Digital Tomosynthesis of the Breast

OBJECTIVES:

"Will tomosynthesis replace conventional mammography?" This is a provocative question and one that we must be consider. Dr. Daniel Kopans, director of breast imaging at Massachusetts General Hospital described positive experiences with tomosynthesis in 400 volunteers at MGH.

As new technologies emerge and are tested the breast imager must learn change and adapt or be left behind.

What is Tomosynthesis?

Digital tomosynthesis creates a 3-dimensional picture of the breast using X-rays. Currently, digital tomosynthesis is available only for research purposes. Digital tomosynthesis of the breast is different from a standard mammogram in the same way a CT of the chest is different from a standard chest x-ray. One is 3-dimensional; the other is flat. Tomosynthesis is a 3D digital technique that removes the effect of superimposed tissue. The principle is to reduce noise of overlapping normal breast tissue, improving detection of breast cancer.

"Looking at a 2D mammogram is like trying to look through a book with clear pages-you can see the letters but it is very hard to read them. It is much better if you can read each page individually," Kopans said.

Digital tomosynthesis takes multiple X-ray pictures of each breast from many angles. The breast is positioned the same way it is in a conventional mammogram, but only a little pressure is applied—just enough to keep the breast in a stable position during the procedure. The X-ray tube moves in an arc around the breast while 11 images are taken during a seven-second examination. Then the information is sent to a computer, where it is assembled to produce clear, highly focused 3-dimensional images throughout the breast. Early results with digital tomosynthesis are promising. Researchers believe that this new breast imaging technique will make breast cancers easier to see in dense breast tissue.

Made possible by a grant from the United States Army and the engineering expertise of the General Electric Company, a Tomosynthesis scanner was built and installed at MGH.

A	B
A) Tomosynthesis slice with patent lesion B) Standard 2D LMLO view with obscure lesion

Standard 2D CC view with obscure lesion	Tomosynthesis slice with patent lesion

Preliminary Studies:

Breast cancer death rates that were constant for 50 years suddenly started to decline in the early 1990s, as the benefits of mass screening became evident. Such benefits are now well accepted, but there is still room for improvement. Out of 100 women with breast cancer today, 80 cases will be detected. Of the cancers that are identified, 68 (85%) will be seen with mammography, while 12 will be detected in clinical breast exams. Yet 20 will go unobserved. Mammography screening is good for detecting cancers that grow moderately quickly, but can miss very fast-growing cancers. Mammography is excellent and beneficial, but we need to continue to research better, faster and more accurate methods of breast cancer detection.

LESION VISIBILITY

In a pilot study of 40 women at MGH (a subgroup of the 400 volunteers), breast tomosynthesis helped improve lesion visibility compared with conventional mammography. With breast tomosynthesis, radiologists were able to see masses and architectural distortion better than with conventional mammography. In the pilot study, six out of seven malignant lesions identified on ultrasound that were occult on mammography were identified by tomosynthesis. At the special focus session, Rafferty presented an image of a young woman with very dense breast tissue and a palpable mass. A malignancy was not visible on the mammogram but was evident on ultrasound as well as on a breast tomosynthesis image. A second malignancy in the same breast was occult on mammography and ultrasound but was demonstrated by tomosynthesis. The reduction of noise from overlapping tissue (using tomosynthesis) also proved to be beneficial in imaging women with fatty tissue.

REDUCING RECALLS

Up to one-quarter of women recalled for follow-up after a mammogram ultimately turn out to have no abnormality. Researchers at Dartmouth Hitchcock Medical Center found that tomosynthesis could dramatically reduce the rate of false-positive mammograms. In a study of 98 patients with abnormal mammograms in 99 breasts, tomosynthesis as a screening follow-up tool was equivalent or superior to diagnostic mammography in nearly 87% of cases. There were 112 screening abnormalities in the study group. If tomosynthesis had been used as an adjunct to mammographic screening, the recall rate could have been cut by 40. On the basis of this very small study, tomosynthesis is equivalent or superior to mammography for diagnosis. It will offer recall reduction and better reflects the reality of the breast, translating into a more accurate diagnosis. Because of the very small scope of this study it is only indicative and must be backed up by larger clinical trials.

It took 20 seconds to acquire a tomosynthesis image and one minute to reconstruct the image. Two views were taken in most cases. Tomosynthesis proved superior to diagnostic mammography in 37 cases, equivalent in 49 cases, and inferior in 13 cases. Tomosynthesis was deemed inferior in some cases which involved calcifications; in addition, quality may have been affected by the lengthy acquisition time and motion artifact. Of the total number of potential abnormalities, five were cancers. Three cancers were detected by both mammography and tomosynthesis, and two cases were occult for both techniques.

The Challenge: This digital technique enabling physicians to “page through” the interior of the breast without obstruction by surrounding superimposed tissue has been hampered in the past because the method was computationally intense and took far too long for use in a clinical setting. Many of the tomosynthesis clinical trials to date have been disadvantaged by these challenges.

Is there a Possible Solution? Designers at Mercury Computer Systems implemented an iterative reconstruction algorithm called Maximum Likelihood Expectation Maximization which initially produced very high quality images. Engineers mapped the algorithm to a specialty graphics processor with a unique programmable rendering pipeline. Careful construction of the software port was required to overcome memory, bandwidth and instruction set limitations while optimizing run-time performance. The latter was reduced from five hours to five minutes. In pre-clinical testing, digital breast tomosynthesis permitted physicians to find cancers earlier and more easily, and to differentiate benign versus malignant lesions while simultaneously reducing false positives.

New and Better Tools:

• Digital Breast Tomosynthesis imaging system, co-developed by Massachusetts General Hospital and Mercury Computer Systems

• GE Healthcare (Waukesha, WI) Senographe 2000 Digital Mammography System

• NVIDIA (Santa Clara, CA) Quadro graphics processor Maximum Likelihood Expectation Maximization algorithm, co-developed by Brandeis University

Is the Solution Effective? The ensemble of observed results, software implementation and acceleration hardware has confirmed that digital breast tomosynthesis improved cancer screening accuracy, lowered costs, increased patient comfort and reduced exposure to ionizing radiation. Deaths from breast cancer dropped twenty percent over past decade, largely due to widespread screening mammography. Digital breast tomosynthesis’ superior ability to display lesions which otherwise might be masked could reinforce early detection and further reduce death rates. The Massachusetts General Hospital-Mercury Computer Systems technique sped up image reconstruction by a factor of sixty.

CONCLUSION:

Before digital breast tomosynthesis can be sanctioned for general use, it must be shown to be more effective in saving lives than conventional mammography. In light of the preliminary studies done in clinical settings such as MGH where digital breast tomosynthesis pinpointed seven cases which were not seen with mammography. Along with much engineering research making the new modality faster, clearer and easier to use; clearly, the stage is set to validate the technology and to propel digital breast tomosynthesis from a research tool into wide-spread clinical practice. Embracing change enhances knowledge and knowledge is power. Claim your power!