Computed tomography: traditional, spiral tomography

Computed tomography is a special type of X-ray examination that is performed by indirectly measuring the attenuation, or weakening, of X-rays from various positions defined around the patient being examined. Essentially, all we know is:

• what leaves the x-ray tube,
• that reaches the detector and
• what is the location of the x-ray tube and detector in each position.

Everything else follows from this information. Most CT sections are oriented vertically relative to the body axis. They are usually called axial or transverse sections. For each section, the X-ray tube rotates around the patient, the thickness of the section is selected in advance. Most CT scanners operate on the principle of constant rotation with a fan-shaped divergence of the beams. In this case, the X-ray tube and detector are rigidly coupled, and their rotational movements around the scanned area occur simultaneously with the emission and capture of X-rays. Thus, X-rays, passing through the patient, reach the detectors located on the opposite side. Fan-shaped divergence occurs in the range from 40 ° to 60 °, depending on the device design, and is determined by the angle starting from the focal spot of the X-ray tube and expanding in the form of a sector to the outer boundaries of the row of detectors. Usually, an image is formed with each rotation of 360 °, the data obtained is sufficient for this. During scanning, attenuation coefficients are measured at many points, forming an attenuation profile. In fact, attenuation profiles are nothing more than a set of signals received from all detector channels from a given angle of the tube-detector system. Modern CT scanners are capable of transmitting and collecting data from approximately 1400 positions of the detector-tube system over a 360° circle, or about 4 positions per degree. Each attenuation profile includes measurements from 1500 detector channels, i.e. approximately 30 channels per degree, assuming a beam divergence angle of 50°. At the beginning of the examination, as the patient table moves at a constant speed into the gantry, a digital radiograph is obtained (a "scanogram" or "topogram"), on which the required sections can be planned later. For CT examination of the spine or head, the gantry is rotated at the desired angle, thereby achieving optimal orientation of the sections).

Computed tomography uses complex readings from an x-ray sensor that rotates around the patient to produce a large number of different depth-specific images (tomograms), which are digitized and converted into cross-sectional images. CT provides 2- and 3-dimensional information that is not possible with plain x-rays and at much higher contrast resolution. As a result, CT has become the new standard for imaging most intracranial, head and neck, intrathoracic, and intra-abdominal structures.

Early CT scanners used only one x-ray sensor, and the patient moved through the scanner incrementally, stopping for each image. This method has been largely replaced by helical CT: the patient moves continuously through the scanner, which rotates and takes images continuously. Helical CT greatly reduces imaging time and reduces plate thickness. The use of scanners with multiple sensors (4-64 rows of x-ray sensors) further reduces imaging time and allows plate thicknesses of less than 1 mm.

With so much data displayed, images can be reconstructed from almost any angle (as is done in MRI) and can be used to construct 3-dimensional images while maintaining a diagnostic imaging solution. Clinical applications include CT angiography (eg, to evaluate pulmonary embolism) and cardiac imaging (eg, coronary angiography, evaluating coronary artery hardening). Electron beam CT, another type of fast CT, can also be used to evaluate coronary artery hardening.

CT scans can be obtained with or without contrast. Non-contrast CT can detect acute hemorrhage (which appears bright white) and characterize bone fractures. Contrast CT uses IV or oral contrast, or both. IV contrast, similar to that used in plain X-rays, is used to image tumors, infection, inflammation, and soft tissue injury and to evaluate the vascular system, as in cases of suspected pulmonary embolism, aortic aneurysm, or aortic dissection. Renal excretion of contrast allows evaluation of the genitourinary system. For information on contrast reactions and their interpretation, see:

Oral contrast is used to image the abdominal area; this helps to separate the intestinal structure from the surrounding structure. The standard oral contrast, barium iodine, can be used when intestinal perforation is suspected (eg, due to trauma); low osmolar contrast should be used when the risk of aspiration is high.

Radiation exposure is an important issue when using CT. The radiation dose from a routine abdominal CT scan is 200 to 300 times higher than the radiation dose received from a typical chest x-ray. CT is now the most common source of artificial radiation for most of the population and accounts for more than two-thirds of total medical radiation exposure. This degree of human exposure is not trivial; the lifetime risk of radiation exposure for children exposed to CT radiation today is estimated to be much higher than that of adults. Therefore, the need for CT examination must be carefully weighed against the potential risk for each individual patient.

