Separating arterial and venous components from 3D dynamic contrast-enhanced MRI studies using factor analysis

• Published in: Magnetic resonance in medicine Vol. 49; no. 5; pp. 928 - 933
• Main Authors: Martel, Anne L., Fraser, Doug, Delay, G. Steve, Morgan, Paul S., Moody, Alan R.
• Format: Journal Article
• Language: English
• Published: New York Wiley Subscription Services, Inc., A Wiley Company 01-05-2003; Williams & Wilkins
• Subjects: Biological and medical sciences; Contrast Media; contrast-enhanced; factor analysis; Factor Analysis, Statistical; Humans; Image Processing, Computer-Assisted - statistics & numerical data; Magnetic Resonance Angiography - methods; Magnetic Resonance Angiography - statistics & numerical data; Medical sciences; MR venography; MRA; Phlebography; Signal Processing, Computer-Assisted; Statistics, Nonparametric; Subtraction Technique; Thrombophlebitis - diagnosis; Thrombophlebitis - diagnostic imaging; contrast-enhanced; factor analysis; MR venography; MRA
• ISSN: 0740-3194; 1522-2594
• DOI: 10.1002/mrm.10462

Dynamic contrast‐enhanced MRI has been used extensively for angiography but in order to generate separate arterial and venous images some form of postprocessing is required. This typically involves the subtraction of one image in a dynamic sequence from another in order to suppress unwanted signal; however, this also has the effect of decreasing the signal‐to‐noise ratio (SNR) of the image. In this study, factor analysis, a technique related to eigenimage filtering, is used to separate arterial and venous components from dynamic contrast‐enhanced images of the legs acquired with a temporal resolution of 30 sec. The SNR of the venous and arterial images extracted from a series of 20 patients using conventional single subtraction, a double subtraction method, and factor analysis were compared. Results show that the use of factor analysis improved the SNR in the venous images by a factor of 2.3 compared with the use of simple subtraction. A subjective comparison of the maximum intensity projection images generated from the venous images was also carried out and showed a significant preference for those generated using factor analysis over those generated using other subtraction methods. Magn Reson Med 49:928–933, 2003. © 2003 Wiley‐Liss, Inc.