Study population
This ongoing prospective study is conducted at the University Medical Center Greifswald and was approved by the Medical Ethical Committee.
Written informed consent was obtained from each patient prior to the study.
Included were male and female patients with a clinical indicated MRI of the lower back and a DEXA.
Patients with contraindication to MRI,
pregnant women and patients with a history of lower spine surgery,
bone lesions or metabolic bone disorder were excluded from the study to prevent signal distortion (reference 1).
Until now,
a total of 48 patients (31 women and 17 men; mean age 64.8 years (95% confidence interval,
61.1-68.6 years)) were included in this study.
Magnetic Resonance Imaging (MRI)
MR imaging was performed at 1.5T on a whole-body MR imager with an integrated spine coil.
Since it is not possible to show trabecular structure itself at an 1.5T MR imager,
we used an indirect method (reference 2).
Our MR imaging protocol included a multiplanar gradient-echo pulse sequence as shown in figure 1 (TR 432ms; TE 9.0,
13.6,
18.3,
22.9,
27.6ms; flip angle 60°; 12 slices,
slice thickness 5mm; slice gap 5mm; FOV 240x240mm; matrix 256x512) for indirect lumbar trabecular thickness measurement previously reported by Wehrli et al.
(reference 3,4).
The MR protocol was slightly adapted to fit our MR imager.
The acquisition time is 1:14min.
Additionally sagittal and transverse T2-weighted and T1-weighted turbo spin echo sequences were used for clinical evaluation of any pathology.
R2* Measurements
In order to calculate the signal decay of multi echo MR sequences,
T2* maps were created (figure 2) using the MRI workstation’s software.
Afterwards mean values of each vertebra were acquired using regions of interest (ROIs) with a standardized size of 1cm2,
which were placed manually inside the vertebral marrow on the R2* map,
using an open source dicom viewer (OsiriX version 3.9).
R2* was calculated using the following formula: R2* = 1000 / T2*.
Dual-Energy-X-Ray-Absorptiometry (DEXA)
BMD of L1 to L4 in anteroposterior projection were created.
L5 was not measured because of the interference with the iliac crest.
Statistical Analysis
Correlation between R2* and BMD values was assessed using the Spearman test.
For reasons of reproducibility,
R2* measurements were done by two radiologists at two different times.
A Bland-Altmann-Plot was used to determine the intra- and interreader variability.