Applied Radiology

Dr. Kuzmiak is an Assistant Clinical Professor and Dr. Pisano is a Professor in the Department of Radiology, University of North Carolina School of Medicine, Chapel Hill, NC.

Magnetic resonance imaging (MRI) of the breast has been in use since the mid-1980s. Primarily, it has been used for evaluating implant rupture and for searching for a primary breast cancer in cases in which the mammogram and physical examination were negative in patients with metastatic disease to axillary lymph nodes. Since MRI allows for improved soft-tissue contrast, enhancement of a lesion, and multiplanar imaging, investigations continue to evaluate its usefulness and role in breast cancer detection.

Breast cancer is often similar in density to the surrounding normal dense breast tissue; therefore, 10% to 20% of cancers are not visualized by conventional film-screen mammography. ¹ Mammography is the only imaging study that has been shown to reduce breast cancer mortality. ² However, mammography has limited sensitivity and specificity in dense breasts, postsurgical/postradiation breasts, and augmented breasts. Recent studies have demonstrated that the sensitivity of breast MRI in lesion detection is as high as 100%; however, breast MRI has a low specificity, ranging from 37% to 86%. ^3,4

To detect malignant breast lesions, an intravenous contrast agent, gadolinium-DTPA, must be administered to the patient. Lesion enhancement is based on the phenomenon of angiogenesis (increased tumor vascularity that is disorganized and has increased permeability). The contrast agent also causes T-1 shortening, which allows for increased conspicuity of the lesion. Therefore, lesions would be expected to enhance rapidly and strongly. Benign lesions such as fibroadenomas, fibrocystic changes, and inflammation can also be vascular and can demonstrate rapid and strong enhancement. Consequently, if a lesion or region demonstrates these findings, malignancy cannot be excluded, which produces a high false-positive rate.

False negatives have also been reported in the literature, but appear to be less common. False negatives have been reported with ductal carcinoma in situ (DCIS), and slowly enhancing lesions, such as invasive lobular and mucinous carcinomas. ^5-7

Imaging protocols

Prior to imaging the patient, there are several imaging variables that need to be considered. First, since there is no universally accepted breast MRI technique, what type of magnetic field strength is your institution going to use? When using a lower field strength, such as 0.5 Tesla, the radiologist needs to be aware that low-field protocols create several issues, including a lower signal-to-noise ratio (SNR), possible problems with chemical shift dispersion, and less T1 tissue variability, which offers less enhancement. Most of the MRI breast literature is based on a 1.5 Tesla field strength.

Second, since there is no consensus on which imaging sequences to use, the radiologist must decide how to determine a balance between spatial and temporal resolution. Also, the slice thickness and interslice gap/no gap between the slices must be chosen, since not all breast cancers are >1 cm in size or have a focal growth pattern. Currently, fat-suppressed T1-weighted, three-dimensional (3D) volumetric images seem to be sensitive to contrast and to allow rapid acquisition so that enhancing lesions can be identified before the enhancement of normal breast tissue.

Third, to increase lesion detection against a fatty background, image subtraction or fat saturation is required. One method for accomplishing this is to subtract the precontrast images from the postcontrast images. This makes it imperative that there is no patient motion between the precontrast and postcontrast images. Alternatively, the fat signal can be eliminated by a variety of fat-suppression techniques. ³ In order to report enhancement relative to the rest of the tissue (minimal, mild, etc.) and for biopsy guidance, the radiologist should be comfortable with viewing fat saturation images.

For breast MR imaging, a dedicated breast coil is required to improve the signal-to-noise ratio. Unilateral and bilateral coils are available. To decrease imaging time, several institutions use a compression plate to decrease the thickness of the breast and reduce respiratory motion. The patient is imaged in a prone position.

The study can be optimized further if the patient has minimal hormone stimulation that can be obtained by imaging the patient in the second week of the menstrual cycle or by stopping hormone replacement therapy for several weeks. ⁷ It is also very important to know the patient's clinical history, including past surgeries, and to have reviewed the patient's recent mammogram and/or ultrasound before viewing the MR images.

