Image artifacts are present in breast MRI just like other breast imaging modalities do. Breast MRI is technically challenging calling for excellent fat reduction, spatial resolution, and post contrast sequence rapid performance (Harvey, Hendrick, Coll, & Nicholson, 2007). Breast MRI artifacts are brought about by technical factors and patients. These artifacts sometimes imitate pathology, and in other times obscure pathology, thus reducing the effectiveness of the diagnosis (Hendrick, 2007). MRI complexities can make the recognition of artifacts more difficult making it harder to understand than presented in mammography. If the artifacts are recognized, there is a possibility that they reduced or eliminated for patients of the future (Hendrick, 2007). The following are some of the common image associated with Breast MRI and the strategies used to minimize them.
Find out more information on how you can easily purchase an essay online
Ghost artifacts are said to be patient-induced due to the motion of the signal-producing tissues in the process of data acquisition. This not only produces blurred images, but also ghostly bright structures in motion (Hendrick, 2007). These artifacts appear as noise with structured noise that propagates in the phase-encoding (PE) direction, despite the bright tissues motion direction as shown in fig 1. Breast immobilization is one way to minimize patient’s motion and the resulting artifacts during MRI (Hendrick, 2007). This process is important because the breasts move either due to the patient respiratory or due to the patient shifting position randomly or seeking comfort.
In figure 1, part A is a transaxial image lacking fat-saturation that depicts the motion ghosting of fat in and out of the breast. In this case phase encoding is left-to-right. Part B shows a signal flare in the tissue near the receiver coil near the breast wall represented in short arrows. Part C shows an enhanced lesion of the same patient, the long arrow shows the invasive ductal carcinoma.
- i) Breast coils are used to combat these movements though this sometimes may lead to signal flaring where the breast tissues are adjacent to the elements of the receiver coil (Hendrick, 2007). Breast coils with adjustable compression plates are useful when immobilizing breasts during the scanning process. They reduce patient’s movement and are more helpful when the scan is occurring in a sagittal plane as ML or mediolateral reduces the slice number needed in each breast (Hendrick, 2007). Nonetheless, breasts that are smaller than the coil volume are immobilized using foam or cloths pad to fill the unoccupied portions. It is therefore ideal for a facility to have a set of foam pad that support and immobilize any breast without deforming the shape of the breast. The immobilization process reduces motion artifacts causing the movement of the breasts.
- ii) Using spatial presaturation reduces swallowing and breathing artifacts and pulse from the arteries.
- iii) Cardiac gaiting and flow compensation can reduce also reduce arterial and CSF pulsation.
This is also known as ‘wraparound’ or ‘image wrap’ an artifact that is apparent when the signal-producing tissues exceed the prescribed field-of-view of PE and FE direction. This artifact occurs because MRI needs a compilation of distinct number of signal values that create an image; this is to say a fixed number of PE and FE steps (Hendrick, 2007). With a discrete number of pixels interprets that the construction of both 2D and 3D cannot properly disclose the difference between tissues that produce signals within the scan field-of-view (FOV) and the corresponding signal producing tissues on the outer of the scan FOV. Normally this causes the addition of signals from the outer FOV to the inner signals from pixels of the FOV originating from the opposite side of the image (Hendrick, 2007). These artifacts add structured noise that makes it to difficult to understand the details in the breast, irregularly simulating pathology. A common phenomenon in PE direction than in the FE direction due to the fact manufacturers of MR implements techniques that wrap images in FE direction. This mainly occurs in the end slices of 3D acquisitions (Ballinger, 2012).
- i) Decreasing resolutions by increasing the FOV
- ii) FOV outside signals can be blocked by using a surface coil
- iii) Phase steps increase in the phase-encoded direction and oversampling the data in the frequency direction (Ballinger, 2012).
- iv) Frequency direction and the phase are swapped to ensure that the phase is in the narrow direction.
Also known as ‘Gibbs’, ‘edge’ or ‘ringing‘ this artifact occurs as a result of finite sampling that is predominant in MRI. Truncation artifacts normally happen to sharp and adjacent to high-contract interfaces (Hendrick, 2007). A discrete number of samples in many instances tend to overshoot the true signals changes on the crossways of sharp interfaces producing ringing artifacts away from the interface. This is because of finite sampling of the image on both in-plane directions. Gibbs phenomenon is the name given to this overshooting enhancing interface contrast and ring associated with it (Bolan, Nelson, Yee, & Garwood, 2005). Therefore, the light and dark banding occurrence adjacent to the sharp interfaces and falling off as the distance from the interface progressively increases.
Part A is 256 X256 matrix of ACR MRI Phantom grid section imaged using a T1W sequence. In this case mild truncation artifacts are present that are visible space light and dark lines on the grid. Part B shows the same section with a 128 by 192 matrix making the traction visible than in A. Part C shows visible truncation artifacts in the breast superior edge shown as parallel dark and light bands.
