The reproducibility of quantitative measurements in lumbar magnetic resonance imaging of children from the general population

Spine (Phila Pa 1976). 2008 Sep 1;33(19):2094-100. doi: 10.1097/BRS.0b013e31817f19f7.


Study design: Quantitative lumbar magnetic resonance imaging (MRI) measurements in children were taken twice and analyzed for intra- and intertester reproducibility.

Objective: To evaluate the reproducibility of a variety of lumbar quantitative measurements taken from MRIs of children from the general population.

Summary of background data: The reproducibility of lumbar quantitative measurements has never been tested on MRIs of children from the general population. This is a prerequisite for a reliable anatomic characterization of the immature spine in normal and pathologic conditions.

Methods: Lumbar quantitative measurements from 40 T2-weighted lumbar (L1-S1) MRIs of children from the general population were taken twice by 1 tester for intratester reproducibility and compared with the same measurements (L4-S1) of the same MRIs taken by a second tester for intertester reproducibility. The following parameters were measured using the iQ-VIEW system (IMAGE Information Systems Ltd., version 1.2.2, Plauen, Germany): Linear measurements--zygoappophyseal facets and interfacet widths, and vertebral body (VB), pedicle and intervertebral discs heights, widths, and lengths. Angular measurements--zygoappophyseal tranverse superior facet angles, sagittal VB, and disc wedging, lumbar lordosis, and sacral inclination. Statistical analysis included the concordance correlation coefficient (CCC), and Bland and Altman's limits of agreement (LOA).

Results: A total of 6160 measurements were analyzed. Good to excellent intratester reproducibility (0.75 < or =CCC < or =0.99) was indicated for all linear measurements (except for pedicle heights) (LOA: -4.76; 5.74 mm) and for angular measurements related to the facets orientation, lumbar lordosis, and sacral inclination (LOA: 11.22 degrees ; 12.34 degrees). VB and disc sagittal wedging manifested poor intratester reproducibility (CCC < 0.75). Good to excellent intertester reproducibility was indicated for 70% of all linear measurements (LOA: -5.83; 4.59 mm) and for all angular measurements in the lumbar spine (L4-S1) (LOA: -9.96 degrees ; 14.62 degrees) except for VB and disc sagittal wedging at L5 (CCC < 0.75).

Conclusion: Quantitative lumbar MRI measurements in children from the general population were found to be reproducible indicating a good visualization of immature vertebral anatomic margins on MRIs and an accurate definition of the measurement protocol.

MeSH terms

  • Adolescent
  • Child
  • Female
  • Humans
  • Lumbar Vertebrae / anatomy & histology*
  • Magnetic Resonance Imaging*
  • Male
  • Observer Variation
  • Reproducibility of Results