Lee Hennie, Brekelmans Geert J F, Visser Leo H
Department of Neurology and Clinical Neurophysiology, St. Elisabeth Ziekenhuis, Tilburg, The Netherlands; Department of Neurology, Gemini Ziekenhuis, Den Helder, The Netherlands.
Department of Neurology and Clinical Neurophysiology, St. Elisabeth Ziekenhuis, Tilburg, The Netherlands.
Clin Neurophysiol. 2016 Jan;127(1):874-879. doi: 10.1016/j.clinph.2015.03.019. Epub 2015 Apr 11.
Quantitative assessment of nerve echogenicity using the MaxEntropy, RenyiEntropy and Yen methods has shown promise in the assessment of ulnar neuropathy at the elbow. The purpose of this study was to determine the clinical usefulness of quantitative echogenicity measurement when using these three automatic thresholding methods in the evaluation of patients with common fibular (CF) neuropathy.
High-resolution ultrasonography (HRUS) images of the CF nerve from 45 patients with CF neuropathy were compared to 42 healthy controls.
With all three methods patients with CF neuropathy had significantly higher mean hypoechoic fraction than healthy controls at the level of the fibular head (FH): MaxEntropy 87.4% versus 71.9% (p<0.001), RenyiEntropy 84.7% versus 68% (p<0.001) and Yen 87.6% versus 68.2% (p<0.001). Patients with CF neuropathy could be differentiated from healthy controls at a hypoechoic fraction cut-off point of 82% (MaxEntropy), 80% (RenyiEntropy) and 80% (Yen) with a sensitivity of 82%, 82% and 84% respectively, and a specificity of 93%, 95% and 83% respectively. Similar results were found for sonographic measurements above the FH.
Quantitative assessment of nerve echogenicity using the MaxEntropy, RenyiEntropy and Yen methods can be used to distinguish between patients with CF neuropathy and healthy controls with high sensitivity and specificity.
These three proven automatic thresholding methods can be used to assess nerve echogenicity in future studies. Quantitative echogenicity assessment with HRUS shows promise for the future as a potential diagnostic tool in daily clinical practice.
运用最大熵法、雷尼熵法和延森方法对神经回声性进行定量评估，已显示出在评估肘部尺神经病变方面具有前景。本研究的目的是确定在评估腓总神经（CF）病变患者时，使用这三种自动阈值化方法进行定量回声性测量的临床实用性。
将45例CF神经病变患者的CF神经高分辨率超声（HRUS）图像与42例健康对照者的图像进行比较。
在腓骨头（FH）水平，使用这三种方法时，CF神经病变患者的平均低回声分数均显著高于健康对照者：最大熵法为87.4%对71.9%（p<0.001），雷尼熵法为84.7%对68%（p<0.001），延森法为87.6%对68.2%（p<0.001）。CF神经病变患者与健康对照者在低回声分数截断点分别为82%（最大熵法）、80%（雷尼熵法）和80%（延森法）时可被区分，敏感性分别为82%、82%和84%，特异性分别为93%、95%和83%。在FH上方的超声测量中也发现了类似结果。
使用最大熵法、雷尼熵法和延森方法对神经回声性进行定量评估，可用于以高敏感性和特异性区分CF神经病变患者与健康对照者。
这三种经过验证的自动阈值化方法可用于未来研究中评估神经回声性。HRUS进行的定量回声性评估作为日常临床实践中的一种潜在诊断工具，未来显示出前景。