To correct the ‘Gibbs,’ more encoding steps can be thus lessening the intensity and narrowing the artifact. Using a higher matrix while increasing scan time can increase the matrix size in the PE direction (Hendrick, 2007). To decrease the truncation artifacts one can use a higher matrix while at the same time increasing the scan time to raise the size of the matrix in the PE direction.
Chemical Shift Artifacts
This artifact occurs due to the hydrogen nuclei in the fat resonant at rather different frequency than water hydrogen nuclei. Compared to fat, water has higher resonant frequency of 3.35 parts per million amounting to 214 Hz at 1.5T leading to two different variety of chemical shift artifacts (Hendrick, 2007). The first kind of chemical shift occurs in all MRI images caused by gradient frequency- encoding application in the process of signal measurement. Because the gradient of FE magnetic field is applicable in signal readout, the location of the shift of the ‘fat image’ relative to the ‘water image’ contains both tissues along frequency direction encoding in any MRI image (Hendrick, 2007). At 1.5T magnetic field strength, the amount of position shifts in terms of pixels largely depends only on the frequency shift from one pixel to the other, referred as pixel bandwidth. The resonant frequency variance between fat and water of 214 Hz amounts to position a shift of 1.8 pixels between the image of water and fat (Hendrick, 2007). This shift occurs between two images occurring in the FE direction.
The image noise is reduced by decreasing the bandwidth of the image – decreasing the pixel bandwidth with a number of frequency encoding. Spreading the shift over more pixels gives unwelcomed effect on the chemical shift artifacts (Bartella & Huang, 2007). For instance, BW can be decreased by the factor of 7, through from 122Hz/Pixel to 17.4 Hz/pixel. In this case, the water and fat chemical shift is reducible by a factor of 7 from 2 to more than 12 pixel as shown in figure 3 part C and D.
Radio Frequency Transmission Artifacts
RF transmission artifacts typically happen due to the radio frequency incomplete shielding in the MRI exam room. Faraday cage provide RF shielding that consist of a wire mesh shell that is overlaid on walls, doors, and windows in a MRI scan room. If there is a leak in the RF shield near the Larmor frequency, the discrete frequencies appear as distinct line of the image (Harvey, Hendrick, Coll, & Nicholson, 2007). The RF transmission artifacts are normally fixed along the frequency-encoding but in the PE direction they appear smeared because of the different measured amplitudes in the varied PE views. Other sources of these artifacts include faulty lighting fittings and the equipment for patient monitoring within the MRI scan room (Copeland & Bland, 2009). The poor fixtures fabricate broad-spectrum RF interferences that normally appear as one or more broad lines that run across the image in the PE direction.
To isolate these artifacts, one can acquire phantom images when the lights are turned on and then off (Harvey, Hendrick, Coll, & Nicholson, 2007). Or turning on the peripheral electrical equipment situated in the scan room then off until the artifacts are eliminated. When these artifacts persist, MRI system service engineering should be contacted to investigate RF shielding integrity.
Clinical Usefulness of Spectroscopy in Breast MRI
Tissues biologically contain important molecules (biochemical) in addition to water and fat protons. The information from the biochemicals are useful for specific diagnosis together with MRI anatomic details. Furthermore, with in-vivo MR spectroscopy metabolic information from volume element (Voxel) is possible (Janannathan, 2011). Using in-vivo MRS information that calls for biopsies are acquirable in a noninvasive manner. Proton (1H) and phosphorous (31P) nucleus due to their high natural abundance and sensitivity are used to report literature on breast (Tabar & Dean, 2000).
For this reason, spectroscopy is used when testing for the presence of any chemical compounds. In suspect lesions, spectroscopy measures any amount of metabolite called choline. Choline elevated levels are a strong indicator of malignancy or cancer. Neither do MRS nor MRI use X-rays that both traditional mammography and CT breast scans (Moy & Mercado, 2010). MRI and MRS can be completed in one sitting. MRI is now becoming a widely used technique for detecting breast cancer, paving ways for combining the MRI and MRS concept. There are clear benefits of recombining the two concepts though it comes with limitations such as the high costs. Not all types of breast cancers demonstrate elevated levels of choline and at least one form of breast cancer does not signal any level of choline (News Medical, 2012). Both in vivo MRS are combined to obtain information on the breast lesions chemical contents. The acquired information is used for clinical applications such as the response to the cancer therapies monitoring and the improvement of the lesion diagnosis accuracy (Bolan, Nelson, Yee, & Garwood, 2005).
Figure 5 illustrates a 3D dynamic imaging high resolution, showing the right breast with breast cancer. Spectroscopy proved quite effective in invasive carcinoma detection that is clearly shown in figure 6.
