郑欢欢,季长风,李琳,刘松,周科峰.术前CT直方图分析预测食管鳞状细胞癌EGFR表达[J].中国医学影像技术,2021,37(4):537~541
术前CT直方图分析预测食管鳞状细胞癌EGFR表达
Preoperative CT histogram analysis in predicting epidermal growth factor receptor expression of esophageal squamous cell carcinoma
投稿时间:2020-03-10  修订日期:2020-09-22
DOI:10.13929/j.issn.1003-3289.2021.04.014
中文关键词:  食管肿瘤  体层摄影术,X线计算机  直方图  ErbB受体
英文关键词:esophageal neoplasms  tomography, X-ray computed  histogram  ErbB receptors
基金项目:十三五南京市卫生青年人才培养工程(QRX17056)。
作者单位E-mail
郑欢欢 南京大学医学院附属鼓楼医院医学影像科, 江苏 南京 210008  
季长风 南京大学医学院附属鼓楼医院医学影像科, 江苏 南京 210008  
李琳 南京大学医学院附属鼓楼医院病理科, 江苏 南京 210008  
刘松 南京大学医学院附属鼓楼医院医学影像科, 江苏 南京 210008  
周科峰 南京大学医学院附属鼓楼医院医学影像科, 江苏 南京 210008 zhoukefeng1977@163.com 
摘要点击次数: 3143
全文下载次数: 722
中文摘要:
      目的 观察CT直方图参数与食管鳞状细胞癌(ESCC)表皮生长因子受体(EGFR)表达状态的相关性,评价其预测ESCC的EGFR表达状态的效能。方法 回顾性分析109例经术后病理证实ESCC患者术前胸部CT,获取其直方图参数。记录EGFR检测结果,将EGFR表达(-/+)归为低表达,EGFR表达(++/+++)归为高表达。比较EGFR高、低表达ESCC间直方图参数的差异。针对差异有统计学意义的参数行受试者工作特征(ROC)曲线分析,评价其预测EGFR表达状态的效能,分析CT直方图参数与EGFR表达的相关性。结果 109例ESCC中,40例EGFR低表达,69例高表达;其中22例EGFR表达(-),18例(+),27例(++),42例(+++)。EGFR高、低表达肿瘤的平扫CT灰度平均值(Mean)、最大值(Max)、第10、25、50、75、90百分位数(10th、25th、50th、75th、90th percentile)、峰度(kurtosis)及增强前后参数差值ΔMean、Δ50th、Δ75th、Δ90th percentile差异均有统计学意义(P均<0.05)。平扫CT肿瘤灰度75th和90th percentile预测ESCC的EGFR表达状态的诊断效能最优,曲线下面积(AUC)均为0.69。平扫CT Mean、Max、Min、5th、10th、25th、50th、75th、90th percentile、kurtosis、ΔMean及Δ5th、Δ10th、Δ25th、Δ50th、Δ75th percentile与EGFR表达呈弱至低度相关(|r|=0.20~0.37,P均<0.05),增强CT直方图参数与EGFR表达无明显相关。结论 平扫CT直方图参数可用于术前预测ESCC的EGFR表达状态,尤以75th和90th percentile效能较佳。
英文摘要:
      Objective To explore the correlations of CT histogram parameters and epidermal growth factor receptor (EGFR) expression status of esophageal squamous cell carcinoma (ESCC), and to analyze the efficacy of CT histogram parameters for predicting EGFR expression status of ESCC. Methods Preoperative chest CT of 109 patients with ESCC confirmed by postoperative pathology were retrospectively analyzed, and histogram parameters were obtained with histogram analysis. EGFR test results were recorded, and EGFR expression (-/+) were classified as low expression, while EGFR expression (++/+++) were identified as high expression. The histogram parameters of ESCC were compared between low and high EGFR expression. receiver operating characteristic (ROC) curve was used to analyze the efficacy of parameters with being statistically different between low and high EGFR expression for predicting EGFR expression status of ESCC. Then the correlations of CT histogram parameters and EGFR expressions were analyzed. Results Among 109 ESCC, 40 cases were with EGFR low expression and 69 were with high expression, including 22 cases expressed EGFR (-), 18 cases (+), 27 cases (++) and 42 cases (+++). Significant differences of the mean value (Mean), the maximum value (Max), the 10th, 25th, 50th, 75th, 90th percentile (10th, 25th, 50th, 75th, 90th percentile), kurtosis of plain CT and parameter differences before and after enhancement ΔMean, Δ50th, Δ75th, Δ90th percentile were found between ESCC with high and low expression of EGFR (all P<0.05). ROC curve results showed that the 75th and 90th percentile of plain CT had the best diagnostic efficiency in predicting high and low EGFR expression, both with the area under the curve (AUC) of 0.69. The Mean, Max, minimum value (Min), the 5th, 10th, 25th, 50th, 75th, 90th percentile, kurtosis of plain CT and ΔMean, Δ5th, Δ10th, Δ25th, Δ50th, Δ75th percentile were all correlated with EGFR expression in weak to low-grade degrees (|r|=0.20-0.37, all P<0.05), while enhanced CT histogram parameters were not significantly correlated with EGFR expressions. Conclusion The histogram parameters of plain CT could be used to preoperatively predict EGFR expression of ESCC, among which the 75th and 90th percentile of plain CT had higher diagnostic efficiency.
查看全文  查看/发表评论  下载PDF阅读器