Development And Validation Of A Clinical Prediction Model For Patients With Intracranial Atherosclerosis-acute Large Vessel Occlusion

Objective: Development and validation of a novel clinical prediction model for more accurate preoperative identification of Intracranial atherosclerosis-acute large vessel occlusion.Methods: Patients with acute large vessel occlusion at Jining No.1People’s Hospital from 2019.01 to 2021.09 were retrospectively included as the training cohort.The 70 patients who met the inclusion and exclusion criteria were included in the validation cohort(2021.10-2022.08).Demographics,onset form,medical history,digital subtraction angiography(DSA)imaging data,and laboratory test data were collected.Preprocedural parameters for the ICAS-LVO risk prediction model were established by Stepwise logistic regression controlling for the confounding effects.Then,we constructed a nomogram model and evaluated its performance via the Hosmer-Lemeshow goodness-of-fit test,the area under the ROC curve(AUC)analysis.Patients in the model validation cohort were compared with the actual results of ICAS-LVO judgments based on the model developed in this study to evaluate the predictive value of the model.Results: The 231 acute LVO patients were included in the final analysis,74(32.3%)patients were ICAS-LVO.A preoperative diagnosis prediction model consisting of five predictors for ICAS-LVO,including fluctuating symptoms,NIHSS<16,atrial fibrillation tapered sign,and ASITN/SIR score≥2.The model depicted an acceptable calibration(Hosmer–Lemeshow test,P=0.451)and good discrimination(AUC,0.941;95% confidence interval,0.910–0.971).The optimal cutoff value for the ICAS-LVO scale was 2 points,with 86.5% sensitivity,91.1% specificity,and 90.5% accuracy.In the validation cohort,the discriminative ability was promising with an AUC value of 0.897,implying a good predictive performance.Conclusion: The ICAS-LVO clinical prediction model constructed from five predictors of symptom fluctuation progression,NIHSS score<16,AF,tapered sign and ASITN/SIR score ≥2 had high sensitivity and accuracy.