Monitoring Response to Treatment
Perhaps the most promising application of using breast MRS is to forecast the response to cancer treatment response to neoadjuvent chemotherapy. The readily available clinical methods like palpation and imaging highly rely on the changes in tumor size that can take several weeks before these changes become detectable. On the contrary, breast MRS detects the intracellular metabolism changes that occur before gross morphological change (Bolan, Nelson, Yee, & Garwood, 2005). The direct effects of chemotherapeutic agents are easily evaluated using MRS.
The recent improvements in MRS includes the use of software in the reduction of large lipids signals that brought the possibility of obtaining stable spectra of even 1.5T MR equipment. When the Cho quantity of the tumor is further assessed, breast MRS can elucidate the biology of breast cancer.
Magnetic Resonance Elastography (MRE)
MRE is a technology that is rapidly advancing for quantitatively assessing tissue mechanical properties (Mariappan, Glaser, & Ehman, 2011). This is a technology considered as imaging-based counterpart to palpation. This is commonly used by physicians when diagnosing and characterizing diseases. As a diagnostic method, the success of palpation that is based on the tissues’ mechanical properties that dramatically affects the presence of the process of disease such as cancer, fibrosis, and inflammation. MRE obtains information on the tissue stiffness through the ascertainment of mechanical waves through the tissue with the use of technique of MRI. Essentially the technique involves three steps: (i) Generation of tissue shear waves, (ii) Acquisition MR images to depict the induced shear waves’ propagation, (iii) Shear waves’ image processing in the generation of quantitative maps of tissues stiffness also known as elastograms (Mariappan, Glaser, & Ehman, 2011).
Figure 7 shows a drawing of a device that is used in the breast MR elastography. The device integrates electromechanical drivers with radiofrequency coil unit generating breast tissue acoustic shear waves using the contact plates on the breast lateral and medial aspects.
Figure 7 shows surgical specimen breast of a 55-year-old women suffering with invasive carcinoma. This is ordinary MR image with low-intensity tumor mass.
Figure 8 on the other hand shows the breast surgical specimen of the same 55- year-old woman as above. But the MR elastogram delineates the harder tissue (in red) with the soft tissues surrounding the breast.
Breast MRI Elastography Adds Information
Breast MR Elastography is MRE is another application investigated with great interest on breast cancer assessment. Breast tumors are stiffer than normal breast tissue or lesions. As a recommended part of the screening routine of the breast cancer, manual palpation helps in the detection of these hard masses (Mariappan, Glaser, & Ehman, 2011). For sensitively detection of tumor modules, Contrast-enhanced MR imaging (CE-MRI) has proven to be a technique to tackle the problem though this may lead to many false positives. CE-MRI provides additional information on these suspicious regions at the same time promising on diagnostic specificity in the future (Warren & Coulthard, 2001).
Ballinger, R. (2012). MRI Artifacts – MRI Tutor. Retrieved September 23, 2012, from Mritutor: www.mritutor.org/lectures/artifacts508.ppt
Bartella, L., & Huang, W. (2007). Proton (H1) MR Spectroscopy of the Breast. Radiographics , Vol 27, P 241-252.
Bolan, P. J., Nelson, M. T., Yee, D., & Garwood, M. (2005). Imaging in breast cancer: Magnetic resonance spectroscopy. Breast Cancer Res , Vol 7(4): 149-152.
Copeland, E., & Bland, K. (2009). The Breast: Comprehensive management of benign and malignant diseases. Philadelphia: PA: Saunders/Elsevier.
Harvey, J. A., Hendrick, E., Coll, J. M., & Nicholson, B. T. (2007). Breast MR Imaging Artifacts: How to Recognize and Fix Them. The Journal of continuing medical eduction in radiology , Vol 32 (5), p 131-145.
Hendrick, R. E. (2007). Breast MRI: Fundamentals and Technical Aspects. Oklahoma: Springer.
Janannathan, N. (2011). Breast tissue characterisation by in-vivo Magnetic Resonance Spectroscopy (MRS). Spectroscopy , 25 (2011), p 251 – 260.
L, T., & Dean, P. (2000). Teaching Atlas of Mammography. Stuttgart: Georg Thieme Verlag.
Mariappan, Y. K., Glaser, K. J., & Ehman, R. L. (2011, July). MAGNETIC RESONANCE ELASTOGRAPHY: A REVIEW. NIHPA Author Manuscript , Vol 23(5): 497–511.
Moy, L., & Mercado, C. L. (2010). Breast MRI, An Issue of Magnetic Resonance Imaging Clinics. Amsterdam: Elsevier Health Sciences.
News Medical. (2012). Magnetic resonance spectroscopy for breast cancer. Retrieved September 18, 2012, from News medical : http://www.news-medical.net/news/2007/07/11/27434.aspx
W.A., B., & Birdwell, R. G. (2008). Diagnostic Imaging: Breast. Canada: Amirsys.
Warren, R., & Coulthard, A. (2001). Breast MRI in Practice. New York:NY: Taylor & Francis Publishing .