Reproducibility of Magnetic Resonance Diffusion Weighted Imaging In Axial Spondyloarthritis Patients and Healthy Subjects.

Background/Purpose: Diffusion weighted imaging (DWI) is a magnetic resonance imaging (MRI) technique where the image contrast is dependent on the diffusion of the extracellular free water molecules. DWI is widely used in oncology imaging, but only few papers address DWI in spondyloarthritis (SpA). From DWI the apparent diffusion coefficient (ADC) can be calculated and hence the diffusion in a region of interest (ROI) can be quantified. The water diffusion reflects the cellularity and thus it may be used as a quantification of inflammatory cells. The aim of the study was to measure the reproducibility of DWI in SpA patients and healthy subjects.

Methods: 25 SpA patients (13 females, mean age 36.1 years (SD 9.6); 12 males, mean age 42.3 (SD10.8) and 24 healthy subjects (13 females, mean age 42.7 (SD 13.2); 11 males, mean age 45.3 (SD 7.5)) were MRI examined at 1.5T two times with a mean interval of 6.8 days (SD 0.9). A 5mm (gap 1.2mm) sagittal single-shot echo-planar imaging DW sequence with four b values (0;50;500;800) and a resolution of 2mm x 1.65mm was performed on Th6 to L5. ADC maps were calculated based on all b values. The MRIs were anonymized and read in random order without information on time point and clinical data. On ADC maps 50mm² circular ROIs were placed in each vertebral corner just inside cortex. ADC was measured on the sagittal slice in the middle of the spine and on the adjacent slice to the right and left to the middle. ADCs from the four ROIs were pooled. Another ROI was placed in the right and left pedicle and in the spinous process from L1 to L5. The inter-reader variability was measured by intra-class correlation coefficient (ICC). Intra-reader variability was measured by a second reading of the examinations at time point (TP) 2.

Results: Overall ICC of the vertebral bodies between TP1 and TP2 was 0.80 (95% CI: 0,77-0.83). At TP1, ICC between the right and middle slice was 0.89 (95% CI: 0.87-0.90) and between left and middle slice 0.89 (95% CI: 0.87-0.90). Intra-reader reliability was 0.94 (95% CI: 0.93-0.95). For the right pedicle ICC was 0.44 (95% CI: 0.33-0.54), for the left pedicle 0.49 (95% CI: 0.39-0.58), for the spinous processes 0.20 (95% CI: 0.07-0.32). Table 1 provides ICCs for each vertebral body.

Table 1. Inter-reader and intra-reader ICCs for each vertebral body.

Table 1: ICC (95% CI) measurements per vertebral body.
Conclusion: MR DWI is overall a reproducible imaging sequence performed in the sagittal plane. The reproducibility is better in the lumbar spine compared to lower thoracic spine. Pedicles and spinous proceses are not reliably imaged by DWI.
References: Gaspersic et al. Skeletal Radiol. 2008;37:123-31. Koh et al. AJR Am J Roentgenol. 2007;188:1622-35