved to be a subhepatic hematoma.

MATERIALS AND METHODS

The CR system currently in use at our institution is a prototype device. With the x-ray tube and xenon detector system fixed, the patient was translated through the gantry at a constant speed of 6 ern/sec, Simultaneously, a highly collimated pulsed fan beam of x rays was passed through the patient. The x-ray tube could be stationed at the top of the gantry to produce anteroposterior and posteroanterior views, rotated 90° for lateral views, or rotated 8° laterally from an anteroposterior position for stereo views. Conventional radiographs were obtained as indicated clinically. The patient was then positioned on the CT table and CR images in the anteroposterior and lateral projec-

October 1979

Fig. 3. a. CR image in a patient with carcinoma of the rectum shows lucent areas superimposed on the liver. b. CT scan of the liver obtained at the same time as (a) shows multiple necrotic metastases corresponding to the lucent areas.

tions were obtained, generally as an aid to localization prior to CT scanning, but as the sole procedure in selected cases. Images could be manipulated using different filter functions to obtain various combinations of smoothing and/or edge enhancement. All images were viewed as diagnostic radiographs and compared to conventional radiographs. Preliminary measurements of high contrast spatial resolution, low contrast perceptibility, and dose were obtained. High contrast resolution (nearly 100 %) was measured using a brass bar pattern ranging from 0.25 to 2 Ip/mm. Low contrast perceptibility was determined for the CR system and a representative medium speed filmscreen (Hi-Plus screens with Cronex 4 film). The contrast

DIGIT AL RADIOGRAPHY

Vol. 133

85

Diagnostic Radiology

Fig. 4. a. Conventional radiograph following transhepatic cholangiography shows multiple large and small stones in the common duct. b. CR image obtained at the same time as a shows large stones in the common duct. The small stones were poorly visualized.

detail phantom for the low contrast measurements consisted of various thicknesses of Plexiglas to obtain contrasts of 16% to 1.5%, with patterns of holes to estimate detail ranging from 19 to 0.3 mm in diameter. CR measurements were made at the clinically used source-object distance of 78 ern, using 1.5-mm beam spacing and maximum dose. Film-screen measurements were obtained at 120 kVp with 15- X 30-cm and 35- X 42-cm x-ray fields and an 8: 1 Bucky grid. Plexiglas scatter material 17-cm thick was used for both fields. Threshold perceptibility was determined as the smallest hole visible to a set of observers at a given contrast. RESULTS

High Contrast Spatial Resolution

The transverse and longitudinal spatial resolutions of the system differed and were dependent on several parameters. Transverse resolution was dependent on: (a) transverse width of the detectors; (b) source-to-object distance; (c) focal spot width; and (d) transverse distance from the central axis (due to an increase in the projected focal spot). Longitudinal resolution was dependent on: (a) table increment per pulse; (b) beam width; (c) sourceto-object distance; and (d) focal spot height. The high

Fig. 5. a. CR image of the spine, AP view, shows a fracture and displacement of the left pedicle of the L4 segment. b. CR image, lateral view, shows a compression fraction of the L4 vertebral body.

86

CHANDRA S. KATRAGADDA AND OTHERS

October 1979

scale (i.e., sensitometry). Preliminary results of low contrast perceptibility studies indicated the CR system to be superior to the film-screen system for the perception of low contrast objects larger than about 3 mm in diameter (less than 2 % subject contrast). The low contrast sensitivity of the CR system fell between the capabilities of CT and conventional radiography for large objects (greater than 3 mm in diameter) (8). Dose

The maximum entrance dose for the CR system at a 78-cm source-to-object distance was 38 rnrad (0.38 mGy), as compared to about 870 mrad (8.7 mGy) for an abdominal radiograph (9).

Clinical Observations

Fig. 6. a. CR image of the left knee after edge enhancement shows a well-delineated soft tissue mass along the medial aspect of the distal femur (arrows). (A CR image of the right knee is included for comparison.) b. CT scan shows the soft tissue mass in the left knee region. The mass is seen containing fluid and with smooth margins. It was proved to be a synovial cyst.

contrast resolution was about 1.25 mm in the transverse direction, primarily determined by the 1.3-mm width of the present detectors. The longitudinal resolution varied from about 1 mm for conditions of .75-mm beam width to about 1.4 mm for conditions of 1.5-mm beam spacing. The medium speed film-screen combination readily resolved the 0.25-mm bars.

