基于CECT影像组学预测非小细胞肺癌患者新辅助化疗后脑转移风险

Predicting the risk of brain metastasis in non-small cell lung cancer patients after neoadjuvant chemotherapy based on CECT radiomics

目的:构建基于增强计算机断层扫描(CECT)影像组学的非小细胞肺癌(NSCLC)患者新辅助化疗后脑转移风险的列线图。方法:选取2018年3月~2021年9月徐州医科大学附属沭阳医院就诊的252例行新辅助化疗的NSCLC作为模型组,2021年10月~2022年7月就诊的110例行新辅助化疗的NSCLC作为验证组。应用Logistic回归分析筛选NSCLC患者新辅助化疗后脑转移的危险因素。利用LIFEx软件提取NSCLC患者新辅助化疗前的CECT影像组学特征,采用LASSO对CECT定量影像学指标的参数进行筛选。建立CECT影像组学、临床特征模型。通过整合优化的CECT影像组学及临床特征模型,采用R软件建立NSCLC患者新辅助化疗后脑转移的列线图。结果:Logsitc回归分析结果显示,Ⅳ期NSCLC、腺癌、低分化及淋巴结转移数目≥4等均是NSCLC患者新辅助化疗后脑转移的危险因素(P<0.05)。模型组和验证组的校正曲线均显示预测值与实际值的一致性较好;模型组的ROC曲线下面积为0.759(95%CI:0.701~0.811),验证组的ROC曲线下面积为0.788(95%CI:0.747~0.829);模型组的决策曲线显示阈值概率为7%~81%时,列线图预测NSCLC患者新辅助化疗后脑转移的净获益值较高,验证组的决策曲线显示阈值概率为4%~85%时,列线图预测NSCLC患者新辅助化疗后脑转移的净获益值较高。结论:基于CECT影像组学的NSCLC患者新辅助化疗后脑转移的列线图模型的准确性及临床应用价值较好,能够用于预测NSCLC患者新辅助化疗后脑转移的风险。

Objective:To construct a nomogram for predicting brain metastases in non-small cell lung cancer(NSCLC)patients after neoadjuvant chemotherapy based on contrast-enhanced computed tomography(CECT)radiomics.Methods:252 NSCLC patients undergone neoadjuvant chemotherapy in our hospital between March 2018 and September 2021 were included in the model group,and another 110 patients visited out Shuyang Hospital affiliated to Xuzhou Medical University between October 2021 and July 2022 were included in the validation group.Logistic regression analysis was used to screen the risk factors for brain metastases in NSCLC patients after neoadjuvant chemotherapy.LIFEx software was used to extract radiomics features from CECT images of NSCLC patients before neoadjuvant chemotherapy,and LASSO was used to screen the parameters of CECT quantitative imaging indicators.Models were established based on CECT radiomics and clinical features.By integrating optimized CECT radiomics and clinical feature models,a nomogram for brain metastasis in NSCLC patients after neoadjuvant chemotherapy was established using R software.Results:Logistic regression analysis showed that stage IV NSCLC,adenocarcinoma,low differentiation and lymph node metastasis number≥4 were all risk factors for brain metastases in NSCLC patients after neoadjuvant chemotherapy(P<0.05).The calibration curves of both the model group and the validation group showed good consistency between the predicted values and the actual values.The area under the ROC curve in the model group was 0.759(95%CI:0.701-0.811),and that in the validation group was 0.788(95%CI:0.747-0.829).Analysis of decision curves revealed that within the specified threshold probability ranges(7%-81%for the model group and 4%-85%for the validation group),the nomogram consistently predicted a high net benefit value for NSCLC patients with brain metastasis after neoadjuvant chemotherapy.Conclusion:The nomogram model based on CECT radiomics demonstrates good accuracy and clinical application value,and can be used for the prediction of brain metastasis in NSCLC patients after neoadjuvant chemotherapy.