Use of contrast media

Breast cancers enhance after intravenous contrast administration (Figures 1 and 2). Numerous authors have evaluated the kinetics or dynamic measurements of contrast medium uptake in breast masses. By doing so, they are attempting to differentiate malignant from benign lesions based on how rapid the initial uptake and washout of the contrast, which is plotted as a kinetic curve. Breast cancers tend to have rapid enhancement and washout while benign lesions tend to have slow progressive enhancement. ^7,8 Heywang-Kobrunner et al ⁷ have seen rapid enhancement with slow washout, as well as less rapid enhancement in breast cancers. Both of these types of enhancement have also been seen at our institution. In addition, benign lesions can also have these types of kinetic curves. ⁷

Since there is such an overlap with kinetic analysis of malignant and benign lesions, Heywang-Kobrunner et al ⁵ and Nunes et al ⁹ have been evaluating the morphologic features of a breast mass after contrast administration. Currently, these researchers have demonstrated higher positive predictive values for malignancy for a spiculated enhancing mass and for an irregular mass with peripheral enhancement. The same appears to be true with ductal enhancement. Also, a mass that was smooth or gently lobulated in contour and had no enhancement or minimal enhancement was most likely found to be benign. ⁹

Indications for breast MRI

Currently, there is not enough data available to justify routine performance of MRI on all patients due to its high cost and lack of well-established accuracy. Research into its utility continues in multicenter studies such as the one led by Sylvia Heywang-Kobrunner, MD at the University of Halle, Germany, and Mitchell Schnall, MD at the University of Pennsylvania, Philadelphia, PA, with the latter results not yet published. MRI of the breast may find its niche in evaluating patients who are recommended for biopsy to evaluate for multicentric or multifocal disease. It may change the patient's staging and the type of surgical procedure selected, such as mastectomy instead of breast conservation surgery. In the literature, it has been reported that up to one-third of patients have other foci of disease than initially suspected. ^3,10

Others who may benefit from breast MRI are those who have undergone breast conservation surgery. Sometimes it is difficult with mammography to determine if the patient has a breast cancer recurrence versus postsurgical changes at the lumpectomy site. With the increased sensitivity of MRI and the use of contrast administration, a malignant lesion would enhance while scar tissue should not. However, scar tissue that is <18 months of age may enhance and create a false-positive result. ⁷ In addition, to evaluate for recurrence, MRI can be helpful in the immediate postoperative period to assess for residual disease.

MRI may be useful in women who are at high risk of developing breast cancer and have dense breasts. Recent reports from Germany assessing this population are promising so far. ¹¹ However, there may be more false-positive results in this subpopulation due to the increased sensitivity of MRI compared with conventional mammography. This patient subgroup tends to be younger. Brown et al ¹² have shown that in a younger population group with increased breast density there are usually more incidental enhancing lesions, and management of these incidental lesions is problematic. At our institution, we have evaluated incidental enhancing lesions with additional mammography and ultrasound, and we have been able to find a lesion(s) retrospectively using these modalities. An appropriate work-up of the lesion(s) is then performed, including biopsy if necessary. Additionally, some institutions do a short-term follow-up MRI for incidental enhancing lesions.

If a suspicious lesion is detected only with MRI, then the lesion can be biopsied or localized for surgical excision only with MRI guidance. It is essential that an institution that provides breast MRI to its patients have the capability to perform the appropriate procedure for the area(s) of concern. MRI-guided procedures are performed with the patient in the prone position and with the patient's breast in a compression system with an open or perforated lateral side plate. All of these procedures must be performed with the introduction of the needle from a lateral approach. After preoperative needle localization, one cannot give the specimen contrast or image it with MRI to look for the original abnormality. At our institution, a specimen radiograph confirms that the localization wire has been removed and determines if an abnormality can be visualized to mark for the pathology department.

Conclusion

Compared with mammography, breast MRI has the advantage of improved tissue contrast and information about vascularity and enhancement of a lesion. However, it has the limitations of greater expense, lack of universal imaging parameters, the low risk associated with the administration of intravenous contrast, and the need for MRI-compatible biopsy or localization systems. Despite these disadvantages, preliminary research demonstrates that MRI of the breast is a promising imaging tool. It appears to be most promising for the detection and staging of breast cancer in patients with a known suspicious lesion, identifying a mammographically occult lesion in patients with axillary lymph nodes positive for adenocarcinoma, distinguishing scar tissue from cancer recurrence, and for the evaluation of residual disease. Further research is still needed to determine if MRI technology can provide valuable clinical information that cannot be obtained with conventional breast imaging methods. AR