Medical imaging continues to inspire elaborate software schemes that will help radiologists characterize tissue by extracting more meaningful information from digital images.One system, called Accent, is designed to help radiologists use MR to make
Medical imaging continues to inspire elaborate software schemes that will help radiologists characterize tissue by extracting more meaningful information from digital images.
One system, called Accent, is designed to help radiologists use MR to make more accurate cancer diagnoses. The system is currently in clinical trials.
The user selects a region of interest (ROI) that includes a known tumor. A clustering method then sorts the pixels within the ROI into separate groups. The user selects one or more clusters from within the ROI to represent the desired tissue. Accent automatically evaluates all the data in the images, highlighting potential sites of metastases.
"Accent uses multivariant statistical techniques to combine the information in all MR sequences of a given region to help physicians evaluate the information in ways that are beyond the ability of human vision," said Dr. Justin Smith, a practicing radiologist and founder of Confirma, the company developing the system.
Digital images are typically displayed on a gray-scale value of zero to 255, even though human physiology limits discrimination to about 140 shades of gray.
"There's nothing else like this out there," said one of the trial's principal investigators, Dr. Jay Tsuruda, a staff neuroradiologist at Inland Imaging in Kirkland, WA.
Accent creates almost a synthetic image by going through MR series and highlighting areas of tissue that have similar characteristics, Tsuruda said.
The new series is just as powerful as, if not more powerful than, the original series because it's annotated, he said. This means radiologists can use it to help decide if something needs further analysis.
"MR spits out tons of images," Tsuruda said. "We acquire so many images so quickly that it's almost overwhelming trying to piece together not only where things are three-dimensionally in anatomy that may be quite complex, but also to analyze what the signal is doing."
Accent could give radiologists another contrast end point to make decisions about whether something is of concern, Tsuruda said, although he stopped short of calling the system computer-aided diagnosis (CAD).
"In a CAD system, you feed in all the scans and it automatically makes a diagnosis for you," he said. "In this case, the system creates a new series and highlights areas of question based on multivaried analysis of signal intensity."
Another system under development, called Insight Segmentation and Registration Toolkit, is available from the National Library of Medicine. It is an open-source package for performing registration and segmentation -- the process of identifying and classifying data found in digital images.
Registration is the task of aligning or developing correspondences between data in two, three, and more dimensions. CT images, for example, may be aligned with an MR scan to combine the information contained in both. The system could prove useful in neurosurgery for defining the safest possible surgical approach.
FDA Approves Fluorescence Imaging System for Detecting Residual Breast Cancer
April 18th 2024The combination of the optical imaging agent Lumisight and the fluorescence imaging device Lumicell Direct Visualization System, collectively known as LumiSystem, reportedly offers 84 percent accuracy with real-time detection of residual breast cancer after lumpectomy procedures.
Study of Ofatumumab for Multiple Sclerosis Shows 'Profoundly Suppressed MRI Lesion Activity'
April 17th 2024The use of continuous ofatumumab in patients within three years of a relapsing multiple sclerosis diagnosis led to substantial reductions in associated lesions on brain MRI scans, according to research recently presented at the American Academy of Neurology (AAN) conference.
Could a Deep Learning Model for Mammography Improve Prediction of DCIS and Invasive Breast Cancer?
April 15th 2024Artificial intelligence (AI) assessment of mammography images may significantly enhance the prediction of invasive breast cancer and ductal carcinoma in situ (DCIS) in women with breast cancer, according to new research presented at the Society for Breast Imaging (SBI) conference.