Elastography stretches horizons of breast ultrasound

October 1, 2006

Initial results convince Europeans experimental technique may reduce benign breast biopsy rates

Initial results convince Europeans experimental technique may reduce benign breast biopsy rates

Researchers at ECR 2006 hailed the potential of ultrasound elastography to dramatically reduce benign breast biopsy rates.

Elastography refers to the measurement of elastic properties of tissues, based on the well-established principle that malignant tissue is harder than benign tissue. The technique is typically performed with ultrasound, but research with MR is also under way.

The procedure, also known as elasticity, was born in the early 1990s. The University of Texas Medical School in the U.S. holds multiple patents on the technique with ultrasound (see www.elastography.com for details). Images are acquired on high-end ultrasound devices equipped with additional software and hardware. Prototype commercial systems became available to researchers after 2000, and vendors have since introduced commercial products, accelerating interest.

Using a research system from Siemens, which is not yet commercially available, U.K. radiologists reported results from four years of routine elastography in breast screening. They concluded that the technique could help halve the rate of benign breast biopsies.

"We can decrease the biopsy rate in benign lesions. This has great importance given the increased incidence of screening abnormalities picked up on breast ultrasound," said Dr. William Svensson, a consultant radiologist at Charing Cross Hospital in London.

With elastography, raw ultrasound is obtained before and after a slight compression of tissue, typically achieved with an ultrasound transducer. Compression may also be performed using vibrations in a technique known as sonoelastography. Elastography measures and displays strain; that is, the change in the dimension of tissue elements at various locations in the region of interest.

An estimate of tissue strain is obtained at each point in the field-of-view, and strain values are then displayed as an image. The strain is estimated from minute differences between these two images.

Research has shown that normal tissue and fat have smaller elasticity profiles, while hard areas, such as cancers, are larger than the gray-scale appearance.

"The hypothesis was that the size of cancer is larger in elasticity imaging than B-mode and with benign lesions the reverse is the case," said Svensson.

The U.K. study included women who were referred for routine breast ultrasound for focal breast abnormalities. In cases where abnormalities were identified on ultrasound, women underwent breast strain imaging.

Of 260 lesions identified, 71 were malignant, nine were intermediate with malignant potential, and the majority, 180 lesions, were benign.

U.K. researchers noted that lesions with an elasticity to gray-scale ratio of less than .75 were benign. Based on this ratio, all of the malignant lesions would have been identified, aside from a few intermediate lesions. That result equates to a sensitivity of 96%.

In the study, specificity reached just 53%, meaning that if the ratio is applied, half of the benign lesions might not need to be biopsied, Svensson said. In addition to the obvious clinical value, there are also economic advantages.

"Benign biopsies cost us a lot of money," he said. "If we can reduce them, that would bring significant savings for the health service."

In another study presented at the ECR, French researchers reported positive findings for ultrasound elastography in a multicenter prospective study of 345 breast lesions in 314 patients.

Patients were imaged on the Hitachi EUB 8500 Logos ultrasound unit with a technique similar to the one used in the U.K. study. Detected lesions were categorized by size and BI-RADS category.

Researchers used the color classification system developed by Dr. Ei Ueno to score elastography images. For example, elastography images that are completely green (soft) are typically benign with a Ueno score 1, while those that are completely blue (stiff) are thought to be malignant and have a Ueno score 5.

For lesions of all sizes, ultrasound elastography achieved sensitivity of 80%, specificity of 93%, positive predictive value of 85.3%, and negative predictive value of 90.3%. Sensitivity was highest for lesions less than 5 mm (90%), while specificity was best for lesions over 10 mm (95%).

For lesions in BI-RADS categories 3 and 4, sensitivity was 68% and specificity was 90%.

Researchers also reported 16 false positives with elastography (such as fibrous mastopathy and sclerosis adenosis) and 26 false-negative findings (such as DCIS).

The technique is most useful for lesions in the indeterminate BI-RADS categories and less useful for lesions in BI-RADS category 5, as false negatives might occur in these lesions. In obvious, suspicious nodules on B-mode imaging, elastographic assessment is either of very little use or not needed, said Dr. Anne Tardivon of the Institut Curie in Paris.

Elastography is easy to perform once practitioners are trained, Tardivon said. It provides good visualization and is not time-consuming.

"Elastography does not replace standard B-mode imaging, but, clearly, it is a useful complementary tool and may reduce biopsies of benign lesions," she said.

At this time, the modality is being used for breast nodules detected on standard ultrasound studies, she said. Its potential role in evaluating patients without an abnormality on B-mode imaging remains to be evaluated.