Quality breast MRI requires much more than field strength

January 10, 2008

Use of dynamic contrast-enhanced MRI to evaluate the breast is rapidly increasing despite the lack of established standards or clearly defined indications and limited randomized clinical trials. The superior sensitivity of MRI for invasive breast cancer has led to the proliferation of its use. Its variable specificity, however, has limited its utility.

Use of dynamic contrast-enhanced MRI to evaluate the breast is rapidly increasing despite the lack of established standards or clearly defined indications and limited randomized clinical trials. The superior sensitivity of MRI for invasive breast cancer has led to the proliferation of its use. Its variable specificity, however, has limited its utility.

Recently published American Cancer Society guidelines recommend annual screening with MRI for women whose lifetime risk of breast cancer is more than 20%. This includes women with a known genetic mutation (BRCA1 or BRCA2) or a strong family history of breast or ovarian cancer and women who have received chest radiation to treat Hodgkin's lymphoma. It is estimated that 2% to 3% of women in the U.S. are BRCA1 or BRCA2 carriers, which would equate to approximately 1.4 million MRI breast examinations annually. Although the number of non-gene carrier women whose overall lifetime risk is greater than 20% is unknown and depends on the risk assessment model used, it is likely to be considerable.

The ACS recommendations lack the data needed to resolve whether women in other categories of increased risk, such as those with a personal history of cancer or a prior biopsy yielding atypical hyperplasia, should receive yearly MRI screening. Given recent media attention and public awareness, many women with less than a 20% lifetime risk will likely request and undergo breast MRI.

Use of MRI for preoperative evaluation in women with newly diagnosed breast cancer continues to grow as well. It has been firmly established that preoperative MRI can detect unsuspected disease and often changes management.1,2 Such use of preoperative MRI to alter surgical management has been criticized, however, because of the lack of studies evaluating its effect on tumor recurrence and mortality. Unsuspected disease detected on MRI prior to breast conserving therapy may be of little clinical consequence and controllable by radiation therapy.3-5

The argument against use of preoperative MRI is primarily based on the low 10-year local recurrence rates in patients with negative surgical margins. The number of women with positive margins at initial lumpectomy who require additional surgery, however, may be as high as 40%.6 By better defining tumor extent and providing 3D information, MRI has the ability to improve surgical outcomes and decrease positive margin rates.7 Preoperative MRI can better guide surgical and radiation therapy planning by demonstrating abnormal axillary adenopathy, internal mammary nodes, occult contralateral disease,8 and pectoralis or chest wall involvement. With the increasing utilization of partial breast irradiation, MRI may prove to be essential in identifying patients likely to fail this treatment due to the presence of occult disease outside of the local radiation field.

The many benefits of preoperative MRI, although individually small, together could potentially have a positive impact on a significant number of patients. Other indications for breast MRI include its use to evaluate response to chemotherapy in patients with known cancer, residual disease in patients with close or positive margins at initial lumpectomy, and scar versus recurrent disease. Given the expanding use of breast MRI and the current uncertainties of cost versus benefit in many cases, it is essential that programs strive for the highest level of quality to reduce false positives and patient anxiety and optimize clinical outcomes.

QUALITY MEASUREMENTS

What constitutes high-quality breast MRI? To answer this question, we must review the fundamental issues involved in its performance and implementation. First, accuracy (sensitivity and specificity) is both the foundation and the limiting factor in the performance of high-quality breast MRI. Second, accurate breast MRI interpretation may be for naught if quality is compromised from either poor integration into existing breast imaging services and patient management or a lack of concise reporting and outcomes tracking.

Accuracy of breast MRI relies on several key elements, including high-quality imaging protocols, an understanding of imaging processing and display, and reader experience.

  • High-quality imaging. A field strength of 1.5T is preferable to 1T because of its higher signal-to-noise ratio and faster acquisition times. Bilateral imaging is preferable; it is helpful in screening cases to be able to compare the opposite breast for symmetry as well as to evaluate the contralateral breast in patients with known cancer. Decisions such as whether to perform fat saturation, acquire in 2D or 3D, or scan in the axial, sagittal, or coronal plane may be largely vendor-dependent.

