The Value Of CT-based Radiomics Combined With Clinical Data In Predicting Synchronous Brain Metastasis From Small Cell Lung Cancer

Objective: To explore the value of predicting synchronous brain metastasis of small cell lung cancer(SCLC)based on CT radiomics combined with clinical indicators.Methods: The clinical and imaging data of 160 patients with SCLC confirmed by biopsy or surgical pathology in Yijishan Hospital of Wannan Medical College from January 2016 to October 2022 were collected retrospectively.According to the results of cranial CT and MRI images,the patients were divided into a brain metastases group of 69 cases and a non-brain metastases group of 91 cases.Using a completely random method,all patients were randomly divided into a training set(n=112 cases)and a validation set(n=48cases)according to a ratio of 7:3.In the training set,principal component analysis(PCA),Kruskal-Wallis(KW)and logistic regression based on the introduction of the minimum absolute value convergence and selection operator,LR-lasso)to reduce the dimensionality and model construction of the radiomics data,and calculate the weight coefficient of each feature,the Radiomic-score of each patient was calculated.Statistically significant variables in clinical and imaging data were used to establish a clinical model through multivariate logistic regression analysis,and combined with the radiomics model to establish a joint model.The area under curve(AUC)was used to evaluate the prediction performance of the model,and the decision curve analysis(DCA)was used to evaluate the clinical application value of the model.Results: The AUC of clinical model,radiomics model and combined model in the training set and validation set were 0.760 and 0.760、0.852 and 0.818、0.909 and 0.926 respectively.DCA showed that the net benefit of the combined model in predicting brain metastases in SCLC patients was superior to that of the clinical model and the radiomics model.Conclusion:1、The expression of CEA,NSE and CYFRA21-1 in serum is the influential factor of brain metastasis of SCLC.2、The serum levels of NSE and CYFRA21-1 are independent risk factors of brain metastasis in SCLC patients.The higher the expression level of NSE and CYFRA21-1,the higher the risk of brain metastasis.3、The predictive value of radiologic model based on Radiologic label for synchronous brain metastasis of SCLC.4 、 The predictive value of the combined model based on the clinical model and radiography model for synchronous brain metastasis of SCLC is significantly higher than that of the clinical model and radiography model.radiomics combined with clinical data is expected to be a novel marker for predicting the risk of synchronous brain metastasis in SCLC.