Magnetic resonance (MR) imaging (MRI) is a noninvasive technique that uses the interaction between the magnetic properties of nuclei and radio waves to portray the structure of biological tissues (Damadian, 1971; Lauterbur, 1973). When placed in an external magnetic field and exposed to radio waves of proper frequency, the hydrogen nuclei within body tissues resonate [i.e., can absorb energy from a tuned radio wave and then, after a delay (relaxation time) emit the energy back at the same frequency]. The energy radiated back by the resonating nuclei, typically hydrogen nuclei (protons) for imaging, is the signal that is used for generating the MR image. The intrinsic differences in hydrogen density between the various components of body tissues (e.g., fat, blood, glandular tissue, muscle, etc.), as well as differences in magnetic relaxation times from voxel to voxel, determine the contrast of the MR image. Since these differences in MR characteristics are greater than differences in electron density and atomic number, the two factors that determine x-ray contrast, there is potentially more contrast in an MR image than in its x-ray equivalent image. Moreover, MR creates planar images, like CT, rather than summation or projection images, like mammography, which further improves the visability of low-contrast lesions.
The spin-lattice (T1) and spin-spin (T2) relaxation times substantially influence the MR signal. They are dependent on such factors as temperature, viscosity, crystalline-lattice structure or other microstructure which affect subtle magnetic interactions within tissues. Relaxation times provide imaging parameters other than simple hydrogen density mapping; indeed they are capable of yielding physiologic as well as anatomic data. For example, conditions such as local hyperemia and necrosis will alter both the temperature and solid-liquid characteristics of tissue, thereby, changing the T1 and T2 values from those of surrounding normal tissues.
In vivo MRI of the breast has been possible since the advent of whole-body MR scanners in the early 1980s, but initial investigations with whole-body imaging coils achieved very limited success (El Yousef et al., 1983; Ross et al., 1982). Subsequently, the development of high-resolution surface imaging coils designed specifically for the breast, along with MR magnet, gradient, computer, and pulse sequence improvements, have resulted in superior breast images capable of demonstrating smaller lesions and finer structural detail (Alcorn et al., 1985; Bydder et al., 1985; Dash et al., 1986; El Yousef and Duchesneau, 1984; El Yousef et al., 1984; 1985; Hornak et al., 1986; Sinha et al., 1993; Stelling et al., 1985; Turner et al., 1988; Wiener et al., 1986; Wolfman et al., 1985). Up until 1986, MRI examinations were done using T1 and T2 weighted-spin echo pulse sequences. Subsequently, there was additional progress with the use of gradient echo sequences, such as fast low-angle shot and fast imaging with steady progression (Kaiser and Oppelt, 1987; Kaiser and Zeitler, 1989). More recently, pulse sequences that obtain three-dimensional blocks of data with fat suppression have been used, producing still further improvement in resolution, contrast, and overall lesion visualization (Harms et al., 1993a; 1993b; Pierce et al., 1991; Rubens et al., 1991).
Current clinical experience with breast MRI indicates its several strengths and weaknesses. Fatty and fibroglandular regions of the breast are clearly distinguished and areas of dense fibroglan-dular tissue are imaged with a greater range of contrast than with either mammography or CT scanning. Large and some small breast masses also are readily portrayed, especially, if surrounded by substantial amounts of fatty tissue with most cancers showing relatively irregular and ill-defined borders and benign lesions demonstrating more smooth and sharply-defined margins (El Yousef et al., 1984). However, even when using the best currently available surface coils, the spatial resolution of nonenhanced MRI is far inferior to that of mammography, so that the tiny clustered calcifications of DCIS and the fine spiculations characteristic of many invasive carcinomas are not imaged with MRI. The lack of inherent contrast difference between breast lesions and normal glandular tissue also makes it difficult for MRI to portray some of the smaller mammographically detected masses, particularly, those invasive cancers that present with vague, ill-defined margins, without the use of contrast agents. Unless new techniques can be developed, it is unlikely that nonenhanced MRI will achieve widespread use for either screening or diagnosis of breast cancer (Dash et al., 1986; Kopans, 1984; Pierce et al., 1991; Turner et al., 1988).
Rather, most investigators have utilized nonenhanced MRI to evaluate breast disease that already has been detected by mam-mography or physical examination. For this use, MRI does appear to be both sensitive and specific in the diagnosis of simple benign cysts, but no more so than the already established and far less expensive methods of aspiration and ultrasonography (Alcorn et al., 1985; Dash et al., 1986; El Yousef et al., 1985; Kaiser, 1990; Rubens et al., 1991; Stelling et al., 1985; Turner et al., 1988). For solid breast masses, nonenhanced MRI cannot reliably distinguish fibroadenomas and post-biopsy scars from malignancies on the basis of either morphological features or T1 and T2 values (Alcorn et al., 1985; Heywang et al., 1986a; Stelling et al., 1987; Turner et al., 1988).
