Artifacts Produced on the Images During CT Scans

Introduction

The term artifacts as used in computer tomography (CT) refer to any inconsistency that occurs between the CT numbers in the rebuilt image and the actual shrinkage coefficient of the image. As images developed through Ct technology are obtained from numerous autonomous detector measurements, they are susceptible to artifacts. One of the major reasons is the assumption made during the process that all measurements are accurate. Any error emerging from the process results in the respective image developing some artifacts. There is a big range of artifacts that occur during the reconstruction of images which range from shading, rings, streaking, and distortion¹. Based on how these artifacts occur, they can be classified into scanner-based artifacts, those caused by the patient and physics-based. Several measures have been put in place to mitigate these effects. This paper aims at giving some of the artifacts witnessed, their causes, and some of the established mechanisms for avoiding or controlling their effects.

Types of artifacts

Physics-based artifacts

In developing images using CT technology, a beam of light from X-ray tubes is issued. This beam is comprised of numerous photons that have different energies. As this light transit across an object being scanned, its average energy increases. This is referred to as beam hardening². The rate of absorption of these photons varies with the mean energy of the beam with the light beam being absorbed faster than the heavier beams. It is this variance in the energy of the beam that leads to the development of cupping and streaking artifacts.

Cupping artifacts

Generally, when X-rays transit through the central part of a consistent cylindrical phantom, they tend to be hardened. This is not the case for those rays passing along the edges of the same phantom. This is because the rays passing along the edges encounter limited materials. Due to the hardening of the beam, its shrinkage rate decreases making it stronger when it gets to the detectors. This is contrary to what is expected had it not been hardened. Consequently, the obtained attenuation profile varies with the actual profile that would be obtained in absence of beam hardening. A close look at the CT numbers reflects cupped shape features³.

Streaking artifacts

For a non-uniform material, streaks or dark bands may result in the middle of two thick objects within an image. This occurs due to rays passing through one object at a specific position being less hardened than when the same rays pass through the two objects at a different position. In most cases, streaking artifacts may be reported in bony areas as well as in areas where the scanning is done using contrast material.

Patient-Based artifacts

Other artifacts emerge during image processing as a result of the patient. These arise if the patient has some metallic elements in him or her such as jewels. In addition, if the patient moves during scanning, the resultant image may develop some artifacts in form of shading or streaking. In instances where the operator is using positron emission tomography in conjunction with CT (PET/CT), truncation artifacts arise due to the variation in the range viewed between the PET and the CT⁴. This mostly occurs for big patients or when a patient is scanned with his or her arms down. When part of the patient goes beyond the field of view, this part is truncated thus not being reflected in the developed CT image.

In addition, truncation leads to streaking artifacts along the edges of the image produced using the PET/CT method leading to overestimation of the shrinkage factor employed in obtaining PET data. The density of metallic objects makes it hard for the beam of rays to penetrate freely. This leads to incomplete shrinkage of the beam of light resulting in an unclear image. Another human-based way in which artifacts are produced comes as a result of an incomplete projection of the entire portion being scanned. For instance, a patient may place his or her hands in the projection field while they are not required. As a result, the resultant image may include features that are not required while others may not be scanned. All these may lead to adverse artifacts rendering the developed image useless.

Scanner-based artifacts

Ring artifacts

The commonly used scanner in image development is termed the third-generation scanner. This comprises a detector integrated into an x-ray tube. There is a calibration for every detector. In case a detector goes out of its calibration, it generates erroneous results during the scanning process⁵. The resulting image contains circular artifacts known as ring artifacts. This is mostly experienced when using numerous solid detectors.

Methods used to remove or reduce the artifacts

Cupping and streaking artifacts

Based on how the artifact arises, there are different mechanisms employed to avoid their occurrence or reduce the severity of their effects. Some of the mechanisms are mechanically done by the operator while others are automatically done with the help of software. To mitigate effects caused by cupping and streaking artifacts, there has been the establishment of a calibration correction mechanism, filtration method as well as software that corrects beam hardening. Filtration entails use of a metallic medium to pre-harden the beam prior to it transiting through the patient being scanned. This is aimed at getting rid of photons with low energy so as to remain with a beam of rays with uniform energy⁶. In addition, a filter referred to as a bowtie issued to harden rims of the beam that traverses the thinner parts of the body.

