GENERAL PRINCIPLES ASSOCIATED WITH GOOD IMAGING TECHNIQUE:
TECHNICAL, CLINICAL AND PHYSICAL PARAMETERS

CT images are the result of the interplay of physical phenomena giving rise to attenuation by the patient of a thin fan beam of x-rays, and complex technical procedures. Each image consists of a matrix of pixels whose CT numbers (measured in Hounsfield Units, HU) represent attenuation values for the volume elements (voxels) within the slice. The quality of the image relates to the fidelity of the CT numbers and to the accurate reproduction of small differences in attenuation (low contrast resolution) and fine detail (spatial resolution). Good imaging performance demands that image quality should be sufficient to meet the clinical requirement for the examination, whilst maintaining the dose to the patient at the lowest level that is reasonably practicable. In order to achieve this, there must be careful selection of technical parameters that control exposure of the patient and the display of the images, and also regular checking of scanner performance with measurement of physical image parameters as part of a programme of quality assurance.

1.

Technical Parameters: Display and Exposure Parameters with an Influence on Image Quality and Dose

1.1 Nominal slice thickness

The nominal slice thickness in CT is defined as the full width at half maximum (FWHM) of the sensitivity profile, in the centre of the scan field; its value can be selected by the operator according to the clinical requirement and generally lies in the range between 1mm and 10mm. In general, the larger the slice thickness, the greater the low contrast resolution in the image; the smaller the slice thickness, the greater the spatial resolution. If the slice thickness is large, the images can be affected by artefact, due to partial volume effects; if the slice thickness is small (e.g. 1-2mm), the images may be significantly affected by noise.

1.2 Inter-slice distance/pitch factor

Inter-slice distance is defined as the couch increment minus nominal slice thickness. In helical CT the pitch factor is the ratio of the couch increment per rotation to the nominal slice thickness at the axis of rotation. In clinical practice the inter-slice distance generally lies in the range between 0 and 10mm, and the pitch factor between 1 and 2. The inter-slice distance can be negative for overlapping scans which in helical CT means a pitch < 1. In general, for a constant volume of investigation, the smaller the inter-slice distance or pitch factor, the higher both the local dose and the integral dose to the patient. The increase in the local dose is due to superimposition of the dose profiles of the adjacent slices. The increase in the integral dose is due to an increase in the volume of tissue undergoing direct irradiation as indicated by a packing factor.

In those cases where 3D reconstruction or reformatting of the images in coronal, sagittal or oblique planes is required, it is necessary to reduce the inter-slice distance to zero or perform a helical scan. In screening or examinations performed with regard to control of disease it can be diagnostically justifiable to have an inter-slice distance corresponding to half the slice thickness or a pitch factor of 1.5-2.

1.3 Volume of investigation

Volume of investigation, or imaging volume, is the whole volume of the region under examination. It is defined by the outermost margins of the first and last examined slices or helical exposure. The extent of the volume of investigation depends on the clinical needs; in general the greater its value the higher the integral dose to the patient, unless an increased inter-slice distance or pitch factor is used.

1.4 Exposure factors

Exposure factors are defined as the settings of x-ray tube voltage (kV), tube current (mA) and exposure time (s). In general, one to three values of tube voltage (in the range between 80 and 140 kV) can be selected. A high tube voltage is recommended for high resolution CT (HRCT) of the lungs and may be used for examination of osseous structures such as the spine, pelvis and shoulder. Soft tissue structures are usually best visualised using the standard tube voltage for the given equipment. In some cases of quantitative computed tomography (QCT), the same slice is examined with two different values of tube voltage, in order to subtract corresponding images and derive information about the composition of particular tissues. At given values of tube voltage and slice thickness, the image quality depends on the product of x-ray tube current (mA) and exposure time (s), expressed in mAs. Absolute values of mAs necessary for an imaging task will depend on the type of scanner and the patient size and composition. For a particular CT model, an increase in radiographic exposure setting (mAs) is accompanied by a proportional increase in the dose to the patient. Relatively high values of radiographic exposure setting (mAs) should therefore be selected only in those cases where a high signal to noise ratio is indispensable.

