As a diagnostic breast radiologist, my focus is to detect breast cancers at their earliest stages so I can provide my patients with the best outcome possible. Over the years, we have seen an evolution in breast imaging. Screening mammography has been able to detect breast cancer in many women who otherwise have no signs or symptoms. Digital mammography and ultrasound are helping to better diagnose breast cancer in its earliest stages. However, no technology is perfect. Mammography sensitivity is only about 50 percent and decreases to 30 percent in women with dense breasts.
Advanced technologies such as magnetic resonance imaging (MRI) are becoming more widely used for both diagnostic and screening purposes. Since 2007, when the American Cancer Society began recommending that women at high risk for developing breast cancer be screened with MRI, we have seen a dramatic increase in the volume of breast MRIs.
As with any technology, MRI has its limitations. The generally accepted belief is that MRI cannot visualize the earliest form of breast cancer, ductal carcinoma in situ (DCIS). DCIS accounts for 40 percent of all screened breast cancers. Yet, DCIS has been the most frequent false negative histology in many of the high-risk screening trials.
In general, the accepted standard for the detection of DCIS has been mammography. Some studies have reported lower detection for DCIS on MRI than mammography. Low-resolution breast MRI is effective at finding high-grade DCIS, but will often miss low-grade DCIS. High-resolution breast MRI can reliably visualize low- and high-grade DCIS. The ability to characterize both low-grade and high-grade DCIS remains critically important, as the tumors are often heterogeneous and depiction of the entire tumor is needed for accurate staging.
The inability to distinguish DCIS from benign breast disease increases the need for biopsy. Many benign lesions can mimic non-mass-like enhancement seen with DCIS. For this reason, the positive predictive value of MRI for non-mass-like lesions is very low in most MRI trials.
Better characterization of DCIS with depiction of microanatomy is expected to be more specific than the general non-mass-like enhancement feature and lead to more accurate interpretations and determination of disease extent. If MRI can reliably visualize the microanatomy of DCIS, the potential for finding cancers early, when they are most treatable, could result in a notable reduction in cost to the healthcare system and stress on patients. In addition, this would be a valuable prognostic tool for surveying women at elevated risk for subsequent disease.
I have spent many years developing a dedicated breast MRI system, in conjunction with Aurora Imaging Technology Inc. The goal has been to push the limits of MRI and optimize an MRI system to image breast tissue. Recently, the development of AuroraEDGE has provided exceptional resolution and morphologic detail (700 micron), integral factors in visualizing and characterizing DCIS. The Aurora dedicated breast MRI system with AuroraEDGE technology utilizes isotropic voxels, allowing image reformation in all planes with equal resolution (700 micron). Any plane can be calculated using original scan data. This eliminates volume averaging, which limits visualization of DCIS that is surrounded by normal breast tissue.
Pre-contrast Aurora scans have the same T1 weighting as the post-contrast images, but also have fluid weighting. When the pre- and post-contrast images are subtracted, it results in a positive scale for enhancing lesions and a negative scale for fluid. The ability to characteristically see tumor and fluid on the same image allows the visualization of microscopic anatomy needed for accurately defining DCIS (patent-pending).
I have been employing this technology at my practice for slightly more than two years now and have seen the clinical benefit it provides. In a recent evaluation of clinical cases, the Aurora System detected 19 of 20 (95 percent) pathology-proven DCIS cases, nearly half of which were intermediate or low-grade. Table 1 shows the details. In addition, two case studies can be seen in the sidebars and images at left and above.
The ability to reliably and reproducibly show all grades of DCIS with breast MRI is undoubtedly a significant advantage in the management and surveillance of breast cancer patients. Early clinical experience utilizing AuroraEDGE technology has shown promise in this area. The high resolution and morphologic detail achieved with the Aurora dedicated breast MRI system may offer advantages in distinguishing DCIS from benign atypical proliferative processes.
Broader study is needed to further explore and analyze the performance characteristics and overall clinical benefit. It is hoped that if we continue to push the limits of breast MRI and optimize its capabilities to image the breast, we will be able to diagnose cancers at their earliest and most treatable stages.
Steven E. Harms, M.D., FACR, is recognized nationally in the area of magnetic resonance imaging (MRI), specifically breast MRI, and was named “Komen Foundation Scientist of the Year” by Susan G. Komen for the Cure in 1998. He was formerly on the faculty at the University of Arkansas for Medical Sciences (UAMS), Baylor University Medical Center and M. D. Anderson Cancer Institute. Harms is a fellow of the Breast Imaging Society and practices at MANA’s The Breast Center in Fayetteville, Ark., and continues to be a clinical professor at UAMS. He also serves as the medical director for Aurora Imaging Technology Inc.
SIDEBAR: Additional Solutions for Breast MRI
Invivo Corp. offers several products to enhance breast magnetic resonance imaging (MRI).
DynaCAD software for breast features several automatic image processing tools, including: multi-vendor/modality viewing capabilities; QuickClick segmentation tool for quick lesion classification; image registration and current prior analysis; interventional planning compatible with various MR coils and vacuum assist biopsy vendors, and optional Oncad morphology detection system.
Oncad is the only morphology detection computer-aided detection (CAD) for MRI. It is automatically processed as a hanging protocol, ready to read.
Invivo also offers Luminescence breast coils (shown above). Features include: high-resolution isotropic sub-millimeter imaging; fast imaging speed with 16 diagnostic channel elements; padded arm rests, headrest and coil ramp for comfort; large aperture volume and access for interventions with seven-channel coil, and integrated lighting for better visibility during interventional procedures.
For more, visit www.invivocorp.com
— Helen Kuhl