[ 1 ], [ 2 ], [ 3 ], [ 4 ]

Multislice computed tomography

Multi-detector spiral computed tomography (multislice computed tomography)

Multi-row detector CT scanners are the latest generation of scanners. Opposite the X-ray tube, there is not one, but several rows of detectors. This allows for a significant reduction in examination time and improved contrast resolution, which allows, for example, for clearer visualization of contrasted blood vessels. The rows of Z-axis detectors opposite the X-ray tube are of different widths: the outer row is wider than the inner one. This provides better conditions for image reconstruction after data collection.

[ 5 ], [ 6 ], [ 7 ]

Comparison of traditional and spiral computed tomography

Conventional CT scans acquire a series of sequential, equally spaced images through a specific body part, such as the abdomen or head. A short pause after each slice is required to advance the table with the patient to the next predetermined position. Thickness and overlap/interslice spacing are predetermined. The raw data for each level are stored separately. A short pause between slices allows the conscious patient to take a breath, thereby avoiding gross respiratory artifacts in the image. However, the examination may take several minutes, depending on the scan area and the patient size. It is important to time the image acquisition after IV CS, which is especially important for assessing perfusion effects. CT is the method of choice for obtaining a complete 2D axial image of the body without the interference of bone and/or air as seen on conventional radiographs.

In spiral computed tomography with single-row and multi-row detector arrangement (MSCT), the acquisition of patient examination data occurs continuously during the table advancement into the gantry. The X-ray tube describes a helical trajectory around the patient. The table advancement is coordinated with the time required for the tube to rotate 360° (spiral pitch) - data acquisition continues continuously in full. Such a modern technique significantly improves tomography, because breathing artifacts and noise do not affect the single data set as significantly as in traditional computed tomography. A single raw data base is used to reconstruct slices of different thicknesses and different intervals. Partial overlapping of sections improves reconstruction capabilities.

Data collection for a full abdominal scan takes 1 to 2 minutes: 2 or 3 spirals, each lasting 10 to 20 seconds. The time limit is due to the patient's ability to hold their breath and the need to cool the X-ray tube. Some additional time is needed to reconstruct the image. When assessing renal function, a short pause is required after the administration of the contrast agent to allow for the excretion of the contrast agent.

Another important advantage of the spiral method is the ability to detect pathological formations smaller than the slice thickness. Small liver metastases may be missed if they do not fall into the slice due to the patient's uneven breathing depth during scanning. Metastases are easily detected from the raw data of the spiral method when reconstructing slices obtained with overlapping sections.

[ 8 ]

Spatial resolution

Image reconstruction is based on differences in the contrast of individual structures. On this basis, an image matrix of the visualization area of 512 x 512 or more image elements (pixels) is created. Pixels appear on the monitor screen as areas of different shades of gray depending on their attenuation coefficient. In fact, these are not even squares, but cubes (voxels = volumetric elements) that have a length along the body axis, corresponding to the thickness of the slice.

Image quality improves with smaller voxels, but this only applies to spatial resolution; further thinning of the slice reduces the signal-to-noise ratio. Another disadvantage of thin slices is the increased radiation dose to the patient. However, small voxels with equal dimensions in all three dimensions (isotropic voxel) offer significant advantages: multiplanar reconstruction (MPR) in coronal, sagittal or other projections is presented on the image without a step contour). Using voxels of unequal dimensions (anisotropic voxels) for MPR leads to the appearance of jaggedness in the reconstructed image. For example, it may be difficult to exclude a fracture.

[ 9 ], [ 10 ]

Spiral step

The pitch of the spiral characterizes the degree of table movement in mm per rotation and the thickness of the cut. Slow table movement forms a compressed spiral. Acceleration of table movement without changing the thickness of the cut or the rotation speed creates space between the cuts on the resulting spiral.

Most often, the spiral pitch is understood as the ratio of the movement (feed) of the table during gantry rotation, expressed in mm, to the collimation, also expressed in mm.