A method for correlating the exposure setting (for a given tube voltage) with the overall image quality is by drawing contrast-detail curves for each available setting. These curves express the minimum size of detail which can still be recognised in the CT image for a given difference in contrast between the detail and the surrounding medium.

1.5 Field of view

Field of view (FOV) is defined as the maximum diameter of the reconstructed image. Its value can be selected by the operator and generally lies in the range between 12 and 50 cm. The choice of a small FOV allows increased spatial resolution in the image, because the whole reconstruction matrix is used for a smaller region than is the case with a larger FOV; this results in reduction of the pixel size. In any case, the selection of the FOV must take into account not only the opportunity for increasing the spatial resolution but also the need for examining all the areas of possible disease. If the FOV is too small, relevant areas may be excluded from the visible image. If raw data are available the FOV can be changed by post-processing.

1.6 Gantry tilt

Gantry tilt is defined as the angle between the vertical plane and the plane containing the x-ray tube, the x-ray beam and the detector array. Its value normally lies in the range between -25° and +25°. The degree of gantry tilt is chosen in each case according to the clinical objective. It may also be used to reduce the radiation dose to sensitive organs or tissues and/or to reduce or eliminate artefacts.

1.7 Reconstruction matrix

Reconstruction matrix is the array of rows and columns of pixels in the reconstructed image, typically 512 x 512.

1.8 Reconstruction algorithm

Reconstruction algorithm (filter, or kernel) is defined as the mathematical procedure used for the convolution of the attenuation profiles and the consequent reconstruction of the CT image. In most CT scanners, several reconstruction algorithms are available. The appearance and the characteristics of the CT image depend strongly on the algorithm selected. Most CT scanners have special soft tissue or standard algorithms for examination of the head, abdomen etc. Depending on clinical requirements, it may be necessary to select a high resolution algorithm which provides greater spatial resolution, for detailed representation of bone and other regions of high natural contrast such as pulmonary parenchyma.

1.9 Window width

Window width is defined as the range of CT numbers converted into grey levels and displayed on the image monitor. It is expressed in HU. The window width can be selected by the operator according to the clinical requirements, in order to produce an image from which the clinical information may be easily extracted. In general, a large window (for instance 400 HU) represents a good choice for acceptable representation of a wide range of tissues. Narrower window widths adjusted to diagnostic requirements are necessary to display details of specific tissues with acceptable accuracy.

1.10 Window level

Window level is expressed in HU and is defined as the central value of the window used for the display of the reconstructed CT image. It should be selected by the viewer according to the attenuation characteristics of the structure under examination.

2. Clinical and Associated Performance Parameters

A series of clinical factors play a special part in the optimal use of ionising radiation in CT. They are described here in order to ensure that an appropriate CT examination is carried out, providing diagnostic quality with a reasonable radiation dose for the patient.

A CT examination should therefore only be carried out on the basis of a justifiable clinical indication, and exposure of the patient should always be limited to the minimum necessary to meet clinical objectives.

Adequate clinical information, including the records of previous imaging investigations, must be available to the person approving requests for CT.

In certain applications, in order to practice CT effectively, prior investigation of the patient by other forms of imaging might be required.

2.1 Supervision

CT examinations should be performed under the clinical responsibility of a radiologist/practitioner according to the regulations (4) and standard examination protocols should be available.

Effective supervision may support radiation protection of the patient by terminating the examination when the clinical requirement has been satisfied, or when problems occurring during the examination (for example, unexpected uncooperation by the patient or the discovery of contrast media residue from previous examinations) cannot be overcome.

Problems and pitfalls: the responsible radiologist/practitioner should be aware of clinical or technical problems which may interfere with image quality. Many of these are particular to specific organs or tissues and may lead to modification of technique. The radiologist/practitioner and the radiographer must be aware of manoeuvres which may be used to overcome such diagnostic or technical problems in order to provide a clinically relevant examination.