Regardless of technique, parameters should be optimized to achieve the common goal of high-quality images. High sensitivity, for the most part, requires appropriate administration of contrast and timing of postcontrast imaging as well as sufficient SNR to allow lesion detection (Figure 1). Typical contrast injections involve the administration of 0.1 mmol/kg of contrast at 2 cc/sec followed by a 20-cc saline flush. Because breast cancer enhancement tends to peak 60 to 120 seconds after contrast injection, postcontrast imaging must be timed to capture this interval, optimally at the center of k-space, where contrast information is the highest.

The primary determinant of breast MRI specificity is the assessment of lesion margins and internal architecture (morphology),9 and this is affected by technical factors relating to spatial resolution such as in-plane resolution (matrix) and slice thickness. Current high-quality breast MRI protocols achieve in-plane resolution of less than 1 mm and slice thickness of less than 3 mm. Specificity may be improved by incorporating the analysis of lesion enhancement patterns over time (kinetics).10 It has been shown that cancers tend to demonstrate rapid early enhancement and wash-out, whereas benign lesions tend to enhance progressively. Spatial and temporal resolution, however, always function as a trade-off (Figure 2).11 Accuracy may also be affected by patient positioning, motion, and artifacts.

The standard dynamic breast MR imaging protocol used at the University of California, San Diego is a 1.5T bilateral 3D gradient recall echo axial acquisition with fat saturation, 512 x 348 matrix, 2.6-mm slice thickness overlapped every 1.6 mm, 30-cm field-of-view, 88 slices, and acquisition time of one minute. A precontrast sequence is followed by five postcontrast scans. Contrast administration finishes 10 seconds prior to the first postcontrast scan. This places the 90-second peak 20 seconds into the second postcontrast scan, which corresponds to the center of k-space. Precontrast right and left T2-weighted spin-echo sequences and a delayed postcontrast high-resolution six-minute bilateral 3D GRE sequence are obtained as well.

The use of higher field strength, such as 3T, may eventually further improve both spatial and temporal resolution as well as enhance MR spectroscopy. There is currently little evidence, however, regarding improved accuracy or added clinical value. It should not be assumed that 3T MRI equates to high-quality breast MRI. Three-T systems still require protocol optimization that may be highly vendor-specific. Further research and coil development are needed.

  • Understanding imaging processing and display. Accurate interpretation requires an understanding of how the MR images being interpreted are processed and the limitations of the tools being used to read them. Because the use of image subtraction (postcontrast minus precontrast) facilitates interpretation by increasing lesion conspicuity, most MRI interpretation involves the reading of subtracted images. But patient motion may significantly degrade the quality of subtracted images through misregistration. It is therefore important to recognize when subtracted images have been affected by patient motion, and this should prompt review of the unsubtracted original images (Figure 3).

The use of computer-aided detection programs to assist in interpreting breast MRI offers several advantages, including registration correction (Figure 4) and quantification of lesion kinetics. Visual estimation of lesion kinetics on an imaging workstation, in general, is not very accurate, as many lesions display mixed kinetics and the accumulation of background parenchymal enhancement may be confusing.

If kinetics will alter lesion categorization and management in some cases, it is important to accurately establish the true kinetics (Figure 5). MR CAD programs can facilitate quantitative kinetic analysis. As with the CAD programs used in mammography, these tools have limitations and must be used appropriately. Simply reading cases based only on the color coding produced by CAD programs may lead to false positives due to artifacts and incorrect lesion characterization as well as to potentially missed cancers whose peak enhancement is below the set threshold of the program. Additional features such as multiplanar reconstruction and volume rendering may also assist in lesion analysis.

  • Reader experience. There is no replacement for exposure to actual supervised case interpretation in gaining experience in breast MRI. Residency training, postgraduate breast imaging fellowships, visiting fellowships, and continuing medical education case review courses are vital in the training process.