Since the dimensions (mm) in the numerator and denominator are balanced, the helix pitch is a dimensionless quantity. For MSCT, the so-called volumetric helix pitch is usually taken to be the ratio of the table feed to a single slice, rather than to the total number of slices along the Z axis. For the example used above, the volumetric helix pitch is 16 (24 mm / 1.5 mm). However, there is a tendency to return to the first definition of the helix pitch.

New scanners offer the option of selecting a craniocaudal (Z-axis) extension of the study area on the topogram. Also, the tube rotation time, slice collimation (thin or thick slice) and study time (breath-hold interval) are adjusted as needed. Software such as SureView calculates the appropriate spiral pitch, usually setting the value between 0.5 and 2.0.

[ 11 ], [ 12 ]

Slice Collimation: Resolution along the Z axis

The image resolution (along the Z-axis or the patient's body axis) can also be adapted to the specific diagnostic task using collimation. Slices of 5 to 8 mm thickness are fully consistent with standard abdominal examination. However, the precise localization of small bone fracture fragments or the assessment of subtle pulmonary changes require the use of thin slices (0.5 to 2 mm). What determines the slice thickness?

The term collimation is defined as obtaining a thin or thick slice along the longitudinal axis of the patient's body (Z axis). The physician can limit the fan-shaped divergence of the radiation beam from the X-ray tube with a collimator. The size of the collimator's opening regulates the passage of rays that hit the detectors behind the patient in a wide or narrow stream. Narrowing the radiation beam improves spatial resolution along the patient's Z axis. The collimator can be located not only immediately at the exit of the tube, but also directly in front of the detectors, i.e. "behind" the patient when viewed from the side of the X-ray source.

A collimator aperture-dependent system with one row of detectors behind the patient (single slice) can produce slices of 10 mm, 8 mm, 5 mm or even 1 mm. CT scanning with very thin sections is called "high-resolution CT" (HRCT). If the slice thickness is less than a millimeter, it is called "ultra-high-resolution CT" (UHRCT). UHRCT, used for examining the petrous bone with slices of about 0.5 mm, reveals fine fracture lines passing through the skull base or the auditory ossicles in the tympanic cavity). For the liver, high-contrast resolution is used to detect metastases, requiring slices of somewhat greater thickness.

[ 13 ], [ 14 ], [ 15 ]

Detector placement schemes

Further development of single-slice spiral technology led to the introduction of multi-slice (multi-spiral) techniques, which use not one but several rows of detectors located perpendicular to the Z axis opposite the X-ray source. This makes it possible to simultaneously collect data from several sections.

Due to the fan-shaped divergence of the radiation, the detector rows must have different widths. The detector arrangement scheme is such that the width of the detectors increases from the center to the edge, which allows for varying combinations of thickness and number of slices obtained.

For example, a 16-slice study can be performed with 16 thin high-resolution slices (for Siemens Sensation 16 this is the 16 x 0.75 mm technique) or with 16 sections of twice the thickness. For iliofemoral CT angiography, it is preferable to obtain a volume slice in one cycle along the Z-axis. In this case, the collimation width is 16 x 1.5 mm.

The development of CT scanners did not end with 16 slices. Data collection can be accelerated by using scanners with 32 and 64 rows of detectors. However, the trend toward thinner slices leads to higher radiation doses for the patient, which requires additional and already feasible measures to reduce radiation exposure.

When examining the liver and pancreas, many specialists prefer to reduce the slice thickness from 10 to 3 mm to improve image sharpness. However, this increases the noise level by approximately 80%. Therefore, in order to maintain image quality, it is necessary to either additionally increase the current strength on the tube, i.e. increase the current strength (mA) by 80%, or increase the scanning time (the mAs product increases).

[ 16 ], [ 17 ]

Image reconstruction algorithm

Spiral CT has an additional advantage: during the image reconstruction process, most of the data is not actually measured in a particular slice. Instead, measurements outside of that slice are interpolated with most of the values near the slice and become slice-specific data. In other words: the results of the data processing near the slice are more important for reconstructing the image of a particular section.