Skull: AP and lateral CR images in 2 patients showed large, depressed skull fracnses clearly. CT performed after CR allowed evaluation of the brain and the degree of fracture depression. In these patients, CR images were comparable to routine skull radiographs. One of the cases is illustrated in Figure 1. Abdomen: CR can routinely produce diagnostic lateral views of the abdomen with the patient supine, while routine radiographic crosstable lateral abdominal views are often suboptimal. CR is frequently useful for evaluation of free intraperitoneal air. In 1 patient, CR showed a circular lucency in the subhepatic region, while CT showed a subhepatic fluid collection; this proved to be a subhepatic hematoma at surgery (Fig. 2), and was not seen on routine radiographs even in retrospect. Liver: A CR image obtained during infusion of 30 % Renografin demonstrated several areas of diminished density within the liver. CT showed these to be due to multiple metastases (Fig. 3). Gallbladder and biliary system: In another patient, transhepatic cholangiography was performed with routine radiographs and CR. Multiple common duct stones were seen with both modalities, but smaller stones were visible on the routine radiographs (Fig. 4). Spine: Several patients with spinal fracture dislocations were examined by CR. In each case, abnormality was demonstrated but fine detail was lacking (Fig. 5). Extremities: In 1 patient, CR, with its capability for edge enhancement through computer manipulation of the data, demonstrated the fat containing soft tissue planes around a large synovial cyst better than conventional radiographs (Fig. 6). CONCLUSION

Low Contrast Perceptibility

The ability to distinguish large objects of differing contrast is dependent on: (a) system noise (quantum plus nonquantum); (b) scatter radiation; and (c) system contrast

As an independent radiographic method, CR can be an effective diagnostic imaging tool. Physical factors which make the system attractive include high scatter rejection, low patient dose, wide dynamic range, and good low

Vol. 133

DIGITAL RADIOGRAPHY

contrast sensitivity for large objects. Current limitations include the long exposure time required to produce an image (about 7 sec.) and the relatively poor high contrast spatial resolution. ACKNOWLEDGWENTS: We are gateful to the General Electric Medical Systems Division for their support and interest in this work. We also wish to thank Frank Rocha, RT, for technical assistance. and Mrs. Marie McMurrey for secretarial assistance.

REFERENCES 1. Barnes GT. Cleare HM, Brezovich IA: Reduction of scatter in diagnostic radiology by means of a scanning multiple slit assembly. Radiology 120:691-694. Sep 1976 2. Dick CE, Soares CG, Motz JW: X-ray scatter data for diagnostic radiology. Phys Med Bioi 23: 1076-1085, Nov 1978 3. Rao GUV, Clark RL. Gayler BW: Radiographic magnification: a critical, theoretical and practical analysis. Part I. Appl RadioI3:37-40. Jan-Feb 1973

87

Diagnostic Radiology

4. Yaffe M, Fenster A, Johns HE: Xenon ionization detectors for fan beam computed tomography scanners. J Comput Assist Tomogr 1:419-428, Oct 1977 5. Motz JW, Danos M: Image information content and patient exposure. Med Phys 5:8-22, Jan-Feb 1978 6. Joseph PM: Image noise and smoothing in computed tomography (CT) scanners. Opt Engr 17:396-399, 1978 7. Wagner RF, Brown DG. Pastel MS: The application of information theory to the assessment of computed tomography. Med Phys 6:83-94, Mar-Apr 1979 8. Cohen G, Wagner LK, Amley SR, et al: Contrast-detail-dose evaluation of computed radiography: comparison with computed torng"aphy (CT) and conventional radiography. Proceedings: Application of Optical Instrumentation in Medicine VII, 173:41-47. Society of Photo-optical Instrumentation Engineers, Bellingham, Wash., 1979 9. Nationwide evaluation of x-ray trends: medical x-ray data. HEW Publication (FDA) 78-8056, U.S. Government Printing Office, 1978

Department of Diagnostic Radiology University of Texas Medical School at Houston 6431 Fannin St. Houston, Texas 77030

Digital radiography using a computed tomographic instrument.

Diaglostic Radiology DigRal Radiography Using aComputed Tomographic Instrument 1 Chandra S. Katragadda, M.D., Stewart R. Fogel, M.D., Gerald Cohen, P...
564KB Sizes 0 Downloads 0 Views