| Objective: Ulnar variance(UV) is defined as the relative length of the distal ulna to that of the distal end of radius. According to previous literature review [1-2], several methods of measuring ulnar variance have been proposed. It remains unclear whether they accurately represent the true ulnar variance of the patient. The purpose of this study was to compare the agreement and reliability of ulnar variance measurements on posteroanterior(PA) radiograph, lateral radiograph, coronal computed tomography(CT) multiplanar reconstruction, sagittal CT multiplanar reconstruction, CT three-dimensional reconstruction, coronal magnetic resonance imaging(MRI), ignoring the articular cartilage(MRI bone-bone), and coronal MRI, accounting for articular cartilage(MRI cartilage- cartilage).Methods:1 Study objects: We assessed 32 volunteers and two wrists in each individual. There were 32 men with a mean age of 19.78(range 18–25) years, right handedness. They had no previous history of acute of the wrist or forearm fracture history.2 Examination methods: All the subjects under PA radiography,lateral radiography, CT and MRI examination. CT imaging was performed on a 320-slice dynamic volume CT scanner made by Toshiba. Sagittal and coronal multiplanar reconstructions were generated. After getting the CT image date, we use the software Mimics 17.0 to develop a three-dimensional model. Using GE Signa HDxt 3.0 T Superconducting type magnetic resonance manchines and knee joint surface coil.The imaging protocol consisted of three dimension fat saturation spoiled gradient echo T1 weighted image(3D-FSPGR T1W), and then send the MRI image to post-process workstation, multiplanar reconstructions were generated. The CT and MRI scans were performed with the wrist in neutral rotation.3 Measurement index: Measure the ulnar variance on 7 images of each wrist: PA radiography, lateral radiography, coronal CT multiplanar reconstruction, sagittal CT multiplanar reconstruction, CT three-dimensional reconstruction, coronal MRI, ignoring the articular cartilage(MRI bone-bone), and coronal MRI, accounting for articular cartilage(MRI cartilage-cartilage), respectively.4 Data analysis: The ulnar variance of each wrist was measured independently using 7 methods by the investigator 1 and investigator 2. The ulnar variance was measured repeatedly by investigator 1 after two weeks.5 Statistical analysis: All statistical analysis was performed using SPSS19.0 and Med Calc 9.6 software [3-4]. If the measurement date meeted the normality, the results were expressed as ±s. Otherwise, expressed by M(QR). The differences in ulnar variance measured with different methods were evaluated with one-way ANOVA or Kruskal-Wallis H analysis of variances. Intraclass correlation coefficient(ICC) was used to assess the repeatability of seven methods, and Agreement between raters and methods were compared using Bland-Altman plots. The significance level was set at P<0.05.Results:1 There was no significant difference in ulnar variance measured by seven methods(P>0.05).2 To assess the intra-rater reliability of UV measurements by one observer with 7 methodsThe intra-rater ICC and 95% confidence interval(CI) were 0.896,(0.829 ~ 0.937); 0.824,(0.814 ~ 0.894); 0.806,(0.681 ~ 0.882); 0.846,(0.747 ~ 0.907); 0.998,(0.997 ~ 0.999); 0.929,(0.878 ~ 0.959); 0.958,(0.927 ~ 0.976). Which measured separately by PA radiography, lateral radiograph, coronal CT multiplanar reconstruction, sagittal CT multiplanar reconstruction,CT three-dimensional reconstruction, coronal MRI(bone-bone), coronal MRI(cartilage-cartilage). All the methods of measurements about intra-rater ICC of UV were greater than 0.8, which means that the reliability is excellent. And among all the methods the CT three-dimensional reconstruction was the highest, and the 95% CI was narrower than the other six methods, which means that the repeatability of UV measured by 7 methods(handled by one observer) was good, but the CT three-dimensional reconstruction was better, so the CT three-dimensional reconstruction is more reliable on the measurement of UV.3 To compare the inter-rater reliability of UV measurements by different observers with 7 methodsThe inter-rater ICC and 95% CI were 0.810,(0.686~0.885); 0.802,(0.730 ~ 0.852); 0.882,(0.806 ~ 0.928); 0.886,(0.813 ~ 0.931); 0.997,(0.995 ~ 0.998); 0.941,(0.899 ~ 0.966); 0.906,(0.838 ~ 0.945). Which measured separately by PA radiography, lateral radiograph, coronal CT multiplanar reconstruction, sagittal CT multiplanar reconstruction, CT three-dimensional reconstruction, coronal MRI(bone-bone), coronal MRI(cartilage-cartilage). All the methods of measurements about inter-rater ICC of UV were greater than 0.8, which means that the reliability is excellent. And among all the methods the CT three-dimensional reconstruction was the highest, and the 95% CI was narrower than the other six methods, which means that the repeatability of UV measured by 7 methods(handled by one observer) was good, but the CT three-dimensional reconstruction was better, so the CT three-dimensional