| 蒋芙蓉,赵静文,刘翔,贾洪博,石蕴玉,汤显,宋家琳.漂移迭代搜索算法提取高频超声声像图肝包膜线并量化评估肝硬化[J].中国医学影像技术,2022,38(5):747~752 |
| 漂移迭代搜索算法提取高频超声声像图肝包膜线并量化评估肝硬化 |
| Drift iterative search algorithm for extracting liver capsule line on high-frequency ultrasonograms and quantitative evaluating liver cirrhosis |
| 投稿时间：2021-07-02 修订日期：2022-01-16 |
| DOI：10.13929/j.issn.1003-3289.2022.05.027 |
| 中文关键词: 肝包膜 肝硬化 迭代算法 超声检查 |
| 英文关键词:liver capsule liver cirrhosis iterative algorithm ultrasonography |
| 基金项目:上海市科学技术委员会科研计划(19ZR1421500)。 |
|
| 摘要点击次数: 2444 |
| 全文下载次数: 652 |
| 中文摘要: |
| 目的 观察漂移迭代搜索算法提取高频超声声像图肝包膜线并用于量化评估肝硬化病情的价值。方法 纳入47例肝硬化患者,按病情将其分为轻度组(n=18)、中度组(n=16)及重度组(n=13);并选取20名健康体检者作为对照组。基于肝脏声像图,采用漂移迭代搜索算法自动定位并提取肝包膜线,观察其连续性、平滑性及腹腔积液;分析漂移迭代搜索算法提取肝包膜线的价值,以及其用于评估肝硬化病情的效能。结果 67组图像中肝包膜线均被精确定位并提取,对照组及轻、中、重度组的提取准确率分别为100%(20/20)、100%(18/18)、93.75%(15/16)及92.31%(12/13)。4组肝包膜线的连续性及平滑性差异均有统计学意义(P均＜0.05),两两比较差异亦有统计学意义(P均＜0.05);腹腔积液仅见于重度组。结论 漂移迭代搜索算法可精确定位并提取高频超声声像图肝包膜线,有助于判断肝硬化严重程度。 |
| 英文摘要: |
| Objective To observe the value of drift iterative search algorithm for extracting liver capsule line on high-frequency ultrasonograms and quantitative evaluating liver cirrhosis.Methods Forty-seven patients with liver cirrhosis were enrolled and divided into mild group (n=18), moderate group (n=16) and severe group (n=13), while 20 healthy subjects were selected as controls (control group). Based on liver ultrasonograms, the drift iterative search algorithm was used to automatically localize and extract the liver capsule line, and the continuity, smoothness and ascites were observed. The value of the drift iterative search algorithm for extracting liver capsule line and its efficacy for quantitative evaluating liver cirrhosis were analyzed.Results The liver capsule line of ultrasonograms of 67 groups were all precisely localized and extracted, the extraction accuracy rate of control group and mild, moderate, severe groups was 100% (20/20), 100% (18/18), 93.75% (15/16) and 92.31% (12/13), respectively. There were significant differences of continuity and smoothness of the extracted liver capsule lines among 4 groups (both P＜0.05), as well as between each 2 groups (all P＜0.05). Ascites was only detected in severe group.Conclusion Drift iterative search algorithm could accurately localize and extract the liver capsule line on high-frequency ultrasonograms, hence being helpful for evaluating the severity of liver cirrhosis. |
| 查看全文 查看/发表评论 下载PDF阅读器 |
|
|
|
