Department of Clinical Services, Physiotherapy Section, St. Olav University Hospital, Trondheim, Norway.
Dev Med Child Neurol. 2010 Aug;52(8):773-8. doi: 10.1111/j.1469-8749.2010.03629.x. Epub 2010 Feb 24.
The aim of this study was to investigate the predictive value of a computer-based video analysis of the development of cerebral palsy (CP) in young infants.
A prospective study of general movements used recordings from 30 high-risk infants (13 males, 17 females; mean gestational age 31wks, SD 6wks; range 23-42wks) between 10 and 15 weeks post term when fidgety movements should be present. Recordings were analysed using computer vision software. Movement variables, derived from differences between subsequent video frames, were used for quantitative analyses. CP status was reported at 5 years.
Thirteen infants developed CP (eight hemiparetic, four quadriparetic, one dyskinetic; seven ambulatory, three non-ambulatory, and three unknown function), of whom one had fidgety movements. Variability of the centroid of motion had a sensitivity of 85% and a specificity of 71% in identifying CP. By combining this with variables reflecting the amount of motion, specificity increased to 88%. Nine out of 10 children with CP, and for whom information about functional level was available, were correctly predicted with regard to ambulatory and non-ambulatory function.
Prediction of CP can be provided by computer-based video analysis in young infants. The method may serve as an objective and feasible tool for early prediction of CP in high-risk infants.
本研究旨在探讨基于计算机的视频分析对早期婴儿脑瘫（CP）发展的预测价值。
前瞻性研究使用一般运动记录，纳入 30 名高危婴儿（13 名男性，17 名女性；平均胎龄 31 周，标准差 6 周；范围 23-42 周），在纠正胎龄 10-15 周时应出现不安运动。使用计算机视觉软件对记录进行分析。从后续视频帧之间的差异中得出运动变量，用于定量分析。CP 状态在 5 岁时报告。
13 名婴儿发展为 CP（8 例偏瘫，4 例四肢瘫，1 例运动障碍；7 例可步行，3 例不可步行，3 例功能未知），其中 1 例有不安运动。运动质心的变异性在识别 CP 方面具有 85%的敏感性和 71%的特异性。通过将其与反映运动幅度的变量相结合，特异性增加到 88%。10 名 CP 患儿中的 9 名（其中 9 名患儿的功能水平信息可用）的步行和非步行功能得到了正确预测。
基于计算机的视频分析可对早期婴儿 CP 进行预测。该方法可能成为高危婴儿 CP 早期预测的一种客观可行的工具。