Radiologists interpreting breast MRI need to be experienced in breast imaging. Prior breast imaging studies such as mammography and ultrasound must be correlated with the MRI findings, along with patient history, to provide the most accurate interpretation and appropriate recommendation. Likewise, breast imaging radiologists must be familiar with cross-sectional imaging, thoracic anatomy, and common findings encountered outside the breast on MRI. Currently, MRI should not be used to override suspicious findings on mammography, ultrasound, or physical exam.

Quality breast MRI also requires concise reporting and appropriate integration into existing breast imaging services and patient management protocols. The use of the American College of Radiology Breast Imaging Reporting and Data System (BI-RADS) Lexicon for MR has helped to standardize lesion categorization and management recommendations.

Incorporation of breast MRI reporting into established mammography reporting and tracking systems is extremely useful. Particularly for preoperative MRI, delays in final treatment due to second-look ultrasound scans and additional biopsies must be minimized through prompt reporting, communication, and scheduling.

FUTURE CHALLENGES

Despite the increasing utilization of breast MRI, quality remains variable. The development of the ACR accreditation process for breast MRI, as well as practice guidelines and new technical standards for the procedure, will help to improve quality. Further clinical research and expanding education will increase radiologists' skills in breast MRI interpretation.

Challenges still remain, however. Indications for breast MRI need to be more firmly established so that referring physicians, radiologists, and patients may make a more informed decision regarding its utilization. As imaging for breast cancer screening continues to evolve into a more tailored process based on breast density, family history, and personal history, the assessment of risk is becoming more and more critical. New questions arise regarding which risk models are most accurate and who should provide risk assessment: referring physicians, genetic counselors, or radiologists. Further clinical studies and refinements in MR technology will, undoubtedly, continue to improve the quality and benefit from breast MRI.

Dr. Comstock is director of breast imaging at the University of California, San Diego.

References

1. Bedrosian I, Mick R, Orel SG, et al. Changes in the surgical management of patients with breast carcinoma based on preoperative magnetic resonance imaging. Cancer 2003;98:468-473.

2. Braun M, Polcher M, Schrading S, et al. Influence of preoperative MRI on the surgical management of patients with operable breast cancer. Breast Cancer Res Treat. 2007 Sep 29 [Epub ahead of print].

3. Morrow M. Magnetic resonance imaging in the preoperative evaluation of breast cancer: primum non nocere. J Am Coll Surg 2004;198:240-241.

4. Morrow M. Magnetic resonance imaging in breast cancer: one step forward, two steps back? Jama 2004;292:2779-2780.

5. Morrow M. Magnetic resonance imaging in breast cancer: Is seeing always believing? Eur J Cancer 2005;41:1368-1369.

6. Smitt MC, Horst K. Association of clinical and pathologic variables with lumpectomy surgical margin status after preoperative diagnosis or excisional biopsy of invasive breast cancer. Ann Surg Oncol 2007:14(3):1040-1044.

7. Comstock CE, Hunt PT, Middleton MS. Effect of pre-operative MRI on mastectomy rates, lumpectomy negative margin rates and time to surgery in women with known breast cancer. RSNA annual meeting, 2007, abstract.

8. Lehman CD, Gatsonis C, Kuhl CK, et al. ACRIN Trial 6667 Investigators Group. MRI evaluation of the contralateral breast in women with recently diagnosed breast cancer. NEJM 2007;356(13):1295-1303 [Epub 2007 Mar 28].

9. Nunes LW. Architectural-based interpretations of breast MR imaging. Magn Reson Imaging Clin N Am 2001;9(2):303-320, vi. Review.

10. Kuhl CK, Mielcareck P, Klaschik S, et al. Dynamic breast MR imaging: are signal intensity time course data useful for differential diagnosis of enhancing lesions? Radiology 1999;211(1):101-110.

11. Kuhl CK, Schild HH, Morakkabati N. Dynamic bilateral contrast-enhanced MR imaging of the breast: trade-off between spatial and temporal resolution. Radiology 2005;236(3):789-800.

12. Comstock CE, Middleton MS, Lundell AL. Accuracy of visual estimates of lesion kinetics during interpretation of dynamic breast MR images on a workstation. RSNA annual meeting, 2005, abstract.