2.2 Patient Preparation

The following patient-related operational parameters play an important role for the quality of the CT examination:

  2.2.1

Cooperation. Patient cooperation should be ensured as far as possible prior to the examination. An explanation of the procedure should be given to each patient. Good communication with and control of the patient is equally necessary during the whole examination.

  2.2.2

Protective Shielding. Relevant protection for sensitive organs outside the imaging field is a lead-purse for the male gonads, if the edge of the volume of investigation is less than 10 - 15 cm away. The protection of female gonads by wrap-around lead has not yet been demonstrated (7,8). Appropriate protection measures must be applied to persons who, for clinical reasons or to ensure cooperation, may need to accompany patients in the examination room during the examination.

  2.2.3

Clothing. The area of examination should be free of external metal or other radio- dense items where possible. Special attention must be given to eliminating any x-ray dense material in the patient's clothes or hair.

  2.2.4

Fasting. Fasting prior to the examination is not essential. Restraint from food, but not fluid, is recommended if intravenous contrast media are to be given.

  2.2.5

Intravenous contrast media. These are needed in some examinations and must be employed in a manner appropriate to the clinical indication, taking into consideration the risk factors.

  2.2.6

Oral or cavitatory contrast media. Oral contrast medium may be required in abdomino-pelvic examinations and must be administered at times and in doses appropriate to the indication. Administration of contrast medium per rectum may be required in some examinations of the pelvis and a vaginal tampon should be used in some examinations for gynaecological applications.

  2.2.7

Positioning and motion. Most CT examinations are carried out with the patient supine. In this position the patient is most comfortable with the knees flexed. Alternate positioning may be required to aid comfort and cooperation, for appropriate display of anatomy, to reduce absorbed radiation to particular organs, or to minimise artefact. Motion should be kept to a minimum to reduce artefacts; typical sources of artefacts are involuntary patient movement, respiration, cardiovascular action, peristalsis and swallowing.

2.3 Examination Technique
  • Scan projection radiograph.

    A scan projection radiograph permits the examination to be planned and controlled accurately, and provides a record of the location of images. It is recommended that this is performed in all cases. In general such imaging provides only a small fraction of the total patient dose during a complete CT procedure (9)

  • Clinical aspects of setting the appropriate technical parameters.

    These parameters must be set according to the area of examination and clinical indication, as follows:

    • Nominal slice thickness is chosen according to the size of the anatomical structure or lesion that needs to be visualised. Staff should be aware of the implications of choice of slice thickness in relation to the image quality and radiation dose to the patient.

    • Inter-slice distance is chosen according to the area under examination and the clinical indication. Staff should be aware of the risk of overlooking lesions which fall in the inter-slice interval during serial CT. In general, the interval should not exceed one half of the diameter of suspected lesions. This problem is absent in helical scanning, when an appropriate reconstruction index is used.

    • Field of view (FOV). Selection of FOV must respect image resolution and the need to examine all areas of possible disease. If the FOV is too small, disease may be excluded from the visible image.

    • Exposure factors: tube voltage (kV), tube current (mA) and exposure time (s) affect image quality and patient dose. Increasing exposure increases low contrast resolution by reducing noise but also increases patient dose. Patient size is an important factor in determining the image noise. Image quality consistent with the clinical indications should be achieved with the lowest possible dose to the patient. In certain examinations image noise is a critical issue and higher doses might be required.

    • The volume of investigation is the imaging volume, defined by the beginning and end of the region imaged. It should cover all regions of possible disease for the particular indication.

    • Reconstruction algorithm: this is set according to the indication and area under examination. For most examinations, images are displayed utilising algorithms suitable for soft tissues; other algorithms available include those providing greater spatial resolution for detailed display of bone and other areas of high natural contrast.

