Fan Shun-Wu, Fang Xiang-Qian, Liu Yun-Jian, Yu He-Jun, Lu Yin-Jiang, Liu Chao
Zhejiang University Faculty of Medicine, Sir Run Run Shaw Hospital, Department of Orthopedics, Zhejiang Province, China.
Wenzhou Medical College, Lishui Central Hospital, Department of Orthopedics, Zhejiang Province, China.
Acta Orthop Traumatol Turc. 2015;49(6):606-13. doi: 10.3944/AOTT.2015.14.0267.
The aim of this study was to evaluate the reliability of high intensity zone (HIZ) and to assess discrepancy in the interpretation, as well as investigate the effects of parameters of HIZ on interobserver variation.
Four spine surgeons made independent observations on lumbar magnetic resonance imaging (MRI) from 207 consecutive patients from 3 institutions. The κ statistic was used to characterize inter- and intraobserver reliability for visual assessments of HIZ. The corresponding MRI was provided to 2 additional spine surgeons for quantitative measurements. The parameters of HIZ, including signal intensity (SI) and area ratio (HIZ%), were used to assess the interobserver variation of HIZ.
The overall interobserver agreement for visual assessments was substantial (κ=0.62 at L4-5 and 0.61 at L5-S1), and intraobserver agreement was excellent (κ=0.84 at L4-5 and 0.86 at L5-S1). Of 93 observed HIZ, 17 instances (18.3%) were agreed upon by all visual observers. The SI with full agreement was significantly brighter than all the others (p<0.01). The HIZ% with 2 agreements was significantly smaller than those with 4 agreements (p=0.04) and 3 agreements (p=0.03). Although fewer observers with consensus were associated with smaller HIZ%, the difference was not significant (p>0.05).
The reliability in the interpretation of HIZ was sufficient for spine surgeons with differing levels of experience. This study highlighted that signal intensity was the primary cause of variability in visual observation.
本研究旨在评估高强度区（HIZ）的可靠性，评估其解读差异，并研究HIZ参数对观察者间差异的影响。
四位脊柱外科医生对来自3家机构的207例连续患者的腰椎磁共振成像（MRI）进行独立观察。κ统计量用于描述观察者间和观察者内对HIZ视觉评估的可靠性。相应的MRI提供给另外两位脊柱外科医生进行定量测量。HIZ的参数，包括信号强度（SI）和面积比（HIZ%），用于评估HIZ的观察者间差异。
视觉评估的总体观察者间一致性较高（L4 - 5节段κ = 0.62，L5 - S1节段κ = 0.61），观察者内一致性极佳（L4 - 5节段κ = 0.84，L5 - S1节段κ = 0.86）。在观察到的93个HIZ中，所有视觉观察者对17例（18.3%）达成了一致。完全一致的SI明显比其他所有情况更亮（p < 0.01）。有两项一致意见的HIZ%明显小于有四项一致意见的（p = 0.04）和有三项一致意见的（p = 0.03）。尽管达成共识的观察者越少，HIZ%越小，但差异不显著（p > 0.05）。
对于经验水平不同的脊柱外科医生，HIZ解读的可靠性是足够的。本研究强调信号强度是视觉观察中变异性的主要原因。
