基于成像的转移性淋巴结体积测量

Imaging-based metastatic lymph node volume measurement

目的评价一种半自动淋巴结体积测量方法对转移性颈部淋巴结体积测量的应用价值。方法18例病理证明的鼻咽癌患者于治疗前行MRT2WI,2名头颈部放射医师使用人工勾画边界法和基于区域变形的半自动图像分割算法测量患者增大的颈淋巴结体积,每一组MR图像由这2名放射医师独立处理。测量结果被记录并进行操作者间差异和可靠性分析。结果共测量45个增大的淋巴结或者淋巴结块。人工勾画所得的淋巴结体积和用区域变形方法所得淋巴结体积间差异无统计学意义(P=0.743),2名放射医师所得的体积数据之间差异无统计学意义(P=0.967)。另外,在操作者和体积测量方法之间没有显著的相关性(P=0.984)。同时,两种方法的操作者间可靠性差异无统计学意义(认同指数分别为0.88±0.09,0.90±0.07;P=0.181)。结论基于区域变形分割算法的半自动测量方法可以作为一种有用的工具来评价转移性淋巴结的体积,应用于淋巴结分期或对治疗效果进行监测。

Objective To evaluate a semi-automated method of volume measurement for metastatic cervical lymph nodes. Methods Two head ＆ neck radiologists measured the volumes of enlarged cervical lymph nodes from pretreatment T2-weighted coronal MR images of 18 patients with histology-proven nasopharyngeal carcinoma. Each MR data set was independently processed by two radiologists using manual tracing and semi-automated, region deformation （RD）-based segmentation algorithm. The data were recorded for analysis of inter-operator variance and inter-observer reliability. Results A total of 45 enlarged lymph nodes （or lymph node masses） were analyzed. There was no significant difference between the manually traced volumes and RD segmented volumes （ P = 0. 743 ） and no significant difference existed between the volumes obtained from both radiologists （ P = 0. 967 ）. In addition, there was no correlation between operators and volume measurement methods （ P = 0. 984）. Furthermore, no significant difference of inter-operator reliability existed between the two methods （ AIM ,0. 88 ±0. 09. AIRD,0. 90 ±0. 07. P =0. 181, respectively）. Conclusion This study validates the use of semi-automated volume measurement based on RD segmentation algorithm for the analysis of cervical nodal volume. This method could be a potentially useful tool for the evaluation of metastatic nodal volume either as a classification method or monitoring treatment response.