We found a match
Your institution may have access to this item. Find your institution then sign in to continue.
- Title
Analysis Of Reliability And Reproducibility Of Manually And Automatically Measured Parameters In Computed Tomography Imaging Of The Injured Spine.
- Authors
Vetter, Sven; Yiheng Chen; Shiyao Liao; Grützner, Paul A.; Matschke, Stefan; Kreinest, Michael
- Abstract
Introduction: Advancements in imaging software products provide increasing automatic measurements of parameters characterizing vertebral body fractures such as the local sagittal angle, scoliosis angle and the disc height. The consistency of these automatically measured parameters compared to manually measured parameters is still unclear. Thus, the aim of the current study was to analyze the reliability of automatically measured parameters in computed tomography imaging of the injured spine. Material and Methods: First, reliability and reproducibility of the manual measurement (mediCAD spine 3D, Hectec, Germany) was analyzed by measuring the above mentioned parameters by three experienced observers at two time points (with an interval of 7 days). Inter-observer reliability as well as intra-observer reproducibility was analyzed by intraclass-correlation. Only if reliability and reproducibility of the manual measurements were good to excellent1,2 (intraclass correlation coefficient (ICC) = 0.70 - 1.00), comparison to the automatic measurement (mediCAD spine 3D) was performed by t-test. Statistical analysis was performed with SPSS (IBM, USA). Results: Inter-observer reliability was excellent for manually measured local sagittal angle (ICC = 0.97) and good for manually measured scoliosis angle (ICC = 0.80). Furthermore, intra-observer reproducibility was excellent for manually measured local sagittal angle (ICC = 0.97) and good for manually measured scoliosis angle (ICC = 0.81). Compared to the manual measurement of the local sagittal angle no significant difference was seen in the automatic measurement (3.2 ± 12.8° vs. 2.6 ± 12.1°; p = 0.208). However, automatic measurement of scoliosis angle differs significantly from manually measured values (3.7 ± 3.1° vs. 4.8 ± 3.6°; p = 0.023). Analyzing the manual measurements of the disc height, inter-observer reliability and intra-observer reproducibility was poor (ICC = 0.61 and ICC = 0.36, respectively). Thus, comparison to automatic measurement was not performed for disc height. Conclusion: Local sagittal angle can be measured reliable by automatic and manual measurement. But other parameters such as scoliosis angle remain with an inaccuracy in the automatic measurement. Other studies3 confirm such discrepancys in the range of 20%. However, some parameters such as the disc height could not be detected reliable even with manual measurement. Difficulties correlating to the identification of anatomical landmarks (such as spondylophytes) could be one reason for this inaccuracy2.
- Publication
Global Spine Journal, 2018, Vol 8, p284S
- ISSN
2192-5682
- Publication type
Article
- DOI
10.1177/2192568218771072