reconstruction is more reliable on the measurement of UV.4 To compare the agreement of UV measurements by different observer with same method according to the Bland-Altman plotsThe mean difference and the 95% limit of agreement were-0.17 mm,(-2.20~1.85)mm;-0.16 mm,(-2.10~1.75)mm; 0.06 mm,(-1.67~1.79)mm;-0.07 mm,(-1.83~1.69)mm; 0.05 mm,(-0.24~0.34)mm; 0.05 mm,(-1.16~1.07)mm;-0.02 mm,(-1.37~1.33)mm. Which measured separately by PA radiography, lateral radiograph, coronal CT multiplanar reconstruction, sagittal CT multiplanar reconstruction, CT three-dimensional reconstruction, coronal MRI(bone-bone), coronal MRI(cartilage-cartilage). The CT three-dimensional reconstruction hadthe smallest limit of agreement, and their range were within(-1~1)mm, besides that, the mean difference was approaching to 0, which means that the agreement of CT among the 7 methods was the highest.5 To assess the agreement of UV measurements before and after 2 weeks by the same observer according to the Bland-Altman plotsThe mean difference and the 95% limit of agreement were 0.17 mm,(-1.34~1.69)mm;-0.14 mm,(-1.67~1.40)mm;-0.05 mm,(-2.03~1.94)mm;-0.28 mm,(-2.13 ~ 1.58)mm; 0mm,(-0.23 ~ 0.24)mm;-0.01 mm,(-1.23 ~1.21)mm;-0.11 mm,(-1.00~0.79)mm. Which measured separately by PA radiography, lateral radiograph, coronal CT multiplanar reconstruction, sagittal CT multiplanar reconstruction, CT three-dimensional reconstruction, coronal MRI(bone-bone), coronal MRI(cartilage-cartilage). The CT three-dimensional reconstruction had the smallest limit of agreement, and their range were within(-1~1)mm, besides that, the mean difference was 0, which means that the agreement of CT among the 7 methods was the highest.6 To compare the agreement of UV by any two of the 7 methods according to the Bland-Altman plots with combined data of observer 1 and 2.6.1 The mean difference and the 95% limit of agreement were 0.43 mm,(-1.85 ~ 2.71)mm for PA radiography and lateral radiography; 0.38 mm,(-1.93 ~ 2.70)mm for PA radiography and coronal CT multiplanar reconstruction; 0.28 mm,(-1.90~2.46)mm for PA radiography and sagittal CT multiplanar reconstruction;-0.16 mm,(-2.05~1.74)mm for PA radiography and CT three-dimensional reconstruction; 0.39 mm,(-2.17~2.95)mm for PA radiography and coronal MRI(bone-bone); 0.46 mm,(-2.08~3.01)mm for PA radiography and coronal MRI(cartilage-cartilage).6.2 The mean difference and the 95% limit of agreement were-0.05 mm,(-2.30 ~ 2.21)mm for lateral radiography and coronal CT multiplanar reconstruction;-0.16 mm,(-2.47~2.15)mm for lateral radiography and sagittal CT multiplanar reconstruction;-0.59 mm,(-2.55 ~ 1.37)mm for lateral radiography and CT three-dimensional reconstruction;-0.14 mm,(-2.70 ~ 2.43)mm for lateral radiography and coronal MRI(bone-bone);-0.07 mm,(-2.58~2.45)mm for lateral radiography and coronal MRI(cartilage-cartilage). 6.3 The mean difference and the 95% limit of agreement were-0.11 mm,(-1.44~1.66)mm for coronal CT multiplanar reconstruction and sagittal CT multiplanar reconstruction;-0.54 mm,(-2.06 ~ 0.98)mm for coronal CT multiplanar reconstruction and CT three-dimensional reconstruction; 0.06 mm,(-2.35 ~ 2.48)mm for coronal CT multiplanar reconstruction and coronal MRI(bone-bone); 0.14 mm,(-2.28 ~ 2.56)mm for coronal CT multiplanar reconstruction and coronal MRI(cartilage-cartilage).6.4 The mean difference and the 95% limit of agreement were-0.43 mm,(-1.89 ~ 1.03)mm for sagittal CT multiplanar reconstruction and CT three-dimensional reconstruction; 0.16 mm,(-2.15~2.48)mm for sagittal CT multiplanar reconstruction and coronal MRI(bone-bone); 0.24 mm,(-2.07~2.54)mm for sagittal CT multiplanar reconstruction and coronal MRI(cartilage-cartilage).6.5 The mean difference and the 95% limit of agreement were 0.57 mm,(-1.45 ~ 2.59)mm for CT three-dimensional reconstruction and coronal MRI(bone-bone); 0.64 mm,(-1.49 ~ 2.78)mm for CT three-dimensional reconstruction and coronal MRI(cartilage-cartilage).6.6 The mean difference and the 95% limit of agreement were 0.07 mm,(-1.09~1.23)mm for coronal MRI(bone-bone) and MRI(cartilage-cartilage).The 95% limit of agreement range of 7 methods which compared by any of each other were out of the(-1~1)mm, and the mean difference was far from the 0. And these results were not acceptable on clinic, caused the agreements of any of each other were not as good as we thought.Conclusions: To measure ulnar variance, there had 7 methods which were PA radiography, lateral radiography, coronal CT multiplanar reconstruction, sagittal CT multiplanar reconstruction, CT three-dimensional reconstruction, coronal MRI(bone-bone), coronal MRI(cartilage-cartilage), although any one of the 7 methods had considerable disagreement with one another. Intra-class correlation coefficients demonstrated excellent inter-rater reliability and intra-rater reliability for each method, all the methods mentioned before could be used on measuring the UV. The CT three-dimensional reconstruction method had the highest reliability and agreement, which more accurate compared with other 6 methods. |
PDF Full Text Request