| |||||||||||||||
|
Home
|
For other articles and previous issues click here. September 19, 2005 CAD
for Ultrasound — New Tool to Help Breast Cancer Diagnosis Two Canadian medical companies plan to jointly market the first-ever computer-aided detection (CAD) system for breast ultrasound. The software solution, which will be marketed under the name Cedara B-CAD, was developed by the Medipattern Corporation of Toronto and will be distributed by the Cedara Software Corporation of Mississauga, Ontario. Both companies anticipate that this system will greatly impact the diagnosis and treatment of breast cancer. Their optimism appears justified. It has already been demonstrated that B-CAD can boost radiologists’ diagnostic confidence by helping them identify sonographic characteristics associated with breast lesions and by helping them analyze solid breast nodules. B-CAD, which received 510(k) clearance from the FDA this past spring, uses pattern recognition methods to help radiologists analyze images and automate the subsequent reporting process. The software employs a complex image analysis scheme to determine more than 20 sonographic lesion characteristics, including differentiation between benign and malignant lesions, according to its developers. Consequently, its developers believe B-CAD will assist radiologists in the analysis of lesions and promote standardized breast ultrasound reporting. Urgent Mission “He had been involved in pattern recognition technology in other industries, so he took the knowledge he had picked up and began applying it to medical imaging technology,” says Kevin Clark, Cedara’s marketing and communications manager. As with any innovation designed to target cancer, a sense of urgency drove the development of the B-CAD software. Collins was well aware of the American Cancer Society statistics that indicate breast cancer is on the rise. Breast cancer incidence has leapt from one in 20 women in 1960 to one in eight today. In 2004, the American College of Radiology (ACR) estimated that approximately 216,000 new cases of noninvasive breast cancer were diagnosed. Breast cancer in his family gave Collins added impetus. He was motivated by the death of his aunt, who succumbed to breast cancer. This compelled him to put his knowledge into developing a solution that would assist with the early detection of breast cancer. “He felt that he had to do something to somehow reduce the number of needless deaths from this terrible disease,” says Sabrina Cannistraro, Cedara’s product manager for women’s health. The result was a core technology in the form of a software platform designed to help healthcare professionals analyze and automate the processing of medical images. B-CAD was developed for automated detection and diagnosis in breast ultrasound, explains Collins. By the end of 2003, Medipattern had its prototype. In 2004, the company developed a relationship with Cedara Software, a division of Merge Technologies Inc. that provides medical imaging technologies. Both parties agreed that Cedara would bring the B-CAD software to the marketplace as well as support its development. Integrating the Stavros Criteria In developing B-CAD, Medipattern worked closely with Stavros, who is currently the director of ultrasound at the Sally Jobe Breast Center in Denver, to incorporate the criteria into the software. “We originally based B-CAD on the criteria and developed algorithms to extract the features associated with those criteria,” says Collins. As the software developed, it evolved to also include the ACR’s Breast Imaging and Reporting Data System (ACR BI-RADS). The original BI-RADS lexicon was developed for mammography and adopted by the Mammography Quality Standards Act (MQSA). As breast ultrasound usage increased, the ACR then developed the BI-RADS Ultrasound Lexicon for breast sonography as a means to standardize the characterization of sonographic lesions. Medical professionals believe using the BI-RADS systems will help the diagnostic capability of breast ultrasound and facilitate acceptance and utilization of the ACR BI-RADS Ultrasound Lexicon by radiologists. Many in the field expect that the MQSA will adopt the ultrasound lexicon as a standardized reporting format. Meanwhile, the B-CAD report form is intended to support compliance with the ACR BI-RADS Ultrasound Lexicon Classification Form. “As the ACR has developed the BI-RAD Ultrasound Lexicon, our system similarly evolved,” says Collins. Luminary Testing • Wendie Berg, MD, PhD, breast imaging consultant and study chair for American College of Radiology Imaging Network protocol 6666; • Marcela Bohm-Velez, MD, Weinstein Imaging Associates, Pittsburgh; • Paula Gordon, MD, FRCPC, clinical professor, department of radiology, University of British Columbia;
