![Contd…
• The value of a CT number is given by the followings
CT number = k[µt-µw/µw]
Where µt= attenuation co-efficient of the tissue in the pixel under
analysis
µw= x-ray attenuation co-efficient of water
k=constant that determines the scale factor for the range of
CT numbers(500)(1000)
• Its value ranges from -1000 to +3000.](https://image.slidesharecdn.com/ctimagequalityartifactsanditremedy-190620151954/85/Ct-image-quality-artifacts-and-it-remedy-9-320.jpg)
Ct image quality artifacts and it remedy
- 1. CT Image Quality Artifacts and It Remedy Yasahawant ku. Yadav NAMS Bir Hospital BSc.MIT 2nd year
- 2. Outline • Image quality • Factors affecting image quality • Artifacts • Types of artifacts • References
- 3. CT image characteristics • CT image is different from the conventional radiographic image. • It is artificially created from data received and is not a projected image. • In radiography, x-rays form an image directly on the image receptor. • With CT imaging systems, the x-rays form stored electronic image that is displayed as a matrix of intensities.
- 5. Image Matrix • In CT image, it is the image area which is subdivided into rows and columns creating small areas having special numerical values. • The small areas known as picture elements • Pixel size= image size (mm)/matrix size (pixels) • The image resolution will be better with a larger image matrix as the pixel size will be smaller
- 6. FOV • is the maximum diameter of the area of the scanned object that is represented in the reconstructed image. ... • SFOV is the parameter that determines how much anatomy is scanned • DFOV determines how much of the scan field of view is reconstructed into an image. • Pixel size = DFOV/matrix size • Voxel (volume element) is 3-D element of anatomy represented by the 2-D pixel.
- 7. Voxel • It is the tissue volume. • It is determined by multiplying the pixel size by the thickness of the ct image slice. • Voxel size(mm3) =pixel size(mm2) x slice thickness(mm) •The depth of voxel or z axis correlates to the slice thickness.
- 8. CT Number • The numeric information contained in each pixel of a CT image. • Also called Hounsfield number or Hounsfield unit. • Is related to the composition and nature of the tissue imaged. • Is used to represent the density of tissue.
- 9. Contd… • The value of a CT number is given by the followings CT number = k[µt-µw/µw] Where µt= attenuation co-efficient of the tissue in the pixel under analysis µw= x-ray attenuation co-efficient of water k=constant that determines the scale factor for the range of CT numbers(500)(1000) • Its value ranges from -1000 to +3000.
- 10. CT numbers for various tissues S.N. Tissue CT number approax. 1 Dense bone +3000 2 Muscle +50 3 White matter +45 4 Gray matter +40 5 Blood C +20 6 CSF +15 7 Water 0 8 Fat -100 9 Lungs -200 10 Air -1000 10
- 11. Image Quality • Image quality (clarity) is the visibility of diagnostically important structure in the CT image. or. • The ability to resolve anatomy in an image. Many factors affect the quality of the image produced. Some of these variables can be regulated by the operator,
- 12. Contd.. • Milliampere (mA) level • Scan time • Slice thickness • Field of view • Reconstruction algorithm • Kilovolt-peak (kVp) • Pitch (helical)
- 13. Contd.. •Image quality in CT depends upon following factors: oSpatial resolution oContrast resolution oTemporal resolution oNoise oLinearity oUniformity oArtifacts
- 14. Resolutions 3 components of resolution. 1. Spatial resolution • Spatial resolution of CT system is its ability to distinguish, as separate images, two objects that are very close to each other • The degree of blurring is a measure of spatial resolution of system • Is usually described as the size of an object that can be viewed. • In CT, it is described by the quantity called spatial frequency.