2.4 Helical or Spiral CT

Helical or spiral CT is obtained by continuous tube rotation coupled with continuous patient transport through the gantry, resulting in volumetric data acquisition. Due to the high speed and ease of image performance with this technique it should be emphasized that helical CT presents particular challenges in radiation protection and it should not be used without clinical justification. Helical CT is in most cases preferable to serial CT because of advantages such as:

  • a possibility of dose saving:
    • the repeating of single scans, which sometimes results from lack of patient cooperation in serial CT, is reduced in spiral CT because of the shorter examination times involved

    • for pitch > 1 the dose will be reduced compared with contiguous serial scanning; there are no data missing as may be the case with the use of an inter-slice interval in serial CT

    • the practice of using overlapping scans or thin slices in serial CT for high quality 3D display or multi-planar reconstructions is replaced by the possibility of reconstructing overlapping images from one helical scan volume data set

  • extremely shortened examination time:
    • makes it possible to acquire continuous patient data during a single breath-hold; problems with inconsistent respiration can thereby be avoided

    • disturbances due to involuntary movements such as peristalsis and cardiovascular action are reduced

    • may optimize scanning with the use of intravenous contrast media (10,11)

  • images can be reconstructed for any couch position in the volume of investigation:
    • anatomical misregistration is avoided

    • equivocal lesions can be further evaluated without additional patient exposure

    • the possibility of displaying the data volume in transverse slices reconstructed at intervals smaller than the x-ray beam collimation results in overlapping slices which, in combination with reduced or eliminated movement artefacts, makes it possible to perform high quality three-dimensional (3D) and multi planar reconstructions with smooth tissue contours. This is used especially in skeletal (12) and vascular imaging (CT angiography) (10).

Helical CT, however, has drawbacks such as:
  • ease of performance may tempt the operator to extend the examination unjustifiably, either by increasing the imaging volume, or by repeated exposure of a region

  • although most image quality parameters are equivalent for contiguous serial CT and helical CT performed with a pitch = 1 (13,14), the performance of helical CT with a pitch greater than 1.5 may imply lower and possibly insufficient diagnostic image quality due to reduced low contrast resolution (10,14)

  • spatial resolution in the z-direction is lower than indicated by the nominal slice width (13,15) unless special interpolation is performed (15)

  • the technique has inherent artefact

When using helical CT in conjunction with intravenous injection of contrast media to provide optimally enhanced images, careful timing of exposure relative to intravenous injection is mandatory.

2.5 Image viewing conditions

It is recommended that initial reading of CT images is carried out from the TV monitor. Display of images and post-processing image reconstruction should be at a display matrix of at least 512 x 512.

Brightness and contrast control on the viewing monitor should be set to give a uniform progression of the grey scale from black to white. A calibrated grey-scale would be preferable.

Settings of window width and window level dictate the visible contrast between tissues and should generally be chosen to give optimum contrast between normal structures and lesions.

2.6 Film Processing

Optimal processing of the film has important implications for the diagnostic quality of the image stored on film. Film processors should be maintained at their optimum operating conditions as determined by the manufacturer and by regular and frequent quality control procedures.

3. Physical Parameters: Physical Measures of Scanner Performance.

The quality of the CT image may be expressed in terms of physical parameters such as uniformity, linearity, spatial resolution, low contrast resolution and absence of artefacts according to IEC recommendations (16). It depends on the technological characteristics of the CT scanner, the exposure factors used and image viewing conditions. Quality may be assessed by quantitative measurement of the parameters listed above, using suitable test phantoms, and by the appearance of artefacts. These measurements should be conducted regularly, in order to guarantee the maintenance of performance of the CT scanner during its whole period of use. It is essential that such technical quality control has been performed when using the criteria presented in these guidelines.

3.1 Test Phantoms

Test phantoms (phantom of a standardised human shape or test objects of a particular shape, size and structure) are used for the purposes of calibration and evaluation of the performance of CT scanners. Performance is checked by acceptance tests after installation and important repairs, and by periodic quality control tests, as established in standardised protocols. A number of test phantoms are commercially available and most manufacturers provide one or more test objects.

The test phantoms should allow for the following parameters to be checked: mean CT number, uniformity, noise, spatial resolution, slice thickness, dose and positioning of couch (16).

3.2 CT Number

The accuracy of CT number is verified by scanning a test object utilising the usual operating parameters and reconstruction algorithms. The CT number is affected by the x-ray tube voltage, beam filtration and object thickness. The CT number of water is by definition equal to 0 HU and the mean CT number measured over the central region of interest (ROI) should be in the range +/- 4HU.