• Stavros and Cindy Rapp, BS, RDMS, RDCS, Sally Jobe Breast Center, Greenwood Village, Colo.; and • Gary Whitman, MD, associate professor of radiology, section of breast imaging, University of Texas M.D. Anderson Cancer Center. Both Medipattern and Cedara believe the testing will provide clinical benchmarking for B-CAD. The companies anticipate the system will be commercially available in the United States in late fall. Reducing Biopsies “Typically, when something is found during a screening mammogram, a woman receives a call back so that a suspicious area can be examined more closely,” says Cannistraro. “Many times, ultrasound is used to examine the suspicious [case], and from that comes a decision whether or not to biopsy. B-CAD looks at the ultrasound image and analyzes different features based on the ACR BI-RAD lexicon. It helps the radiologist determine whether a biopsy needs to be performed. Still, the radiologists always make their own decisions in the end. The software offers a suggestion, not a final say.” If that step can reduce the number of unnecessary biopsies, it would be a significant advancement. Needle biopsies are invasive procedures that can be physically and emotionally traumatic for a woman. Further, core biopsies can also lead to scar tissue, making future imaging difficult. “A biopsy is a frightening thing,” emphasizes Clark. “It is one thing for a woman to have a mammogram or an ultrasound examination. But when things are elevated to the level of biopsy, there’s a tendency for a woman to believe that something malignant has been found. Of course, that’s not always the case.” Implementation Options It could also be integrated into DICOM-compliant workstations and ultrasound machines. B-CAD will also be available within Cedara’s I-ReadMammo, a soft-copy review workstation for mammography and multimodality breast images. I-ReadMammo offers a combination of CAD systems for multiple breast imaging modalities in a streamlined package. It provides viewing capabilities for mammography and multimodality breast data and works with all the major brands of imaging equipment. According to Collins, the B-CAD will be available to original equipment manufacturers (OEMs) for integration with imaging systems or inclusion in next-generation ultrasound equipment. Even before its market introduction, B-CAD has generated a great deal of interest from potential end users and OEM customers. Improving Workflow “The whole process, from loading of the image to generating a report, is very fast, so it speeds up the workflow,” says Collins. “In that way, it is also a business solution. People at busy centers don’t want to spend a lot of time doing things like fetching information from a database. The software doesn’t use a database, or case-based reasoning, or other concepts. It is an application that runs very quickly.” Eventually, B-CAD will be available in two systems: B-CAD 1.0 and 2.0. The first release, B-CAD 1.0, looks at shape and orientation from the BI-RAD lexicon. The radiologist will have to do the analysis on the rest of the features. B-CAD will act as a second read and give its BI-RADS assessment. “In that case, it is the second opinion that is really important, not necessarily the automatic feature extraction,” says Cannistraro. “The radiologists who use 1.0, will be selecting those other features themselves. Therefore, the suggestion that the B-CAD provides is the key part of the software.” The 2.0 software will automatically find all the different features in the lesion and then give its suggestions. By the end of this year, B-CAD will become available in the United States. Utilizing pattern recognition and complex algorithms, the software system can help facilitate earlier detection of cancer by enhancing the diagnostic confidence of radiologists and decrease the number of unnecessary biopsies. It may also encourage increased adoption of breast ultrasound, which has proven to be especially useful in evaluating masses in young, pregnant, or lactating women. It may also widen acceptance and utilization of the ACR BI-RADS Ultrasound Lexicon. The system’s developers believe B-CAD can have a significant impact on the diagnostic capability of breast ultrasound. Looking to the future, even as B-CAD awaits its introduction into the U.S. market, Medipattern and Cedara already envision other uses for the pattern recognition platform that went into its design. They’re looking to extend its usefulness into other modalities and early detection in other diseases. — Dan Harvey is a freelance writer based in Wilmington, Del., and frequent contributor to Radiology Today. |
 |
|
3801 Schuylkill Rd • Spring City, PA 19475 Publishers of Radiology Today All rights reserved. |