Teplyi Valerii, Grebchenko Kateryna
Department of Surgery #2, O.O. Bogomolets National Medical University, Kyiv, Ukraine.
Skin Res Technol. 2019 Mar;25(2):194-199. doi: 10.1111/srt.12634. Epub 2018 Oct 17.
The application of new techniques of the scars' correction requires the objective evaluation of their vascularization.
To evaluate the effectiveness of digital program ImageJ in assessing neovascularization of pathologic scars.
In this cross-sectional study, a total of 25 patients with pathologic scars were enrolled. Vessel selection in the first set of digital images of their scars was performed by computer processing started from thresholding with subsequent manual correction. In the second set of the same pictures, Vessel Analysis plugin was used. Comparison of both approaches was performed by three independent investigators. The time required for images processing was measured.
The average time that image processing and calculation have taken in the first group (753.3 ± 88.02 seconds) was statistically longer (P < 0.0001) than in the second one (358.1 ± 105.91 seconds). Independent investigators scored the precision of vessel selection in the first group as 80.4 ± 9.82, in the second group as 72.6 ± 10.53 (P < 0.0001). Kolmogorov-Smirnov test demonstrated higher precision of vessel selection by method that involves manual correction (P < 0.001). The results of Vascular Density measurements were obviously overestimated in the second group. More expedient looks calculation of the Vascular Length Density: ratio of skeletonized vasculature area to total area. Skeletonization avoids overestimation of Vascular Density, but the density of the vessel mesh can be determined.
Computer processing of the scars' digital photographs using ImageJ software gives cheap, technically easy and not cumbersome way of superficial scars' vascularization objectifying. Vessel selection with subsequent manual correction has advantage of higher precision.
瘢痕矫正新技术的应用需要对其血管化进行客观评估。
评估数字程序ImageJ在评估病理性瘢痕新生血管化方面的有效性。
在这项横断面研究中，共纳入25例病理性瘢痕患者。对其瘢痕的第一组数字图像进行血管选择时，先通过阈值处理启动计算机处理，随后进行手动校正。在同一组图片的第二组中，使用血管分析插件。由三名独立研究人员对两种方法进行比较。测量图像处理所需的时间。
第一组图像处理和计算的平均时间（753.3±88.02秒）在统计学上比第二组（358.1±105.91秒）更长（P<0.0001）。独立研究人员对第一组血管选择的精度评分为80.4±9.82，第二组为72.6±10.53（P<0.0001）。Kolmogorov-Smirnov检验表明，涉及手动校正的方法在血管选择上具有更高的精度（P<0.001）。第二组血管密度测量结果明显高估。血管长度密度的计算看起来更方便：骨骼化血管区域与总面积的比值。骨骼化可避免血管密度的高估，但可确定血管网的密度。
使用ImageJ软件对瘢痕数字照片进行计算机处理，为浅表瘢痕血管化的客观化提供了一种廉价、技术上简单且不繁琐的方法。后续进行手动校正的血管选择具有更高精度的优势。
