This poster is published under an
open license. Please read the
disclaimer for further details.
Keywords:
Musculoskeletal bone, Musculoskeletal joint, Cone beam CT, MR, Comparative studies, Cost-effectiveness, Arthritides
Authors:
E. Trajcevska, C. Owen, R. Buchanan, R. B. Zwar; Melbourne/AU
DOI:
10.1594/ecr2015/C-1889
Results
Overall,
more erosions were seen on CBCT than MRI with statistically significant results in several bones.
The mean number of bone erosions detected using CBCT compared to MRI is shown in Table 1.
Table 1: Mean number of erosions per bone (mean ±standard deviation).
Clinically significant p values are highlighted by bold text.
R1: Reader 1. R2: Reader 2. MC: Metacarpal.
Reader 1 detected a total of 470 erosions on MRI and 662 erosions on CBCT. Reader 2 detected a total of 728 erosions on MRI and 429 on CBCT.
Reader 1 detected a statistically significant higher mean number of erosions using CBCT compared to MRI in 7 of the 15 bones assessed.
Reader 2 detected a statistically significant higher mean number of erosions in CBCT in 12 of the 15 bones assessed. Overall,
the mean number of erosions for all bones was statistically significantly higher on CBCT compared to MRI for both readers with a p value of 0.0001.
Of note is the differences seen when assessing the pisiform bone (Fig. 5) and the fourth and fifth metacarpal bases (Fig. 6).
Reader 1 detected a mean number of 1.91 erosions in the pisiform on CBCT and only 0.41 on MRI.
Reader 2 detected a mean number of 1.64 erosions in the pisiform on CBCT and also only 0.41 on MRI.
When assessing the base of the fifth metacarpal,
Reader 1 detected a mean number of 1.27 erosions on CBCT and only 0.59 on MRI.
Reader 2 detected a mean number of 2.59 erosions in the base of the fifth metacarpal on CBCT and only 0.77 on MRI.
Fig. 5: Pisiform: Sagittal imaging. A. PD MRI. B, C. CBCT. Comparison of imaging of the same patient.
CBCT demonstrates clear pisiform erosions which are not appreciated on MRI.
The mean bone involvement score detected using CBCT compared to MRI based on OMERACT-RAMRIS scoring system is shown in Table 2.
Table 2: Proportion of bone involvement scores (mean score per bone; mean ±standard deviation), based on the OMERACT-RAMRIS scoring system. Bone erosion was scored 1- 10, according to the proportion of bone involvement (in increments of 10%).
Clinically significant p values are highlighted by bold text.
R1: Reader 1. R2: Reader 2. MC: Metacarpal.
As demonstrated,
the percentage of bone involvement scoring is also greater overall in CBCT compared to MRI.
Reader 1 detected a statistically significant higher bone involvement score using CBCT compared to MRI in 8 of the 15 bones assessed.
Reader 2 detected a statistically significant higher bone involvement score in CBCT in 10 of the 15 bones assessed.
The large standard deviations are attributed to the cohort assessed including a wide range of disease states with bone involvement scores ranging from 0 to 9.
Seven patients demonstrated bone oedema on MRI.
Direct,
non-blinded comparison demonstrated no correlation in trabecular appearance on CBCT.
This may be a subject for further future study.