魏义圆,史燕杰,李晓婷,孙应实.基于多层螺旋CT简易评分模型鉴别局灶型食管癌与食管平滑肌瘤[J].中国医学影像技术,2020,36(8):1197~1201
基于多层螺旋CT简易评分模型鉴别局灶型食管癌与食管平滑肌瘤
Simple scoring model based on multi-slice spiral CT in differentiation between focal esophageal carcinoma and esophageal leiomyoma
投稿时间:2020-03-24  修订日期:2020-06-08
DOI:10.13929/j.issn.1003-3289.2020.08.018
中文关键词:  食管肿瘤  平滑肌瘤  体层摄影术,X线计算机
英文关键词:esophageal neoplasms  leiomyoma  tomography,X-ray computed
基金项目:
作者单位E-mail
魏义圆 北京大学肿瘤医院暨北京市肿瘤防治研究所医学影像科 恶性肿瘤发病机制及转化研究教育部重点实验室, 北京 100142  
史燕杰 北京大学肿瘤医院暨北京市肿瘤防治研究所医学影像科 恶性肿瘤发病机制及转化研究教育部重点实验室, 北京 100142  
李晓婷 北京大学肿瘤医院暨北京市肿瘤防治研究所医学影像科 恶性肿瘤发病机制及转化研究教育部重点实验室, 北京 100142  
孙应实 北京大学肿瘤医院暨北京市肿瘤防治研究所医学影像科 恶性肿瘤发病机制及转化研究教育部重点实验室, 北京 100142 sys27@163.com 
摘要点击次数: 1193
全文下载次数: 559
中文摘要:
      目的 观察多层螺旋CT(MSCT)简易评分模型鉴别局灶型食管癌与食管平滑肌瘤的效能。方法 对46例局灶型食管癌和21例食管平滑肌瘤患者术前均行胸部增强CT检查,观察二者病变位置、密度、大小、病灶MPR最长径/轴位最厚厚度比值、强化程度及方式、瘤-空气界面、瘤周脂肪间隙及肿大淋巴结;对存在差异的CT征象进行Logistic回归分析,建立简易评分模型,采用受试者工作特征曲线(ROC)其鉴别诊断效能。结果 二者肿瘤MPR最长径/轴位最大截面厚度比值[优势比(OR)=0.154,95% CI(0.033,0.722))、肿瘤增强CT值[OR=0.928,95% CI(0.866,0.994)]及瘤-空气界面[OR=0.028,95% CI(0.004,0.184)]存在显著差异(P均<0.05)。据此建立简易评分模型,肿瘤增强CT值>65.5 HU赋值1分,肿瘤MPR最长径/轴位最大截面厚度比值>1.61为2分,瘤-空气界面不光滑为4分,以2.5分为最佳阈值,≥ 2.5分诊断为局灶型食管癌,简易评分模型曲线下面积(AUC)为0.945[95% CI(0.891,0.999)],其鉴别诊断效能优于单一特征(P<0.05)。结论 MSCT简易评分模型有助于鉴别局灶型食管癌与食管平滑肌瘤,可显著提高诊断效能。
英文摘要:
      Objective To explore the efficiency of simple scoring model based on multi-slice spiral CT (MSCT) for differential diagnosis of focal esophageal carcinoma and esophageal leiomyoma. Methods Totally 46 patients with focal esophageal carcinoma and 21 with esophageal leiomyoma who underwent preoperative chest enhanced CT were retrospectively analyzed. The lesion's location, density, size (the thickest wall on axial position and the longest diameter on multi-planar reconstruction), the ratio of the longest diameter on MPR to the thickest wall on axial position, enhancement degree, tumor-air surface, peritumoral fat space and enlarged lymph nodes were assessed, and Logistic regression analysis was used to select MSCT signs significantly different between two diseases to establish a simple scoring model. Receiver operating characteristic curve (ROC) was used to evaluate the diagnostic efficiency of this model. Results The ratio of the longest diameter on MPR to the thickest wall on axial position (odds ratio[OR]=0.154, 95%CI[0.033, 0.722]), enhanced CT value (OR=0.928, 95%CI[0.866, 0.994]), tumor-air surface (OR=0.028, 95%CI[0.004, 0.184]) were significantly different between focal esophageal carcinoma and esophageal leiomyoma(P<0.05), and a simple scoring model then was established. The enhanced CT value of tumor >65.5 HU was defined as 1 point, the ratio of the longest diameter of MPR to the thickest wall on axial position >1.61 as 2 points, and the unsmooth tumor-air surface as 4 points. Taken 2.5 points as the cutoff value for diagnosing esophageal carcinoma, the area under curve (AUC) of this model was 0.945 (95%CI[0.891, 0.999]), better than that according to single features (P<0.05). Conclusion The simple scoring model based on MSCT was helpful to differentiation of focal esophageal carcinoma from esophageal leiomyoma and improving diagnostic efficiency.
查看全文  查看/发表评论  下载PDF阅读器