For calibration correction, scanners are calibrated with varied ranges of phantoms when being manufactured. This facilitates calibrating detectors with reparations aligned for beam hardening impacts of varied parts of the victim. As the human body does not match the developed phantom, there are incidences of limited cupping artifacts that are reflected in the reconstructed image despite the effort made to mitigate this. Beam hardening correction software comprises a repetitive algorithm⁷. This is in most cases used when one is reconstructing a bony area. The software facilitates in reducing cases of blurred images in regions that contain intersections of bones and soft tissues. This is mostly used in brain scans. The software also aids in avoiding the formation of dark bands in heterogeneous regions. Apart from the aforementioned mechanisms, cupping and streaking artifacts can be mechanically reduced or avoided. This can be done by ensuring that the patient is properly positioned before scanning. Furthermore, the operator may alter the gantry to ensure that its focus on the field to be scanned.

Avoiding metallic artifacts and motion artifacts

To avoid artifacts that arise from metallic elements found within the patients, they are asked to put off all removable metallic objects they may be wearing. However, there are non-removable metals that may be found within the patients such as surgical clips, dental fillings, and prosthetic gadgets. In this case, the gantry angle is altered to ensure that regions containing these elements are not scanned. If it becomes hard to exclude these areas when scanning, the beam energy is increased to reduce attenuation. There is special software that has been developed to facilitate scanning patients with non-removable metallic objects⁸. Different interpolation methods are used to compensate for overhanging in the shrinking profile. Despite this software aiding in removing streaking artifacts, it does not give a clear image of the intersections of the metal and the body which in most cases is the focal point of the entire scan. Additionally, beam hardening correction software is used together with this to avoid artifacts that arise as a result of beam hardening.

To avoid motion artifacts, patients are asked to be still during the entire scanning duration. For cases of pediatric patients who are not capable of remaining still, sedative drugs are administered to the patient. There has also been the establishment of scanning machines that take the shortest time possible for areas that are prone to movement. This is to ensure that scanning is done before the patient moves. For the case of movement caused by respiration, patients are requested to hold their breath during the scanning process. Other inbuilt methods used to reduce chances of motion artifacts include cardiac gating, over and underscan techniques, and software correction.

Specialists have identified that most of the artifacts arise at the beginning and the end of the scanning process. For a rotational scan, an extra 10% of the rotation is added to the scan thus going round for 396⁰. Afterward, an average is computed for areas that are projected twice⁹. This aids in mitigating the effect of motion artifacts. Advancement in technology has led to the development of software that automatically reduces incidences of motion artifacts occurring during scanning. The scanner mechanically uses condensed weighting to areas where the scanning begins and ends. This is to avoid these regions playing a significant role in the interpretation of the reconstructed image. For patients with rapid heartbeats, reconstructed image may exhibit numerous artifacts which if not well interpreted may be perceived to be diseases leading to the patient being prescribed drugs he or she does not deserve to take¹⁰. To avoid the occurrence of such incidences, there has been the establishment of techniques where images of the heart are developed from a portion of the entire cardiac cycle. This is done when the patient has a low number of heartbeats. In such instances, good images are obtained by employing electrocardiographic gating together with sophisticated image reconstruction techniques.

Avoiding scanner-based artifacts

The occurrence of ring artifacts in an image indicates that there is a problem with how the detectors are calibrated. This calls for the operator to recalibrate or repair them to avoid the occurrence of artifacts in the reconstructed image. To avoid the occurrence of ring artifacts, operators ensure that they have identified to correct field that needs to be scanned. This is by ensuring that they tailor their calibration data based on the body structure of the patient being scanned. Most of the present scanners have solid detectors¹¹. This makes them susceptible to producing ring artifacts in the reconstructed images. However, technology has led to the development of software that identifies and corrects variations that arise in these detectors.

Conclusion

Artifacts that arise during image reconstruction using computer tomography may adversely affect the quality of the image produced. This may lead to doctors perceiving the patient as to be having diseases that may not be real. It is with this respect that scanner designers have produced varied design techniques used in manufacturing the scanners to control and even avoid the occurrence of these artifacts. Some of the techniques that are currently used to reduce the chances of artifacts occurring when scanning is based on the type of artifacts to be controlled and the anatomy of the body being scanned. Operators have to ensure that calibration data used in every scanner is aligned with the body anatomy of the patient. In addition, positioning of the patients is done to ensure that the scanner projects the required fields. This helps in ensuring that unwanted regions are not scanned thus avoiding cases of truncation. At times, it becomes hard to manually eliminate the occurrence of artifacts. This is controlled through the use of special software that has been developed for this purpose. The combination of these techniques helps in reconstructing images clear images that are free from artifacts thus making it possible for doctors to identify and effectively treat the diseases various patients are suffering from.

