Modalities and methods

Computed tomograpy

 

The invention of computed tomography (CT) by Sir Godfrey Hounsfield in the early 1970s, was considered by many to represent the greatest step forward in radiology since the discovery of the X-rays. Together with Allen Cormack, Hounsfield was awarded the Nobel Prize in 1979 for his achievement. The first CT scanners were designed for head studies only, but soon whole body scanners became available as well. Today, CT scanning can be used for imaging of any part of the body.

The technological advances in both hardware and software have been tremendous and have greatly enhanced the applications and image quality of CT scanning since the introduction of the first CT scanner (EMI-Scanner) in 1972. Even though CT has met competition, first from ultrasonography, and then from magnetic resonance imaging, there are still many indications where CT is the imaging method of choice. The diagnostic role of CT will be further discussed in the clinical chapters.

Physical principles

On their way through tissues, X-rays are attenuated, partly due to absorption of energy, partly due to scattering. The attenuation may be expressed by the equation:

I = I₀e^{- md} (1)

where I is the intensity of the transmitted radiation (i.e., the radiation exiting from the tissue), I₀ is the intensity of the incident radiation (entering the tissue), is the so-called total linear attenuation coefficient of the tissue, and d is the travelled distance of the radiation through the tissue (tissue thickness). The attenuation coefficient, m is determined by the atomic number and electron density of the tissue; the higher the atomic number and electron density, the higher the attenuation coefficient. Atomic number and electron density are thus the two parameters determining the X-ray attenuating properties of a tissue. Note that the attenuation coefficient is also dependent upon the X-ray energy (see Fig. 9 in the Radiation physics chapter).

All imaging techniques and modalities using X-rays, are based upon the fact that different tissues provide different degrees of X-ray attenuation. The radiographic film used in full-size radiography has a very high spatial resolution, provided there are sufficiently large differences in attenuation between the structures being imaged. In this respect, full-size radiography is superior to all other radiological modalities. One of the major disadvantages of the film, is that its sensitivity to small differences in attenuation is low, i.e., its contrast resolution is poor. A radiographic film may roughly differentiate between only four different "substances" in the body: bone/calcification, soft tissue/fluid, fat, and gas (in decreasing order of attenuation). It is impossible to differentiate between different soft tissues, or between soft tissue and fluid. The ability of the radiographic film to show structural detail is further diminished by the projectional nature of the technique, resulting in considerable overlap of structures. Traditional tomography may improve the display of structural details, but even tomographic images contain (blurred) information from overlapping structures contributing to a reduction in contrast resolution.

With CT, only thin tissue slices are exposed to X -rays. There is no disturbing superimposition or blurring of structures located outside the selected tissue planes. The result is a contrast resolution far superior to projection X -ray techniques. The technical developments that have occurred in CT vary between manufacturers, and several generations of CT scanners have evolved. The generation number (first, second, third, fourth generation scanner, etc.) refers to the fundamental tube-detector structure of the scanner. Most scanners today have a basic tube-detector 

 

Figure 5. Schematic drawing of tube-detector system of 3rd generation CT scanner. The X-ray tube emits a sharply collimated fan beam of X-rays which passes the patient and reaches an array of detectors. Tube and detector array rotate together around the patient; one exposure often comprises 360o rotation.

system as illustrated in Fig. 5 (3rd generation). A thin, collimated, fan shaped beam of X-rays, perpendicular to the long axis of the body, is emitted from the tube. The fan may be wide enough to encompass the entire body diameter. By altering the collimation, the fan thickness may be varied from e.g. 1-10 mm. The thickness of the tissue slice exposed varies accordingly. The fan beam of X-rays transmitted through the patient is detected not by a film, but by an array of special detectors. The detector number may be approximately 700, and their location and distribution are adjusted to the fan beam width (Fig. 5). Detectors in current use are either solid-state crystals of various compositions (e.g. sodium iodine) or hollow chambers filled with pressurised xenon. In the detectors, X-ray photons generate electrical signals. The higher the intensity of the primary beam reaching the detector, the stronger the electrical signal. By recording the intensity of the transmitted radiation (I in Eq. 1), the attenuation of the primary beam may be calculated. CT detectors are approximately 100 times more sensitive than radiographic film in detecting differences in radiation intensity, and are therefore equally more sensitive in detecting differences in attenuation.

A CT examination starts with a digital projection radiograph (" scanogram ", "scout view") of the anatomical region to be slice imaged. The projection image is obtained by moving the examination table with the patient through the fan beam without rotating the tube or detectors (Fig. 6). The projection image is used for selection of slice locations, which are shown as superimposed lines ("slice level annotation").

 

	Figure 6. CT scanner. The X-ray tube and detectors are located inside the frame-work or gantry, surrounding the patient. The gantry may be angled around a horizontal axis to a maximum of approx. 20o. The examination starts with a projection image (scanogram): the patient table is fed through the gantry opening (aperture) during exposure without movement of the tube-detector system. A vertical tube-detector orientation (tube at 6 or 12 o' clock) yields frontal projections, and a horizontal tube detector orientation (tube at 3 or 9 o'clock) provides lateral projections. (Phototechnical Department, Rikshospitalet, Oslo.)
	Figure 7. The imaged slice of tissue divided into volume elements, voxels. The attenuation in each voxel determines the brightness (shade of grey) of the corresponding pixel in the final two-dimensional image.
	Figure 8. Scale of Hounsfield units (HU). The approximate scale locations of different substances are indicated. (By "tissue" is meant most fat-deficient soft tissues and parenchymal organs.) Reference points: -1,000 HU for air, 0 HU for water.

