摘要
目的探讨基于MRI纹理特征的肺腺癌脑转移瘤诊断及对肺腺癌脑转移EGFR基因突变的预测价值。方法选取154例肺腺癌患者,根据是否脑转移将患者分成脑转移组(n=71)和非脑转移组(n=83)。比较两组患者MRI纹理特征指标,评估能量、熵、惯性矩、相关和逆差矩对肺腺癌脑转移的诊断价值。针对脑转移组患者行EGFR基因检测,将患者分成EGFR突变组(n=33)和EGFR野生组(n=38),单因素和多因素Logistic回归分析确定MRI纹理特征对肺腺癌脑转移患者EGFR基因突变的影响因素。结果脑转移组患者熵和相关等明显高于非脑转移组,脑转移组患者能量、惯性矩和逆差矩明显低于非脑转移组。根据ROC曲线可得,能量的临界值为0.00121,敏感度为69.01%,特异度为71.08%,AUC为0.801(95%CI:0.750~0.852);熵的临界值为10.42,敏感度为73.24%,特异度为69.88%,AUC为0.760(95%CI:0.707~0.812);惯性矩的临界值为7139.86,敏感度为64.79%,特异度为60.24%,曲线下面积(AUC)为0.708(95%CI:0.645~0.772);相关的临界值为0.000026,敏感度为59.15%,特异度为60.24%,AUC为0.631(95%CI:0.564~0.698);逆差矩的临界值为0.0301,敏感度为74.65%,特异度为78.31%,AUC为0.816(95%CI:0.768~0.864)。回归分析在最佳临界值是对应的敏感度为80.28%,特异性为80.72%,AUC为0.869(95%CI:0.831~0.907),回归分析的敏感度和特异性明显高于MRI纹理特征的单独诊断。Logistic分析显示,能量、熵、惯性矩、相关和逆差矩是肺腺癌脑转移EGFR基因突变的影响因素(P<0.05)。结论能量、熵、惯性矩、相关及逆差矩等MRI纹理特征在肺腺癌脑转移瘤中异常表达,单独及回归分析联合诊断可应用于提高肺癌脑转移的诊断效能;且基于MRI纹理特征分析可以用于预测肺腺癌脑转移EGFR基因突变。
Objective To explore the diagnosis of lung adenocarcinoma with brain metastases based on MRI texture features and predictive value of EGFR gene mutation in lung adenocarcinoma with brain metastases.Methods A total of 154 patients with lung adenocarcinoma admitted were selected,and patients were divided into BMS group(n=71)and non-BMS group(n=83)according to whether they had BMS.MRI texture characteristics in the two groups were compared to evaluate the diagnostic value of energy,entropy,moment of inertia,correlation and deficit moment for lung adenocarcinoma brain metastases.EGFR gene detection was performed for patients in BMS group,and patients were divided into EGFR mutation group(n=33)and EGFR wild group(n=38).Univariate and multivariate Logistic regression analysis were performed to determine the influencing factors of MRI texture features on EGFR gene mutations in patients with lung adenocarcinoma with brain metastasis.Results The entropy and correlation in BMS group were significantly higher than those in the non-BMS group,and the energy,moment of inertia and deficit moment in BMS group were significantly lower than those in the non-BMS group.According to the ROC curve,the critical value of energy,sensitivity,specificity and AUC were 0.00121,69.01%,71.08%and 0.801(95%CI:0.750-0.852).The entropy critical value was 10.42,the sensitivity was 73.24%,the specificity was 69.88%,and the AUC was 0.760(95%CI:0.707-0.812).The critical value of moment of inertia was 7139.86,the sensitivity was 64.79%,the specificity was 60.24%,and the AUC was 0.708(95%CI:0.645-0.772).The critical correlation value was 0.000026,the sensitivity was 59.15%,the specificity was 60.24%,and the AUC was 0.631(95%CI:0.564-0.698).The critical value of the negative moment was 0.0301,the sensitivity was 74.65%,the specificity was 78.31%,and the AUC was 0.816(95%CI:0.768-0.864).At the optimal critical value of regression analysis,the corresponding sensitivity,specificity and AUC were 80.28%,80.72%and 0.869(95%CI:0.831-0.907).The sensitivity and s
作者
李平
王鸿礼
王伟亮
LI Ping;WANG Hongli;WANG Weiliang(Department of Radiology,Suzhou Municipal Hospital,Suzhou Hospital Affiliated to Nanjing Medical University,Suzhou,Jiangsu Province 215001,P.R.China)
出处
《临床放射学杂志》
北大核心
2023年第3期498-503,共6页
Journal of Clinical Radiology
关键词
磁共振纹理特征
肺腺癌脑转移瘤
EGFR基因突变
模型建立
Magnetic resonance texture features
Lung adenocarcinoma brain metastases
EGFR gene mutation
Model establishment
