Correlation Between EGFR Gene Mutation And Lung Adenocarcinoma Brain Metastasis And Construction Of A Risk Prediction Model For Brain Metastasis Of Lung Adenocarcinoma

Objective1.To explore the correlation between epidermal growth factor receptor（EGFR）mutation and brain metastasis in lung adenocarcinoma;2.To construct a brain metastasis risk prediction model for patients with EGFR mutation lung adenocarcinoma,and evaluate the applicability of the model.Methods1.A total of 488 patients with lung adenocarcinoma from March 2018 to October2022 were collected in the Second Affiliated Hospital of Anhui Medical University,including 188 cases of EGFR wild and 300 cases of EGFR mutation.The chi-square test was used to analyze the correlation between brain metastases and the clinicopathological features of lung adenocarcinoma.The proportion of brain metastases and the number of brain metastases in patients with EGFR wild-type and EGFR mutant,exon 19 deletion（19 Del）and L858R mutation in 21 exon（21 L858R）were compared.The median metastasis time of brain metastasis after EGFR19 Del and 21 L858R mutation patients receiving EGFR-TKIs targeted therapy was analyzed,and a logistic regression model was established to predict the risk factors for brain metastasis of lung adenocarcinoma.2.The sensitivity of different EGFR-mutant lung adenocarcinoma cell lines PC9（19 Del）and H3255（21 L858R）to gefitinib was detected by in vitro CCK8,and the semi-inhibitory concentration IC₅₀was calculated.3.The data of 300 patients with EGFR mutation lung adenocarcinoma were randomly divided into modeling dataset（n=210 cases）and validation dataset（n=90cases）according to the 7:3 ratio through caret package in R4.2.2.The rms package in R language was used to construct a risk prediction model for lung adenocarcinoma brain metastasis by modeling dataset,and the ROC curve,calibration curve,goodness-of-fit test,and external verification methods were used to test the model and evaluate the model performance.Results1.Of the 410 patients with lung adenocarcinoma,a total of 182（7.3%）cases had brain metastasis.The incidence of brain metastases in age ≤60 and age>60 was 43.2%and 32.7%,respectively（X²=5.621,P=0.018）,and the incidence of brain metastases in patients with TNM stage N0-1 and N2-3 was 18.9%and 43.4%,respectively（X²=23.659,P=0.000）.There were no significant differences between EGFR mutation and EGFR wild type,and between 19 Del and 21 L858R in the proportion of liver,bone,adrenal gland,pleura metastases,and there was no significant difference in the proportion of EGFR mutation and EGFR wild type brain metastasis,but EGFR mutation had more multi-lesion characteristics（P=0.000）,and the proportion of brain metastases in patients with 21 L858R was higher than that of 19 Del（P=0.001）,and the number of metastases in the former was significantly higher than that of the latter（P=0.005）.There were 103 patients with newly diagnosed brain metastases,35 cases with EGFR wild,and 68 cases with EGFR mutant（P>0.05）,of which the number of19 Del and 21 L858R were 20 and 48,respectively（P=0.002）.2.After EGFR-TKIs targeted therapy,18 patients with 19 Del and 28 patients with21 L858R developed brain metastases.The median time of brain metastasis after EGFR-TKIs treatment were 14.5 months（7.0,21.5）and 7 months（4.0,11.5）,respectively（P=0.006）.The IC₅₀of PC9 and H3255 to gefitinib were 0.037±0.008μmol/L and 0.150±0.04μmol/L,respectively（P=0.007）.3.In this study,Logistic multivariate analysis was used in the modeling set to screen risk factors lymph node stage N,age and EGFR mutation as model predictors,and the risk prediction model of lung adenocarcinoma brain metastasis with EGFR mutation was successfully constructed.The area under the curve（AUC）of the forecast model was 0.809（95%CI:0.753～0.865）,the sensitivity was 80.2%,the specificity was72.1%,the Brier score was 0.171,the goodness-of-fit test P=0.565,and the Calibration curve showed that the model consistency was acceptable.Moreover,the predictive model had an AUC of 0.813（95%CI:0.731～0.896）,a sensitivity of 84.8%,and a specificity of 66.7%in the validation set dataset.Conclusions1.Age≤60 years,lymph node N2-3 stage and EGFR 21 L858R were independent risk factors for brain metastasis of lung adenocarcinoma.2.Compared with EGFR wild type,the number of brain metastases in patients with EGFR mutation is more likely to be multiple metastases,and patients with EGFR 21L858R mutation are more likely to have multiple lesions than EGFR 19 Del brain metastases.3.Patients with 21 L858R still have less time to develop brain metastases after EGFR-TKIs than 19 Del mutations,which may be related to drug sensitivity.4.We successfully constructed a risk prediction model for lung adenocarcinoma brain metastasis with EGFR mutation,which has certain clinical value for predicting the risk of lung adenocarcinoma brain metastasis.