Reference List

Barrett, Julia F, Keat, Nicholas, Platten, Daniel, Lewis, Martin A & Edyvean Stephen. “Cardiac CT scanning.” MHRA Report 03076. London: Medicines and Healthcare Products Regulatory Agency, 2003.

Herman, Gabor T. Fundamentals of computerized tomography: Image reconstruction from projection, 2nd edition. New York: Springer, 2009.

Hsieh, John.” Image artifacts: appearances, causes and corrections.” Computed tomography: principles, design, artifacts and recent advances. Bellingham, Wash: SPIE Press, 2003. 167-240.

Kachelriess, Marc, Watzke, Oliver, Kalender, Willi A. “Generalized multi-dimensional adaptive filtering for conventional and spiral single-slice, multi-slice, and cone-beam CT.” Med Phys 28, (2001): 475-490.

Kamel, Ehab M, Burger, Cyrill, von Schulthess, Gustav K & Goerres, Gerhard W. “Impact of metallic dental implants on CT-based attenuation correction in a combined PET/CT scanner.” Eur Radiol 13, (2003): 724–728.

Kinahan, Paul E, Townsend, David W, Beyer, Thomas & Sashin, David. “Attenuation correction for a combined 3D PET/CT scanner.” Med Phys 25, (1998): 2046–2053.

Osman, Medhat M, Cohade, Christian, Nakamoto, Yuji & Wahl, Richard L. “Respiratory motion artifacts on PET emission images obtained using CT attenuation correction on PET-CT.” Eur J Nucl Med Mol Imaging 30, (2003): 603–606.

Seeram, Euclid. “Image quality.” Computed tomography: physical principles, clinical applications and quality control. 2nd ed. Philadelphia, Pa: Saunders, 2001. 174-199.

Taguchi, Katsuyuki, Aradate, Hiroshi. “Algorithm for image reconstruction in multi-slice helical CT.” Med Phys 25, (1998): 550-561.

Udupa, Jayaram.K & Gabort, Herman T. 3D Imaging in Medicine, 2nd Edition. Ney York: CRC Press, 2000.

Wilting, James E &Timmer, John. “Artifacts in spiral-CT images and their relation to pitch and subject morphology.” Eur Radiol 9, (1999): 316-322.

Footnotes

1. John Hsieh.” Image artifacts: appearances, causes, and corrections.” Computed tomography: principles, design, artifacts and recent advances (Bellingham, Wash: SPIE Press, 2003), 167-240.
2. Marc Kachelriess, Oliver Watzke, Willi Kalender A. “Generalized multi-dimensional adaptive filtering for conventional and spiral single-slice, multi-slice, and cone-beam CT.” Med Phys 28, (2001): 475-490.
3. Euclid Seeram. “Image quality.” Computed tomography: physical principles, clinical applications, and quality control. 2nd ed (Philadelphia, Pa: Saunders, 2001), 174-199.
4. James Wilting E & John Timmer. “Artifacts in spiral-CT images and their relation to pitch and subject morphology.” Eur Radiol 9, (1999): 316-322.
5. Katsuyuki Taguchi & Hiroshi Aradate. “Algorithm for image reconstruction in multi-slice helical CT.” Med Phys 25, (1998): 550-561.
6. Paul Kinahan E, David Townsend W, Thomas Beyer & David Sashin. “Attenuation correction for a combined 3D PET/CT scanner.” Med Phys 25, (1998): 2046–2053.
7. Ehab Kamel M, Cyrill Burger, Gustav von Schulthess K & Gerhard Goerres W. “Impact of metallic dental implants on CT-based attenuation correction in a combined PET/CT scanner.” Eur Radiol 13, (2003): 724–728.
8. Medhat Osman M, Christian Cohade, Yuji Nakamoto & Richard Wahl L. “Respiratory motion artifacts on PET emission images obtained using CT attenuation correction on PET-CT.” Eur J Nucl Med Mol Imaging 30, (2003): 603–606.
9. Jayaram Udupa K & Herman Gabort T. 3D Imaging in Medicine, 2nd Edition. New York: CRC Press, 2000.
10. Julia Barrett F, Nicholas Keat, Daniel Platten, Martin Lewis A & Stephen Edyvean. “Cardiac CT scanning.” MHRA Report 03076 (London: Medicines and Healthcare Products Regulatory Agency, 2003), 48.
11. Gabor Herman T. Fundamentals of computerized tomography: Image reconstruction from projection, 2nd edition (New York: Springer, 2009), 87-94.