An interesting phenomenon follows from this. The patient dose (in mGy) is defined as mAs per rotation divided by the helix pitch, and the dose per image is equal to mAs per rotation without taking into account the helix pitch. If, for example, the settings are 150 mAs per rotation with a helix pitch of 1.5, then the patient dose is 100 mAs, and the dose per image is 150 mAs. Therefore, the use of helical technology can improve the contrast resolution by choosing a high mAs value. This makes it possible to increase the image contrast, tissue resolution (image clarity) by decreasing the slice thickness and to select a pitch and helix interval length such that the patient dose is reduced! Thus, a large number of slices can be obtained without increasing the dose or the load on the x-ray tube.

This technology is especially important when converting the obtained data into 2-dimensional (sagittal, curvilinear, coronal) or 3-dimensional reconstructions.

The measurement data from the detectors are passed, profile by profile, to the detector electronics as electrical signals corresponding to the actual attenuation of the X-rays. The electrical signals are digitized and then sent to the video processor. At this stage of the image reconstruction, a "pipeline" method is used, consisting of preprocessing, filtering and reverse engineering.

Preprocessing includes all corrections made to prepare the acquired data for image reconstruction. For example, dark current correction, output signal correction, calibration, track correction, radiation hardening, etc. These corrections are made to reduce variations in the operation of the tube and detectors.

Filtering uses negative values to correct for the image blurring inherent in reverse engineering. If, for example, a cylindrical water phantom is scanned and reconstructed without filtering, its edges will be extremely blurry. What happens when eight attenuation profiles are superimposed to reconstruct the image? Since some portion of the cylinder is measured by two superimposed profiles, a star-shaped image is obtained instead of a real cylinder. By introducing negative values beyond the positive component of the attenuation profiles, the edges of this cylinder become sharp.

Reverse engineering redistributes the convolved scan data into a 2-dimensional image matrix, displaying the corrupted slices. This is done profile by profile until the image reconstruction process is complete. The image matrix can be thought of as a checkerboard, but made up of 512 x 512 or 1024 x 1024 elements, commonly called “pixels.” Reverse engineering results in each pixel having an exact density, which on the monitor screen appears as different shades of gray, from light to dark. The lighter the area of the screen, the higher the density of the tissue within the pixel (e.g., bone structures).

[ 18 ], [ 19 ]

Effect of voltage (kV)

When the anatomical region being examined has a high absorption capacity (e.g. CT of the head, shoulder girdle, thoracic or lumbar spine, pelvis or simply an obese patient), it is advisable to use higher voltage or, alternatively, higher mA values. By selecting a high voltage on the x-ray tube, you increase the hardness of the x-ray radiation. Accordingly, the x-rays penetrate the anatomical region with a high absorption capacity much more easily. The positive side of this process is that the low-energy components of the radiation that are absorbed by the patient's tissues are reduced without affecting the image acquisition. For examination of children and when tracking the KB bolus, it may be advisable to use a lower voltage than in standard settings.

[ 20 ], [ 21 ], [ 22 ], [ 23 ], [ 24 ], [ 25 ]

Tube current (mAs)

The current, measured in milliampere seconds (mAs), also affects the radiation dose received by the patient. A large patient requires a higher current in the tube to obtain a good image. Thus, a more obese patient receives a higher radiation dose than, for example, a child with a significantly smaller body size.

Areas with bone structures that absorb and scatter radiation more, such as the shoulder girdle and pelvis, require a higher tube current than, for example, the neck, the abdomen of a thin person, or the legs. This dependence is actively used in radiation protection.

Scan time

The shortest possible scan time should be selected, especially in the abdomen and chest, where cardiac contractions and intestinal peristalsis may degrade image quality. CT imaging quality is also improved by reducing the likelihood of involuntary patient movements. On the other hand, longer scan times may be necessary to collect sufficient data and maximize spatial resolution. Sometimes the choice of extended scan times with reduced current is used deliberately to extend the life of the x-ray tube.

[ 26 ], [ 27 ], [ 28 ], [ 29 ], [ 30 ]

3D reconstruction

Because spiral tomography collects data for an entire region of the patient's body, visualization of fractures and blood vessels has improved significantly. Several different 3D reconstruction techniques are used:

[ 31 ], [ 32 ], [ 33 ], [ 34 ], [ 35 ]

Maximum Intensity Projection (MIP)

MIP is a mathematical method by which hyperintense voxels are extracted from a 2D or 3D data set. Voxels are selected from a data set acquired at different angles and then projected as 2D images. The 3D effect is obtained by changing the projection angle in small steps and then visualizing the reconstructed image in rapid succession (i.e., in a dynamic view mode). This method is often used in contrast-enhanced blood vessel imaging.