An exposed tissue slice may be considered divided into a number of equally large volume elements, so-called voxels (Fig. 7). To calculate the X-ray attenuation in each voxel, the attenuation recorded by each detector needs to be measured at several projections. This is done by simultaneous rotation of the X-ray tube and detector array in the slice plane during exposure (Fig. 5). In the two-dimensional image of the tissue slice (the CT scan), each voxel is represented by a picture element, or pixel, the size and location of the pixel being determined by the size and location of the voxel in the scan plane. The scan is presented on a monitor, where each pixel is given a shade of grey or brightness, according to the attenuation in the corresponding voxel. Pixels representing high attenuating voxels (e.g. bone) are bright, and pixels representing low attenuating voxels (e.g. adipose tissue) are relatively dark.

CT scanning allows measurement of tissue attenuation in a simple manner, and these measurements may have some diagnostic value. For example, fatty infiltration of the liver may be diagnosed by measuring abnormally low attenuation in the liver parenchyma. The attenuation is usually given a numerical value, an attenuation number or CT number. The numbers are set on an arbitrary linear scale which in modem CT scanners ranges from approximately -1,000 to +3,000 (Fig. 8). The unit for CT attenuation is named the Hounsfield unit (HU). The CT scanner is calibrated to give water an attenuation number equal to 0, and air an attenuation number equal to -1,000. Due to the computer technology used in CT scanning, the attenuation scale is based upon the binary system (see the Digital radiography chapter). The scale comprises 4,096 attenuation numbers (12 bits, 2¹²), and the exact values range from -1,024 to +3,071. (Older scanners will show only 2,048 numbers, ranging from -1,024 to +1,023. The numbers for water and air are the same, however.) Bone tissue has attenuation numbers ranging from approximately 800 HU in normal cortical bone to approximately 3,000 HU in the temporal bone pyramids. Most parenchymal tissues have values in the range of 40-80 HU, and pure fatty tissue has attenuation numbers in the order of -100 HU. Theoretically, these arbitrary numbers are directly proportional to the linear attenuation coefficients of the tissues; it should be noted, however, that the measurements of these numbers suffer from inaccuracies and inconsistencies caused by artefacts. For diagnostic purposes, attenuation numbers should therefore be used with caution.

Although CT scans have a contrast resolution far superior to traditional radiography, the spatial resolution is inferior. The spatial resolution is determined by the voxel size, i.e., by the pixel size and slice thickness.The smaller the voxels, the higher the spatial resolution. An ordinary number of voxels (and pixels) in a square field-of-view (FOV, the area being imaged) is 256 x 256 or 512 x 512 (the matrix size). If an area displayed in a 512 x 512 matrix measures 250 x 250 mm, each pixel will have a size of approximately 0.5 x 0.5 mm (250 mm : 512). The slice thickness is often 5-10 mm, but may be as thin as l mm. Thin slices are good for spatial resolution, but require a higher radiation dose to retain image quality (signal-to-noise ratio). Thin slices are also impractical when a large anatomical coverage is needed. The number of slices could become very high, further increasing the total radiation dose to the patient. The examination time will also increase with the number of slices. The choice of slice thickness is therefore a compromise between the demands for high spatial resolution, low radiation dose, and short examination time.

A recently introduced, entirely new scanning concept, named spiral CT, has dramatically increased the efficiency of CT scanning with respect to anatomic coverage per unit time. During exposure, there is a continuous linear movement of the table through the primary fan beam, and simultaneously a continuous rotation of the tube and detector array. The result is a spiral shaped trajectory of the fan beam through the patient. A large anatomical area may thus be covered during one breath-hold. By providing thin and contiguous "slices" (i.e., densely packed, thin windings in the spiral), spiral CT may yield very high quality three-dimensional reconstructions. Combined with intravenous bolus injections of contrast medium and subtraction techniques, CT angiograms may be reconstructed, providing projection images of the three-dimensional vascular tree.

Contrast media in CT scanning

Contrast in CT scanning is determined by differences in attenuation, but despite the superior sensitivity of CT in detecting attenuation differences, these differences are often too small for diagnostic purposes. In the vast majority of CT examinations, various contrast media are therefore added to enhance attenuation differences. The various contrast media are reviewed in the chapter Contrast media, and their diagnostic usefulness is discussed in the clinical chapters. The reader is therefore referred to these sections.

Cine CT

Even spiral CT scanners are too slow to give a detailed depiction of the cardiac structures. There is, however, a very special CT instrument having exposure times of 50 ms and a maximum exposure rate of 17 images per second. This speed is sufficiently high to "freeze" cardiac motions; sharply delineated images of the heart at different phases of contraction, may be acquired without the use of ECG trigging or gating. The scanner has been named cine CT, ultra fast CT, and millisecond CT. The X-ray source is a large electron gun with several massive parallel anode targets, oriented in semicircular rings around the patient. The intense electron beam is electronically steered along the tungsten anode rings. The X-ray beam thus created, sweeps through the patient in a fan shape, and is detected by a fixed array of detectors. There is thus no movement of "tube" or detectors.

Cine CT instruments are more expensive than conventional CT scanners, and they are so far not widely used. Competition from spiral CT and MR imaging will also probably contribute to a restricted use in the future. In addition to cardiac studies, cine CT has played a role in scanning of small children. The short exposure time "freezes" patient movements, and the use of sedation, which otherwise is necessary at most paediatric CT examinations, may be avoided.

 

Hans-Jørgen Smith