Abstract
Measurements of length and section area on MR images are made conventionally by the manipulation of a tracker ball to move a cursor to define end points or delineate boundaries, This operator-dependent method lacks the objectivity that computer-based image analysis techniques, such as edge-detection algorithms, can give, me describe an image analysis routine, based on earlier work by Vincent and Soille using algorithms of mathematical morphology, for automatically extracting the boundary of lumbar intervertebral discs on MR images, This is applied to the automatic measurement of the maximum disc space height and section area of non-degenerate intervertebral discs on MR images obtained for healthy male subjects of two distinct age ranges, i.e. twenty 20- to 30-year-olds and twenty 50- to 60-year-olds. Throughout the levels L1 to L4, intervertebral disc height is found to be significantly greater (P = .0003 to .0161) in the older than in the younger age group. The data support a model of the spine in which the disc adopts a more concave form as a result of microfracturing of the end plates during adult life.
Original language | English |
---|---|
Pages (from-to) | 880-886 |
Number of pages | 7 |
Journal | Journal of Magnetic Resonance Imaging |
Volume | 7 |
Issue number | 5 |
Publication status | Published - 1997 |
Keywords
- image analysis
- intervertebral disc
- lumbar spine
- MRI
- mathematical morphology
- top-hat algorithm
- watershed algorithm
- SPINE