These disappointing results have led to additional avenues of investigation, especially to the use of the paramagnetic metal ion chelate, gadolinium diethylene triamine penta-acetic acid (Gd-DTPA). This MRI-specific contrast agent serves as an indirect indicator of tissue perfusion, since it accumulates at a faster rate in more highly vascularized lesions than in normal tissues. Therefore, similar to results observed with CT scanning following iodide administration, many breast cancers also demonstrate differential enhancement after intravenous infusion of Gd-DTPA (Adler and Wahl, 1995; Dao et al., 1993; Harms et al., 1993a; 1993b; Heywang et al., 1986a; 1986b; 1989; Kaiser and Zeitler, 1989; Pierce et al., 1991; Revel et al., 1986; Rubens et al., 1991; Stack et al., 1990). Enhancement is found not only for invasive carcinomas, but also for some cases of DCIS that present mammographically only by virtue of clustered microcalcifications (Davis and McCarty, 1997; Gilles et al., 1993; 1995; Nunes et al., 1997a; 1997b; Orel et al., 1997). A particularly helpful aspect of contrast-enhanced MRI is that breast cancer displays much higher levels of enhancement than benign post-biopsy scar tissue, permitting differentiation often not possible by either mammography or physical examination (Dao et al., 1993; Gilles et al., 1993; Harms et al., 1993a; 1993b; Heywang et al., 1989; 1990).
Another potentially useful application of contrast enhanced MRI is in determining the extent of tumor for breast cancer patients who desire breast conservation therapy. MRI indeed is more sensitive than mammography in this regard, often identifying nonpalpable multifocal and multicentric tumor deposits not depicted at mammography (Harms et al., 1993a; 1993b; Orel and Schnall, 2001; Orel et al., 1994; 1995; Weinreb and Newstead, 1995). However, the utility of MRI in treatment planning is limited by its relatively low specificity and the lack of general availability of MR-guided localization devices to permit tissue diagnosis for lesions detected only at MRI (Adler and Wahl, 1995; Frankel and Sickles, 1997; Weinreb and Newstead, 1995).
Although most cancers exhibit considerable contrast enhancement, many fibroadenomas also demonstrate similar substantial amounts of enhancement. An interesting variation in contrast-
enhancement technique involves dynamic fast-sequence imaging. Some investigators suggest that this approach may permit differentiation between carcinoma and fibroadenoma, whereas cancers typically show rapid enhancement within the first 2 min after contrast injection, the enhancement seen in fibroadenomas progresses more slowly (Boetes et al., 1994; Hulka et al., 1995; Kaiser, 1990; Kaiser and Zeitler, 1989; Stack et al., 1990). Unfortunately, there are reports also indicating the unreliability of dynamic fast sequence contrast enhancement; some fibroadenomas and other benign conditions have been shown to enhance just as rapidly as do most cancers (Frankel and Sickles, 1997; Gilles et al., 1994; Harms et al., 1993a; 1993b; Orel et al., 1994).
As with the use of ultrasonography (Section 8.1.3), there also has been recent interest in using MRI to identify complications of breast implants, especially, leakage and rupture of gel-filled sili-cone implants (Brem et al., 1992; Gorczyca et al., 1992a; Harms et al., 1992; Schneider and Chan, 1993). The combination of a T2 weighted fast spin-echo technique, T2 weighted fast spin-echo technique with water suppression, and T1 weighted spin-echo technique with fat suppression or, alternatively, a modified three-point Dixon (1984) technique may reliably differentiate silicone from native breast tissues, thereby, permitting identification of both extracapsular and many intracapsular ruptures and leaks (Gorczyca et al., 1992a; 1992b; Schneider and Chan, 1993). The relative efficacy of MRI versus ultrasonography has not been determined. However, ultrasonography is more readily available and considerably less expensive, while MRI is less operator dependent and has the imaging advantage of being able to portray deep structures through the full-thickness of an implant. Indeed, MRI is more accurate than either mammography or ultrasonography in identifying implant disruption and is widely used as the ultimate imaging procedure to assess implant integrity when there is clinical suggestion of disruption, particularly when other imaging modalities do not provide clear supporting evidence (Berg et al., 1995; Gorczyca et al., 1994).
Ultimately, the major promise of MRI in the diagnosis of breast disease is its potential to image the radiographically dense breast with uniquely high contrast, perhaps also permitting the differentiation of benign from malignant tissue on the basis of the physiological information transmitted via contrast enhancement techniques combined with new pulse sequences. Another advantage of MRI is that it does not involve ionizing radiation. Indeed, in its current clinical form, it causes no known genetic damage (Wolff et al., 1980) and there is no indication of other significant hazards (Budinger, 1979). Breast MRI is now an accepted diagnostic adjunct to mammography and breast ultrasound, especially in cancer staging to aid in determining extent of disease, and in evaluating cancer recurrence (Heywang-Kobrunner et al., 2001; Morris, 2002; Orel and Schnall, 2001). There is also mounting evidence that breast MRI is useful beyond mammogra-phy for evaluation of the contralateral breast in women with a known breast cancer (Fischer et al., 1999; Lee et al., 2003; Liber-man et al., 2003). Recent studies have found that breast MRI detects cancer in the contralateral breast in four to five percent of women with a known breast cancer, even after negative mammog-raphy and physical examination.
Several recent studies have found that MRI appears to be more sensitive than mammography among women with an inherited susceptibility to breast cancer (Kriege et al., 2004; Kuhl et al., 2000; Stoutjesdijk et al., 2001; Tilanus-Linthorst et al., 2000; Warner et al., 2001). Unlike screening mammography trials, none of these studies measured breast cancer mortality as an endpoint. In addition the design of these studies might have artificially increased the sensitivity of MRI with respect to mammography. Nevertheless, these studies do suggest that MRI screening may benefit women at extremely high risk for development of breast cancer. There are several reasons why MRI is not advised for screening all other women. These include need for intravenous contrast injection, an extremely high cost of equipment and examination, limited availability of equipment, and a high FP biopsy rate.
Was this article helpful?