An algorithm optimized to run on chips developed for video gaming promises a 50-fold acceleration in the postprocessing needed to register 3D data sets from different imaging modalities or exams conducted at different times.
IBM's Cell Blade together with an algorithm developed by IBM and the Mayo Clinic register imaging data 50 times faster than can be done with a conventional workstation. The alignment of images sets the framework for their fusion, which simplifies the detection of tumor changes that may indicate the likelihood of malignancy or patient response to chemo- or radiotherapy. IBM is negotiating with several large imaging vendors to bring the algorithm and Cell-powered computing hardware to the marketplace.
An algorithm optimized to run on chips developed for video gaming promises a 50-fold acceleration in the postprocessing needed to register 3D data sets from different imaging modalities or exams conducted at different times.
A test run of data sets acquired at the Mayo Clinic in Rochester, MN, showed that registering key reference points from two data sets can be accomplished in less than a second, compared with more than a minute on a conventional workstation, according to Dr. Bradley Erickson, a neuroradiologist and director of the radiology informatics lab at the Mayo Clinic. Registering two entire volumetric data sets can be done in less than 20 seconds, compared with several minutes.
"This has the potential to make radiologists faster and more accurate," said Erickson, who with a colleague from IBM will present findings from initial tests today at the IEEE (Institute of Electrical and Electronics Engineers) International Symposium on Biomedical Imaging.The alignment of images sets the framework for their fusion, which simplifies the detection of tumor changes that may indicate the likelihood of malignancy or patient response to chemo- or radiotherapy. Such registrations are now impaired by the time needed to do them, particularly since computers do not always succeed on the first try. "If you start from the wrong point, the data sets might not converge to the right solution," said Shahrokh Daijava, IBM's software lead for the company's next-generation systems. "If every (attempt to converge) takes five minutes and several attempts are necessary, the radiologist will get frustrated and give up." Running an algorithm developed through a collaboration between the Mayo Clinic and IBM on an IBM BladeCenter QS20 outfitted with a Cell processor solves that problem. Computations performed on a Cell-based component, called the IBM Cell Blade, register data 50 times faster than can be done with a conventional workstation. The cornerstone of the advance is the Cell Broadband Engine processor developed as part of a joint venture by IBM, Toshiba, and Sony Group to handle the kind of rapid image rendering demanded by high-performance video games. Cell drives Sony's PlayStation 3.Together, the algorithm and Cell-powered BladeCenter find the best spatial positioning for registering 2D images, translate this to all the slices in the data sets, and then very rapidly register the data. Daijava and Erickson will report at the IEEE conference that registrations performed on 98 data sets using a conventional processor took seven hours, whereas the same registrations took just eight minutes, 36 seconds on the IBM BladeCenter QS20.Although the tests so far have focused on a dedicated workstation, the final implementation need not be so constrained. Cell processors might be integrated into a network as part of a thin-client server or even a PACS connecting desktop PCs, according to Erickson.IBM is now negotiating with several large imaging companies that might bring the algorithm and Cell-powered computing hardware to the marketplace. "IBM needs to partner with a big vendor like GE or Hitachi or Siemens that has been in this field for a long time," he said. "We are negotiating with more than one. But in the end it will be just one partner."
Comparative AI Study Shows Merits of RapidAI LVO Software in Stroke Detection
February 6th 2025The Rapid LVO AI software detected 33 percent more cases of large vessel occlusion (LVO) on computed tomography angiography (CTA) than Viz LVO AI software, according to a new comparative study presented at the International Stroke Conference (ISC).
Study: Mammography AI Leads to 29 Percent Increase in Breast Cancer Detection
February 5th 2025Use of the mammography AI software had a nearly equivalent false positive rate as unassisted radiologist interpretation and resulted in a 44 percent reduction in screen reading workload, according to findings from a randomized controlled trial involving over 105,000 women.
New CT Angiography Study Shows Impact of COVID-19 on Coronary Inflammation and Plaque
February 5th 2025Prior COVID-19 infection was associated with a 28 percent higher progression of total percent atheroma volume (PAV) annually and over a 5 percent higher incidence of high-risk plaque in patients with coronary artery lesions, according to CCTA findings from a new study.