RATIONALE AND OBJECTIVE: An application of independent component analysis (ICA) is reported for the detection and characterization of breast lesions in dynamic contrast-enhanced MRI. The ICA technique, which uses a novel statistical algorithm to separate mixed signal sources of unknown nature, enables the extraction of spatial and temporal features of dynamic MRI data. MATERIALS AND METHODS: Six patients with confirmed lesion diagnosis (three with malignancy and three with benign condition) participated in this study. T1-weighted MRI covering both breast volumes was dynamically acquired in every 90 seconds after the bolus injection of contrast media using three-dimensional fast low-angle shot sequence (3-D FLASH). RESULTS: With the application of the ICA, differential signal responses were delineated from benign and malignant lesions, and areas with high temporal correlations with the extracted signal components were selectively visualized. CONCLUSIONS: The results suggest that ICA allowed for an identification of lesion morphology and lesion-specific dynamic enhancement patterns.
RATIONALE AND OBJECTIVE: An application of independent component analysis (ICA) is reported for the detection and characterization of breast lesions in dynamic contrast-enhanced MRI. The ICA technique, which uses a novel statistical algorithm to separate mixed signal sources of unknown nature, enables the extraction of spatial and temporal features of dynamic MRI data. MATERIALS AND METHODS: Six patients with confirmed lesion diagnosis (three with malignancy and three with benign condition) participated in this study. T1-weighted MRI covering both breast volumes was dynamically acquired in every 90 seconds after the bolus injection of contrast media using three-dimensional fast low-angle shot sequence (3-D FLASH). RESULTS: With the application of the ICA, differential signal responses were delineated from benign and malignant lesions, and areas with high temporal correlations with the extracted signal components were selectively visualized. CONCLUSIONS: The results suggest that ICA allowed for an identification of lesion morphology and lesion-specific dynamic enhancement patterns.
Authors: Matthew Marzelli; Krisztina Fischer; Young Beom Kim; Robert V Mulkern; Seung-Schik Yoo; Hyunwook Park; Zang-Hee Cho Journal: Int J Imaging Syst Technol Date: 2008-06-13 Impact factor: 2.000
Authors: Thorsten Twellmann; Anke Meyer-Baese; Oliver Lange; Simon Foo; Tim W Nattkemper Journal: Eng Appl Artif Intell Date: 2008-03 Impact factor: 6.212