- 15. Contd… • Slice Thickness • Pitch Slice Thickness • Inversely related to spatial resolution • Increased slice thickness, scan or viewing, decreases detail in an image and more partial volume averaging
- 17. Pitch • Inversely related to spatial resolution • Increases in pitch will result in a decrease in image detail and increase in partial volume averaging,
- 19. Spatial frequency • Relates the no. of line pairs in a given length expressed as centimeters or millimeters. • One line pair consists of the line and an interspace of the same width as the line. • Unit;-line pair per centimeter(lp/cm) or line pair per millimeter(lp/mm) • Screen-film radiography- 8 lp/mm and CT- 2 lp/mm
- 20. Contd… • A low spatial frequency represents large objects & high spatial frequency represents small objects. • The absolute object size that can be resolved is equal to one-half the reciprocal of the spatial frequency at the limiting resolution
- 21. MTF Modulation Transfer Function • Spatial resolution can be calculated from an analyzing the spread of information within the system.(0.1-0.62/mm) • This data analysis is known as modulation transfer function. • This is the most commonly used method of describing spatial resolution ability. • It is the plot of the ratio of image to object contrast (image fidelity) at each spatial frequency
- 22. Limiting Resolution on a given CT system at an MTF equal to 0.1 • Imaging system B has 0.1 MTF at 5.2lp/cm, where as A can manage only 3.5lp/cm. • Therefore ,B has better spatial resolution than A.
- 23. Contd… • MTF is always less than 1. • If MTF for 2 different systems should be interpreted, • Spatial frequency at an MTF equal to 0.1(limiting resolution) • MTF curve that extends farther to right indicates higher spatial resolution. • Curve that is higher at low spatial frequencies indicates better contrast resolution.
- 24. Factors affecting spatial resolution • x-ray tube focal spot size • Sampling: detector size (aperture) & Sample spacing(detector pitch) • Slice thickness: slice sensitivity profile • Reconstruction • Pixel size • Matrix size • Field of view • Patient motion
- 25. X-RAY FOCAL SPOT SIZE: • Focal-spot size does affect CT resolution but to a lesser extent than in radiography(anode heat storage capacities:1MHU -8 MHU). • There are usually two available focal spot sizes on CT scanners: Fine = 0.7 x 0.7 mm high-resolution and Broad = 1.2 mm • Larger focal spot cause more geometric unsharpness in the detected image & reduce spatial resolution.
- 26. Contd.. • Smaller focal spots give higher resolution but limit the maximum mA deliverable • Mag. Factor in CT are higher than in radiography because of fixed diameter gantry. • Mag. Factor range from about 1.6-2.7
- 27. Flying focal spot:- • Two focal spot positions for the same projection angle shifted by half the detector width. • The position of the focal spot is altered rapidly in the transaxial plane (as well as the Z-axis). This can double the number of projections sampled and improve spatial resolution • Flying focal spot of ~ 0.6 – 0.9
- 29. Sampling :- • Detector size (aperture): The use of smaller detectors increases the cut-off frequency of the image & improves the resolution at all frequencies • Detector pitch (spacing): In addition to a small aperture, closely spaced measurements are required for good resolution. • For fixed FOV, as the detector pitch decreases the no of rays increases & better resolution. • The no of views influences the convey of higher spatial frequencies in the image without artifacts.
- 30. For two objects to be seen as separate the detectors must be able to identify a gap between them.
- 31. Quarter detector shift • By shifting the detector elements by a distance of a quarter of the size of the detector elements, the theoretical achievable spatial resolution becomes twice as good.
- 32. Slice thickness:- • Thinner slice thickness allow better spatial resolution • In single slice CT, slice thickness equal to beam collimation. • In multidetector CT, slice thickness equal to width of the detector in slice thickness direction.
- 33. • Effect of slice thickness. Axial reconstructed CT images of a coronary CT angiogram in a 42-year-old man show (A) a small amount of noise at a slice thickness of 0.9 mm. Increasing the slice thickness to (B) 3 mm or (C) 5 mm reduces this noise, although spatial resolution is also decreased, resulting in reduced edge definition compared to thinner slice reconstructions.
- 34. RECONSTRUCTION FILTER :- • A reconstruction filter or algorithm or kernel is applied during filtered back projection reconstruction to remove the blur from images. • The appearance and the characteristics of the CT image depend strongly on the algorithm selected. • More detail can be achieved by using sharp filter, provides greater spatial resolution, used for bone and pulmonary parenchyma.