3.3 Linearity

Linearity concerns the linear relationship between the calculated CT number and the linear attenuation coefficient of each element of the object. It is essential for the correct evaluation of a CT image and, in particular, for the accuracy of QCT. Deviations from linearity should not exceed +/- 5HU over specific ranges (soft tissue or bone).

3.4 Uniformity

Uniformity relates to the requirement for the CT number of each pixel in the image of a homogeneous object to be the same within narrow limits over various regions of the object such as a cylindrical 20 cm diameter phantom of water-equivalent plastic. The difference in the mean CT number between a peripheral and a central region of a homogeneous test object should be < 8HU. Such differences are largely due to the physical phenomenon of beam hardening.

3.5 Noise

Picture element (pixel) or image noise is the local statistical fluctuation in the CT numbers of individual picture elements of a homogeneous ROI. Noise is dependent on the radiation dose and has a marked effect on low contrast resolution. The magnitude of the noise is indicated by the standard deviation of the CT numbers over a ROI in a homogeneous substance. It should be measured over an area of about 10% of the cross-sectional area of the test object. Image noise diminishes with the use of a slightly flattened convolution kernel, with simultaneous reduction of spatial resolution and an increase in low contrast resolution. Image noise is inversely proportional to the square root of the dose and to the slice thickness. For example, if the dose is halved then the noise will only increase by about 40%. Conversely, a reduction in slice thickness requires a proportionate increase in dose in order to avoid an increase in noise. The medical problem under study and the corresponding image quality required should determine what level of image noise and what patient dose are reasonably practicable.

3.6 Spatial Resolution

Spatial resolution at high and low contrast are interdependent and critical to image quality and good imaging of diagnostically important structures.

The spatial resolution at high contrast (high contrast resolution) determines the minimum size of detail visualised in the plane of the slice with a contrast > 10%. It is affected by the reconstruction algorithm, the detector width, the slice thickness, the object to detector distance, the x-ray tube focal spot size, and the matrix size.

The spatial resolution at low contrast (low contrast resolution) determines the size of detail that can be visibly reproduced when there is only a small difference in density relative to the surrounding area. Low contrast resolution is considerably limited by noise. The perception threshold in relation to contrast and detail size can be determined, for example, by means of a contrast-detail curve. In such determinations, the effects of the reconstruction algorithm and of the other scanning parameters have to be known. Dose and the corresponding image noise greatly affect low contrast resolution.

3.7 Slice Thickness

The slice thickness is determined in the centre of the field of view as the distance between the two points on the sensitivity profile along the axis of rotation at which response has fallen to 50%. Certain deviations in thickness should not be exceeded because of the effect of slice thickness on image detail; for example, with a nominal slice thickness > 8mm, a maximum deviation of ± 10% is acceptable; tolerable deviations for smaller slice thickness of 2-8 mm and < 2 mm are ± 25% and ± 50%, respectively.

The use of post-patient collimation, which is inherent in some CT equipment to reduce the slice sensitivity profile, leads to significant increases in the patient dose for a series of contiguous slices (9).

3.8 Stability of CT numbers

Stability is defined as the maintenance over time of constancy of CT number and of uniformity. It can be checked by means of a suitable test object, containing at least three specimens of different materials, e.g. water, Polymethylmethacrylate (PMMA) and Teflon. Deviations should not exceed +/- 5 CT numbers with respect to initial mean values. A similar tolerance should be applied in the verification of uniformity, as measured in three ROI's, each containing approximately 100 pixels and placed respectively at the centre, at the periphery, and in a position intermediate between the centre and the periphery of the reconstructed image.

3.9 Positioning of couch

The accuracy of positioning of the patient couch is evaluated by moving the loaded couch a defined distance relative to the gantry and subsequently moving it back to the start position (16). Positional accuracy includes both deviation in longitudinal positioning and also backlash. Maximum tolerances of ±2 mm apply to both criteria. These also apply to mobile CT equipment.