Construction Of Risk Factors And Disease Prediction Model For Non-hypertensive And Non-diabetic Intracranial Aneurysm

Objective: We aim to investigate the risk factors of intracranial aneurysm without hypertension and diabetes formation and construct a disease prediction model.Methods : 99 patients diagnosed with intracranial aneurysm(aneurysm group)and 61 healthy subjects(control group)were retrospectively analyzed(2018.01～2021.12).Neutrophil to lymphocyte ratio(NLR),low-density cholesterol(LDL-C),high-density cholesterol(HDL),apolipoprotein AI(Apo-AI),apolipoprotein B(Apo-B),and apolipoprotein-AI/B(Apo-A1/Apo-B)were compared.The logistic regression model ROC curve and DCA decision score curve were used to analyze the correlation and predictive value of each factor with intracranial aneurysm formation.The stability of the model was verified by Cross validation.Results: The Logistics multivariate regression analysis showed that age(OR=1.157,95%CI: 1.092-1.226,P=0.000),NLR(OR=2.353,95%CI:1.005-5.510,P=0.049),HDL-C(OR=0.112,95%CI: 0.024-0.534,P=0.006),and Apo-AI(OR=0.049,95%CI: 0.003-0.769,P=0.032)are statistical difference,while LDL-C and Apo-B,Apo-A1/Apo-B were not.The new model is constructed after combining multivariable.The area under the ROC curve was0.897(95%CI: 0.848-0.947).When the cut-off value was 0.662,the sensitivity and specificity were 79.8% and 91.8% respectively.DCA decision score curve shows that the area under the curve of the joint model is larger than the area under the single factor curve,indicating that the model bias is small and the effect is good.Cross validation showed that the AUC of the training set was0.898(95%CI: 0.843-0.953)and that of the test set was 0.887(95%CI: 0.887).In0.771～0.990,the AUC of test sets did not decrease significantly,and the disease prediction model constructed in this study had clinical extrapolation.Conclusions: Age and NLR are risk factors for intracranial aneurysm formation,while Apo-A1 and HIGH density lipoprotein cholesterol are protective factors.Establishing a disease prediction model through logistics regression has high accuracy and can predict the possibility of intracranial aneurysm formation..