R2 Technology has evolved the sensitivity of its ImageChecker about as far as possible and is now focusing R&D on providing physicians with qualitative information potentially useful in evaluating mammograms.“The more information we give
R2 Technology has evolved the sensitivity of its ImageChecker about as far as possible and is now focusing R&D on providing physicians with qualitative information potentially useful in evaluating mammograms.
The more information we give radiologists, the more we can improve their ability to detect breast cancer at the earliest stages, when the outcomes for women are best, said James Pell, president and CEO of the Los Altos, CA, company.
The sensitivity milestone was underscored by FDA supplemental approval of R2s version 2.2 software, which achieves 98% sensitivity identifying microcalcifications and 86% identifying masses. Moreover, increased sensitivity was not attained at the expense of specificity, according to Sangeeta Gupte, R2 marketing manager. Overall, the new software marks 20% fewer suspicious lesions, Gupte said.
Software features were improved to ensure that marks made in the digital mammograms are clinically meaningful and represent indications that the radiologist wants and needs to see, she said. This was accomplished through advanced algorithms running on an upgraded computing platform and operating system.
The workstation, called ImageChecker M1000, is intended as a second reader for radiologists, directing attention to possible lesions after the physician has conducted a primary reading of the mammograms.
Pattern-matching algorithms electronically mark suspicious lesions in the digital images.
The M1000 was tested against data obtained from the clinical trials that supported ImageCheckers initial FDA approval in 1998. Company executives believe that for every 100,000 cancers detected by mammography, use of the advanced form of ImageChecker could result in early detection of an additional 14,200.
About 100 ImageCheckers are in service, most of them in the U.S. None, however, are included in the full-field digital mammography system being sold by GE Medical Systems, despite an alliance R2 has with that company. The reason, according to Gupte, is the FDA.
We do have a contract with GE and, basically, it says that once GE gets approval to have physicians read off soft copy, then we will incorporate (ImageChecker) into the system, she said. Right now GE only has approval for hard-copy reading.
The R2 ImageChecker now in operation digitizes mammography films and then reads the electronic images. These systems have an average selling price of $190,000 to $200,000, depending on monitor type and capacity of the automatic film feeder.
Several other companies make similar workstations. Qualia Computing of Beavercreek, OH, has developed Second Look, a computer-assisted detection system that runs on PCs. Its effectiveness in detecting breast cancer is now being evaluated in clinical trials. Quebec-based CADx has exclusive worldwide marketing rights to Second Look. Another competitor, Scannis of Foster City, CA, is in the early stages of clinically evaluating its Mammex TR CAD workstation.
Ideally, CAD systems will eventually clarify what leads to the identification of suspicious lesions, rather than just displaying, as they do now, triangles or asterisks surrounding suspected lesions. Addressing this challenge is the next step for R2 engineers.
Were not exactly sure how were going to do that, Gupta said. Thats why its a work-in-progress.
Study Assesses Lung CT-Based AI Models for Predicting Interstitial Lung Abnormality
September 6th 2024A machine-learning-based model demonstrated an 87 percent area under the curve and a 90 percent specificity rate for predicting interstitial lung abnormality on CT scans, according to new research.