- 35. Contd.. • For example, a standard filters are designed to produce images with a maximum resolution of 6 lp/cm, Whereas a bone filter may image with a resolution of 10 lp/cm
- 36. • Image (A) is appropriate for evaluation of the enhanced areas in the liver, evaluation of these areas in image (B) is hampered due to image noise that results from the sharp reconstruction filter.
- 37. Contrast Resolution • Contrast resolution refers to the ability of a system to differentiate, on the image, objects with similar densities. • CT not only eliminates the superimposition of structures also can image tissues that vary only slightly in density and atomic number. • It is generally accepted that the contrast resolution of screen-film radiography is approximately 5%, whereas CT demonstrates contrast resolution of about 0.5 at 5 mm thickness. • CT is excellent in terms of contrast resolution because of scatter radiation rejection of pre-patient & pre-detector collimators.
- 38. Factor affecting contrast resolution mAs Directly related Doubling of the mAs of the study increases the SNR by √2 or 41% , and the CR improves PIXEL SIZE Directly As FOV increases so as pixel dimension and the no of x-rays passing through each pixel increases SLICE THICKNESS Directly Thicker slices uses more photons and have better SNR RECONSTRUCTION FILTER Inversely Bone filter produce lower CR and soft tissue filter improves contrast PATIENT SIZE Inversely larger pt attenuate more photons resulting in detection of fewer photons causes reduction in SNR as well as contrast resolution GANTRY ROTATION SPEED Inversely Faster the gantry rotation speed lesser the contrast resolution
- 39. Temporal resolution • Temporal resolution is the ability to resolve fast moving objects in the displayed CT image. • Good temporal resolution avoids motion artifacts and motion induced blurring of the image. • A good temporal resolution in CT is realized by fast data acquisition (fast rotation of the X-ray tube).
- 40. Contd.. • Temporal resolution can be improved further by using dedicated reconstruction algorithms (cardiac CT with a segmented reconstruction) or by using a dual source CT scanner. • There are no simple methodologies available yet that allow for measuring temporal resolution in a clinical setting.
- 41. Noise (Quantum mottle) • CT no are the average values i.e. pixels have a range of values greater than or less than CT no, these variation of pixel values represent image noise. • If all pixel values were equal, the noise would be zero. • large variation represents high image noise. • Appears on the image as graininess. • It is due to statistical fluctuation of no. of photons absorbed by the detector, not related to the mathematical reconstruction.
- 43. Contd.. • Noise is the percentage standard deviation of a large no. of pixels obtained from a water bath image. Noise(ơ)= where xi is each CT value, x is the average of at least 100 values & n is the no of CT values averaged.
- 44. Factors affecting noise 1. Incident x ray intensity (kVp , mAs and filtration) ↑mAs→↑SNR & ↑ contrast. Dose increases linearly with mAs per scan. 2. Quantum detection efficiency of detector. 3. Slice thickness( decrease in slice thickness increases noise) 4. Pixel size ( decrease in voxel size increases noise). Reducing pixel size increases spatial resolution (if dose levels are kept same) SNR decreases & contrast resolution also decreases due to SNR per pixel drop.
- 45. Contd.. 5. Reconstruction filter or kernel: low pass filter reduces and high pass filter increases noise (bone filter). 6. Patient size: For the same x-ray technique, larger patients more attenuation, ↓ SNR • In spiral scan noise also depends on interpolation algorithm.(360˚ LI and 180˚ LI) • wide window widths also help to disguise noise. For this reason, it is a common practice in CT to increase the window width on images of obese patients.
- 47. 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.
- 48. Contd.. • To check linearity calibration should be done frequently by Cat Phan phantom or 5 pin performance test phantom • Each of the 5 pins are made up of diff. plastic material having known physical and x-ray attenuation properties
- 49. Contd… • Deviations from linearity should not exceed +/- 5HU over specific ranges (soft tissue or bone). • Linearity is typically measured semiannually.
