Automated detection and evaluation of lumbar discs in MR images

Abstract

A method for computerized detection of disk pathologies is described. It is based on: (1) automated localization of individual disks in a magnetic resonance (MR) image, (2) analysis of the pattern of changes in grey intensities along the disk, and (3) verification of findings based on the analysis of trends recorded for subsequent images in the study. In preliminary studies there was good correlation between abnormal intensity distribution and clinical evidence of disk abnormalities, suggesting that the method will be useful in automated, unconstrained analysis of MR images of the spine.< <ETX xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">&gt;</ETX>