[ 36 ], [ 37 ], [ 38 ], [ 39 ], [ 40 ]

Multiplanar Reconstruction (MPR)

This technique makes it possible to reconstruct images in any projection, be it coronal, sagittal or curvilinear. MPR is a valuable tool in fracture diagnostics and orthopedics. For example, traditional axial slices do not always provide complete information about fractures. A very thin fracture without displacement of fragments and disruption of the cortical plate can be detected more effectively using MPR.

[ 41 ], [ 42 ]

Surface Shaded Display, SSD

This method reconstructs the organ or bone surface defined above a given threshold in Hounsfield units. The choice of the imaging angle, as well as the location of the hypothetical light source, is key to obtaining an optimal reconstruction (the computer calculates and removes shadow areas from the image). The bone surface clearly shows the fracture of the distal radius demonstrated by MPR.

3D SSD is also used in surgical planning, as in the case of a traumatic spinal fracture. By changing the angle of the image, it is easy to detect a compression fracture of the thoracic spine and assess the condition of the intervertebral foramina. The latter can be examined in several different projections. The sagittal MPR shows a bone fragment that is displaced into the spinal canal.

Basic rules for reading CT scans

• Anatomical orientation

The image on the monitor is not just a 2-dimensional representation of the anatomical structures, but contains data on the average tissue absorption of X-rays, represented by a matrix of 512 x 512 elements (pixels). The slice has a certain thickness (d _S ) and is the sum of cuboid elements (voxels) of the same size, combined into a matrix. This technical feature is the basis of the partial volume effect, explained below. The images obtained are usually viewed from below (from the caudal side). Therefore, the right side of the patient is on the left in the image and vice versa. For example, the liver, located in the right half of the abdominal cavity, is represented on the left side of the image. And organs located on the left, such as the stomach and spleen, are visible in the image on the right. The anterior surface of the body, in this case represented by the anterior abdominal wall, is defined in the upper part of the image, and the posterior surface with the spine is at the bottom. The same principle of image formation is used in conventional radiography.

• Partial Volume Effects

The radiologist determines the slice thickness (d _S ). For examination of the thoracic and abdominal cavities, 8-10 mm is usually selected, and for the skull, spine, orbits and pyramids of the temporal bones - 2-5 mm. Therefore, structures can occupy the entire slice thickness or only part of it. The intensity of voxel coloring on the gray scale depends on the average attenuation coefficient for all its components. If the structure has the same shape throughout the slice thickness, it will appear clearly outlined, as in the case of the abdominal aorta and inferior vena cava.

The partial volume effect occurs when the structure does not occupy the entire thickness of the slice. For example, if the slice includes only part of the vertebral body and part of the disk, their contours are unclear. The same is observed when the organ narrows inside the slice. This is the reason for the poor clarity of the kidney poles, the contours of the gallbladder and urinary bladder.

• Difference between nodular and tubular structures

It is important to be able to distinguish enlarged and pathologically altered lymph nodes from vessels and muscles included in the cross-section. It can be very difficult to do this from just one section, because these structures have the same density (and the same shade of gray). Therefore, it is always necessary to analyze adjacent sections located more cranially and caudally. By specifying in how many sections a given structure is visible, it is possible to resolve the dilemma of whether we are seeing an enlarged node or a more or less long tubular structure: the lymph node will be determined only in one or two sections and will not be visualized in adjacent ones. The aorta, inferior vena cava, and muscles, such as the iliac-lumbar, are visible throughout the craniocaudal series of images.

If there is a suspicion of an enlarged nodular formation on one section, the doctor should immediately compare adjacent sections to clearly determine whether this "formation" is simply a vessel or muscle in cross section. This tactic is also good because it allows for the rapid establishment of the effect of a private volume.