- 50. Uniformity • At any time that a uniform object( e.g. water bath) is imaged, the pixel values should be constant in all regions of reconstructed image. This characteristics is spatial uniformity. • But practically it is different at different time. • An internal software package allows plotting of CT no along any axis of the image as histogram or line graph. • If all values are within 2 SD of the mean value (±2ơ), the system is said to exhibit acceptable spatial uniformity.
- 51. • Axial image of a homogenous phantom used • for the assessment of CT number uniformity At any time for a uniform phantom, the CT No measurement should not change with the location of the selected region of interest (ROI)
- 52. Artifacts • Artifacts are distortions or errors in the image that are unrelated to the object scanned . • Undesirable optical density on image. 1. Systematic error between the CT numbers in the reconstructed image 2. Detector measurement imbalance 3. Networking 4. External objects
- 53. Classification ON THE BASIS OF APPEARANCE APPEARENCE CAUSE A. STREAKS - Improper sampling data, partial volume average, Pt motion, Beam hardening, Noise, spiral/ helical, Mechanical failure B. SHADING - Partial volume averaging , Beam hardening, Spiral/ Helical, Scatter radiation, Off focal radiation , Improper projection C. RINGS/ BANDS - Bad detector channel
- 54. Classification ON THE BASIS OF ORIGIN A. Physics based- caused by physical process involved in the acquisition of CT data -Beam Hardening -Partial volume effect -Photon starvation -Under sampling -Edge gradient artifact B. Patient based- -Metal artifacts -Motion artifacts -Out of field
- 55. Contd.. C. Scanner based- caused by imperfection in scanner function - Ring artifacts - Tube arching D. Helical and Multisection - produced by the image reconstruction process - Rod artifacts - cone beam artifacts - wind mill E . Artifacts due to multiplanar and 3-D reformation - Stair step - Zebra artifacts
- 56. BEAM HARDENING ARTIFACTS • As the x-ray beam passes through an object ,it becomes harder that means its mean energy increases because the lower energy photons are absorbed more rapidly than the higher energy photons. Two types of artifacts can result from this effects • A . Cupping artifacts • B . Streaks and dark band
- 57. Cupping artifact • X-ray passing through the middle portion of a uniform cylindrical phantom are hardened more than those passing through edges because they are passing through more material
- 58. Streaks and dark bands • If a high density materials severely reduces transmission, the detector may record no transmission and streaks and dark band appear.
- 59. Remedy Built in features for minimizing beam hardening:- Manufacturers minimize beam hardening by using filtration, calibration correction and beam hardening correction software. Avoidance of beam hardening by the operator:- It is possible to minimize beam hardening by the operator where it is possible to avoid scan of bone region either by means of patient positioning or by tilting the gantry. Filtration :- a flat piece of metallic material is used to pre-harden the beam by filtering out the lower energy components before it passes through the patient
- 60. Contd… • It is important to select the appropriate scan field of view to ensure that the scanner uses the correct calibration and beam hardening correction data and, on some systems, the appropriate bowtie filter.
- 61. Partial volume effect Partial volume artifacts occur when a scanned object is partially intruded into scan plan. When a voxel contains tissues with highly different attenuation coefficient, the total attenuation in voxel is due to contribution of both of them. Partial volume artifacts can be reduced with thinner slice acquisition and computer algorithm
- 62. • Fig; CT images of three 12mm diameter acrylic rods supported in air parallel to and approax.15cm from the scanner.1st image obtained with the rods partially intruded into the section width shows partial volume artifact and 2nd image of same rods fully intruded into the section width shows no partial volume artifacts
- 63. Photon starvation artifact • Occurs in highly attenuating region, due to insufficient photons passing through the widest part of patient. • Image appears noisy with streaks. • Can be minimized by automatic tube current modulation (increase in tube current in widest part in the section without unnecessary dose through narrower parts) • Adaptive filtration-smoothing of attenuation profile in areas of high attenuation before image reconstruction.
- 64. Sampling or aliasing artifact • Too large interval b/w projection(under sampling) can result in misregistration by the computer of information relating to sharp edges and small object. • This causes an effect called aliasing where fine striations (lines) appear in the image. • Aliasing does not have too serious effect on the diagnostic quality of an image • Under sampling artifacts can be avoided by increasing the possible largest no. of projection per rotation and by using high resolution technique.
