Algor Software Helps University of Kansas Professor to Improve Early Breast Cancer Detection Techniques

Breast cancer is the most common type of cancer in women, and is second only to lung cancer as a cause of cancer death among women. Approximately 184,300 women will be diagnosed with breast cancer in 1996, and 44,300 women with breast cancer will die.¹

The good news is that breast cancer mortality rates have been declining in recent years, in part because of early detection. Michael Insana, Ph.D. of the University of Kansas is using Algor software as part of a long-term study funded by the National Institutes of Health (NIH) to improve early detection techniques.

Dr. Michael Insana of the University of Kansas uses Algor software to conduct analyses of "phantom materials" that imitate the elastic properties of breast tissues. His analyses and experiments will hopefully lead to valuable discoveries about the best way to detect breast tumors in the early stages of development. To see a close-up of the image on Dr. Insana's screen, click here.

Early Detection Techniques

Currently, X-ray mammography and physical (tactile) breast examinations are the recommended methods for detecting breast tumors in the early stages. The accuracy of physical examinations, either by women themselves or by doctors, depends on the examiner's finger sensitivity. In addition, lumps that are very small or located next to the chest wall may not be detected. X-ray mammography can detect a tumor up to two years before it would be felt by a physical examination, but it is not usually useful for young women.² This is because young women have denser tissues, which does not give a clear X-ray image and because the radiation that is used to create the image can actually contribute to the development of breast cancer.³

Ultrasound is, therefore, being researched as an additional diagnostic method. It has the advantage of being harmless, non-intrusive and cost-effective. Dr. Insana's project is to make ultrasound detection more effective by optimizing signal processing so that medical professionals will get the clearest image possible.

Determining the Material Properties

Dr. Insana's first task was to determine the material properties of various kinds of breast tissues with the following testing cycle that uses an existing ultrasound device.

1. An ultrasound image of a breast is taken.
2. The tissues are compressed.
3. Another ultrasound image is taken.
4. The two images are compared.

By comparing the two ultrasound images, Dr. Insana determined the elastic properties of various kinds of breast tissues. Lumps are less elastic and mobile than healthy tissues.

Modeling "Phantom Materials"

Based on the elastic properties from the ultrasound tests, scientists in Dr. Insana's laboratory have formulated gelatin-like "phantom materials" that behave like breast tissues. Volumes of these materials are then modeled for analysis and testing. The model consists of several layers, each of which has different material properties.

Phantom materials have several advantages over using human test subjects. First, human tissues are more complex. By constructing models with phantom materials, researchers can learn to recognize true tumors from other kinds of tissues. Second, inconvenience to both researcher and subject is eliminated.

Existing ultrasound machines use a probe with a small surface area which must be passed over a surface several times to construct a complete image. One of Dr. Insana's goals is to establish whether tumors can be detected more easily when the tissues are palpitated between plates (as in a traditional mammography). To that end, he is conducting linear stress analyses to discover the displacement of a phantom tumor within surrounding material and any deformation of the volumes. Boundary conditions are established which replicate palpitation between plate and pressure from ultrasound probes.

A 2-D stress image of a tumor nine times stiffer than the surrounding tissues. The surface was compressed 1.5% to observe this stress field. The circular tumor appears diamond-shaped and there are four characteristic "wings" around the tumor. FEA helps Dr. Insana to understand these artifacts common in clinical imaging. The effects of boundary conditions can also be seen in the image, where the top surface is free to slide and the bottom is bound.

Dr. Insana's analyses are already yeilding preliminary results. The image on the left shows an Algor linear stress analysis of two hard spheres in a gelatinous materials. The small spere represents a benign tumor, the large one, a malignant. When the lesions are compressed, the stress patters around the lesion are very complex. The figure on the right is an elastogram of the same model of the two spheres. Comparing the image of the Algor analysis and the elastogram, researchers can evaluate how well the stress patterns are produced in the elastogram and teach physicians how to read elastograms. Click on the images to see enlarges versions of both images.

Optimizing Ultrasound Configuration

Ultrasound machines are currently used in a number of medical applications. Therefore, many design issues are already solved. Dr. Insana is focusing on how best to process the ultrasound signals so that medical professionals will get the information they need from an examination. A five-year NIH grant will continue to fund Dr. Insana's research for three more years. The grant also funds related projects at the University of Texas, University of Montreal and the University of Kansas. All are Algor users.

About Algor

Although Dr. Insana has only been using Algor software for a few months, he already has a few favorite features.

"Using the axisymmetric elements has been really great," said Dr. Insana. "I can very quickly model cross-sections and get results which can be extrapolated into 3-D.

"Algor software also handles multiple material properties very well," continued Dr. Insana. "Test cases are often comprised of layer-upon-layer of different kinds of tissues. I don't know of any other package which would enable me to do this kind of analysis."

In the future, Dr. Insana will be incorporating Merlin, Algor's surface mesh enhancement and generation, for automatic meshing. Merlin is expected to meet the challenges of meshing the complex layers used in some of his models.

Sources

1. American Cancer Society's Breast Cancer Statistics
2. Breast Cancer Answers, Frequently Asked Questions
3. American Cancer Society Breast Resource Center, Finding Breast Cancer