Researchers at the University of Chicago are developing new imaging techniques using computer-aided diagnosis (CAD) and quantitative imaging analysis for mammograms, ultrasound and MRIs.
These new techniques will help identify specific tumor characteristics such as size, shape and sharpness.
This represents a change in the way that CAD is used, said lead researcher Maryellen Giger, Pritzker Professor of Radiology/Medical Physics and director of the Imaging Research Institute at the University of Chicago, since CAD is currently used to provide a "second opinion" for radiologists looking for suspicious regions on mammograms.
According to Giger, the use of quantitative image analysis, as well as CAD, is also expanding beyond use as a screening tool and towards risk assessment, diagnosis, prognosis, response to therapy and using data to identify how tumor characteristic apply to disease states.
In an article published at the Professional Society for Optics and Photonics (SPIE), Giger writes that quantitative image analysis could be used to merge relevant tumor features "into diagnostic, prognostic, or predictive image-based biomarkers, and could estimate the probability of a particular disease state." She adds that it could also be used to retrieve and compare similar tumor cases with a particular tumor, and enable researchers to examine the relationships among image-based tumor characteristics across populations.
Giger also points out that quantitative analysis of digital mammograms can yield information on breast density and parenchymal patterns. "While results are promising across screen-film and full-field digital mammograms, we still need clinical validation within a large screening program," she wrote. "We are also extending such analysis to breast ultrasound and MRI, which we are investigating in screening women deemed to be at high risk for future breast cancer."
In order to calculate image-based signatures of breast lesions for diagnostic and prognostic assessments, Giger says she and her colleagues investigated tumor and parenchymal features from breast images across modalities, with the idea that eventually radiologists will be able to use the information in a diagnostic workup to determine the likelihood of whether a tumor is cancerous.
After that an assessment of prognosis and potential treatment options is necessary, which can be accomplished, Giger says, by relating computer analyses of imaged tumors to pathology and molecular classifiations: "We use the computer-extracted features of the tumor to assess its aggressiveness, as a ' virtual biopsy,' which could be used with clinical biomarkers to help determine patient management."
SPIE also recently announced that Giger has been named the editor-in-chief of a new journal, the Journal of Medical Imaging, which, according to SPIE, will be launched in early 2014.
Mammo plus CAD increases cancer detection at cost of more false positives
CAD boosts ID rate for lung nodules
Computer-aided detection could hinder image interpretation