- 65. Metal artifact • Presence of highly attenuating metal objects in the field of view such as dental restorations, orthodontic bands, surgical plates and pins cause streaks in image. • The primary reason that streaks occur from metal objects is because the objects exceed the attenuation values that CT system can faithfully image. • It can be avoided by removal of the metal object, by tilting gantry, use of special software correction.
- 66. Use of special software correction
- 67. How to reduce metal artifacts • High kVp , high tube current , low pitch , standard or smooth reconstruction filter minimize artifact due to metal. • Volumetric rendering techniques provide semitransparent views of bones that tend to reduce metal artifacts. • Use of wide windows to review or to film images (3000–4000 HU WW, 800 WL) facilitates visualization of structures adjacent to metal hardware and reduce the effects of metal artifacts.
- 68. 140 kV peak, smooth filter, bone window 120 kV peak, smooth filter, bone window. 120 kV peak, hard filter, soft tissue window
- 69. Motion artifacts • Motion artifacts occur when the patient (anatomy in different locations during different parts of the scan) moves during the acquisition. • Motion can be voluntary or involuntary. • Small motions →blurring, & larger physical displacements →artifacts that appear as double images or image ghosting. • The reconstructed image will display an object in motion as a streak in the direction of motion.
- 70. How to minimize • Use of immobilization device or positioning aids. • Sedation in children. • Short scan time and cardiac gating techniques. • Post processing Software correction.
- 71. Ring artifact • Commonly associated with third generation (rotate- rotate ). • Each detector and its associated electronics has a certain amount of drift, causing the signal levels from each detector to shift over time. • Each detector is responsible for the data corresponding to a ring in the image. • Ring artifacts arise from errors, imbalances, calibration drifts, or other measurement inaccuracies in an element of a detector array relative to its neighbors. • Selecting the correct scan field may reduce the artifact by using calibration data that fit more closely to pt. anatomy.
- 72. TUBE ARCHING ARTIFACT Tube arching occurs when there is a short circuit within the tube, typically from cathode to tube envelope Causes momentary loss of x-ray output Now a days such incidence is decreased due to replacement of glass enveloped tube with metal ceramic x- ray tube
- 73. Zebra artifact • Faint stripes may be seen in MPR and 3D images because of helical interpolation. • process gives rise to a degree of noise in homogeneity along the z axis. MIP image obtained with helical CT shows zebra artifacts
- 74. Wind mill artifact • Seen in thin-slice images reconstructed from high pitch helical multi-slice CT images • arise from not satisfying the Nyquist sampling criteria in the patient longitudinal direction. • Type of aliasing artifact. • The term windmill comes from the spiraling appearance of shading artifact.
- 76. References 1. CT for technologist 2. Previous slides 3. https://pubs.rsna.org/doi/full/10.1148/rg.246045065
- 78. Questions 1. It is more common practice to manipulate mAs rather than kVp when modifying the radiation dose why ? - First choice of mA is more flexible typically range from 20 to 800mA - Adjusting the mA instead of kVp effect the image quality more straightforward and predictable
- 79. 2. What is the relation of spatial frequency with spatial resolution and contrast resolution? 3. What is contrast resolution and temporal resolution? 4. Why ring artifacts occur in 3rd generation CT?
Editor's Notes
- #12: One definition of image quality in CT refers to how accurately the CT image reproduces the three dimensional attenuation distribution of the x-ray beam through the patient
- #22: The MTF describes the percentage of an object’s contrast that is recorded by the imaging system as a function of its size (spatial frequency).
- #51: The Hounsfield unit (HU) value of water is 0 HU while that of air is -1000 HU. These are defined for the CT value scale and are checked regularly during normal quality QC tests of the CT system to check for uniformity. The HU value for water should not differ over time. Uniformity is obtained by measuring the CT value of water using an appropriate water phantom. The accepted tolerance range for different measures over time of the mean CT value of the water phantom is usually +/- 4 HU and +/- 2 HU for uniformity.