• Densitometry (measurement of tissue density)

If it is not known, for example, whether the fluid found in the pleural cavity is effusion or blood, measuring its density facilitates differential diagnosis. Similarly, densitometry can be used for focal lesions in the liver or kidney parenchyma. However, it is not recommended to draw a conclusion based on the assessment of a single voxel, since such measurements are not very reliable. For greater reliability, it is necessary to expand the "region of interest" consisting of several voxels in a focal lesion, any structure or volume of fluid. The computer calculates the average density and the standard deviation.

Particular care should be taken not to miss hardening artifacts or partial volume effects. If a lesion does not extend across the entire slice thickness, the density measurement includes adjacent structures. The density of a lesion will only be measured correctly if it fills the entire slice thickness (d _S ). In this case, it is more likely that the measurement will involve the lesion itself rather than adjacent structures. If d S is larger than the diameter of the lesion, such as a small lesion, this will result in a partial volume effect at any scan level.

• Density levels of different types of fabrics

Modern devices are capable of covering 4096 shades of gray scale, which represent different levels of density in Hounsfield units (HU). The density of water was arbitrarily taken as 0 HU, and air as - 1000 HU. A monitor screen can display a maximum of 256 shades of gray. However, the human eye can distinguish only about 20. Since the spectrum of human tissue densities extends wider than these rather narrow limits, it is possible to select and adjust the image window so that only tissues of the desired density range are visible.

The average window density level should be set as close as possible to the density level of the tissues being examined. The lung, due to its increased airiness, is best examined in a window with low HU settings, while for bone tissue the window level should be significantly increased. The image contrast depends on the window width: a narrowed window is more contrasty, since 20 shades of gray cover only a small part of the density scale.

It is important to note that the density level of almost all parenchymatous organs lies within the narrow limits between 10 and 90 HU. The lungs are an exception, so as mentioned above, special window parameters must be set. With regard to hemorrhages, it must be taken into account that the density level of recently clotted blood is approximately 30 HU higher than that of fresh blood. The density then drops again in areas of old hemorrhage and in areas of thrombus lysis. Exudate with a protein content of more than 30 g/L is not easily distinguished from transudate (with a protein content below 30 g/L) with standard window settings. In addition, it must be said that the high degree of density overlap, for example in lymph nodes, spleen, muscle and pancreas, makes it impossible to establish the tissue identity based on density assessment alone.

In conclusion, it should be noted that normal tissue density values also vary among individuals and change under the influence of contrast agents in the circulating blood and in the organ. The latter aspect is of particular importance for the study of the genitourinary system and concerns the intravenous administration of contrast agents. In this case, the contrast agent quickly begins to be excreted by the kidneys, which leads to an increase in the density of the renal parenchyma during scanning. This effect can be used to assess renal function.

• Documenting research in different windows

Once the image is obtained, it is necessary to transfer the image to film (make a hard copy) to document the examination. For example, when assessing the condition of the mediastinum and soft tissues of the chest, a window is set so that the muscles and fat tissue are clearly visualized in shades of gray. In this case, a soft-tissue window with a center of 50 HU and a width of 350 HU is used. As a result, tissues with a density from -125 HU (50-350/2) to +225 HU (50+350/2) are represented in gray. All tissues with a density lower than -125 HU, such as the lung, appear black. Tissues with a density higher than +225 HU are white, and their internal structure is not differentiated.

If it is necessary to examine the lung parenchyma, for example, when nodular formations are excluded, the window center should be reduced to -200 HU, and the width increased (2000 HU). When using this window (pulmonary window), low-density lung structures are better differentiated.

To achieve maximum contrast between the gray and white matter of the brain, a special brain window should be selected. Since the densities of gray and white matter differ only slightly, the soft tissue window should be very narrow (80 - 100 HU) and high-contrast, and its center should be in the middle of the brain tissue density values (35 HU). With such settings, it is impossible to examine the skull bones, since all structures denser than 75 - 85 HU appear white. Therefore, the center and width of the bone window should be significantly higher - about + 300 HU and 1500 HU, respectively. Metastases in the occipital bone are visualized only when using a bone window, but not a brain window. On the other hand, the brain is practically invisible in the bone window, so small metastases in the brain matter will not be noticeable. We should always remember these technical details, since in most cases images in all windows are not transferred to film. The doctor conducting the examination views the images on the screen in all windows so as not to miss important signs of pathology.

[ 43 ], [ 44 ], [ 45 ]