RT Image October 4, 2010 : Page 14

A Conversation with . . . Brent Lewis and Paul Merrild CAD-MRI continues to provide efficiency and access COMPUTER-AIDED DETECTION (CAD), AS IT APPLIES TO RADIOLOGY, can be defined as the use of a software application to aid in the clinical review, evaluation, and reporting of radiological images. But as the use and appli-cations of CAD continue to grow, so does some confusion. Brent Lewis and Paul Merrild of Merge Healthcare spoke with rt image to clear up the discrepancies and provide insight into this developing area. Lewis is vice president of regulatory affairs at the Hartland, Wis.-based company, which develops solutions to automate healthcare data and diagnostic work in order to improve electronic records. Merrild is senior vice president of marketing and business development. Brent Lewis: In the last 2 years, panel meetings and guidance documents from the FDA have identified two categories of CAD devices. Computer-aided detection (CADe) devices incor-porate pattern recognition and data analysis capabilities to direct attention to regions of image data that may reveal abnormalities. Computer aided diagnosis (CADx) provide a likelihood of malignancy for a specific finding. These recent evolutions in CAD terminology have resulted in confusion over what constitutes a CAD device. For example, while mammography CAD devices fall under the CADe definition, the breast MRI-CAD counterparts have different intended uses and designs that exclude them from the FDA definition. CAD is implemented in MRI-CAD devices through workflow automation and complex quantitative analysis to provide reading and reporting efficiencies, as well as confidence in clinical decision making. Q A Paul Merrild: CAD for MRI was first developed to assist physi-cians with the manual processing and reporting associated with breast MRI studies. Q A |14| October 4, 2010 rt image: How did MRI-CAD come to be? rt image: How would you describe CAD? Paul Merrild is senior vice president of marketing and business development at Merge. Because this complex study produces thou-sands of images, technologists and physi-cians sometimes spent hours processing and reviewing a single study. A few years ago, many sites we worked with said they didn’t believe they could offer breast MRI studies because they could not spend the time neces-sary to process and read them. Concerns also existed with standardization; it was diffi-cult to standardize interpretation with multiple physicians across multiple reading sites. Once CAD was implemented, customers reported a 60-percent to 80-percent reduc-tion in reading time and a 99-percent reduc-tion in technologist handling time. Sites con-sistently report that CAD has brought stan-dardization to their practices. With CAD, studies are processed and interpreted the same way every time, which has been impor-tant as more and more patients are being monitored with MRI. In addition, in the past few years, peer-reviewed studies have shown an increase in specificity using MRI-CAD. Today, more imaging facilities are offering breast MRI due to the reading efficiencies and diagnostic confi-dence recognized through CAD. MRI-CAD also assists with MR-guided breast interventions. Biopsies under MRI can be time-consuming and difficult to perform. MRI-CAD systems help physicians quickly calculate coordinates for interventions. Brent Lewis is vice president of regulatory affairs at Merge Healthcare. Q A rt image: How has the technology evolved? Lewis: MRI-CAD has developed into a robust tool for analysis and reporting not only for breast MRI studies, but also for liver and prostate MRI studies. The same principles apply: liver and prostate MRI studies are complex and require significant time to process and report. CAD assists the radiologist by reducing the amount of time spent with manual analysis and interpretation. CAD also provides access to data that was previously unavail-able or difficult to obtain. For example, prior to MRI-CAD for the liver, most sites struggled with motion artifact. CAD’s motion correction has greatly improved study analysis and, in some cases, has Merge Healthcare

Q&A: A Conversation with ... Brent Lewis and Paul Merrild

Previous Page  Next Page

Publication List