Establishment of nomogram prediction model for neurological complications after interventional treatment of ruptured intracranial aneurysms / 中国综合临床

ABSTRACT

Objective:To explore the risk factors of neurological complications after interventional treatment of ruptured intracranial aneurysms(RIAS), and to establish a predictive model of nomogram.Methods:The clinical data of 89 patients with RIAS who underwent endovascular treatment in Nanyang Second General Hospital Affiliated to Xingxiang Medical University from January 2016 to January 2019 were retrospectively studied.The clinical imaging data were collected and followed up for 6 months.The patients were divided into two groups no neurological complications group (61 cases) and neurological complications group (28 cases). To analyze the clinical indicators and the possible related factors of neurological complications after RIAS interventional therapy.A nomogram was established to score the influencing factors, and a scoring prediction model was constructed; the clinical calibration of the model was evaluated by consistency index (C-index) and calibration curve, and the clinical differentiation of the model was evaluated by nomogram relying on ROC curve.Results:Multivariate logistic regression analysis showed that Hunt-Hess classification ( OR=4.927, 95% CI 1.189-20.426, P=0.028), Fisher classification ( OR=4.633, 95% CI 1.012-21.208, P=0.048 ), aneurysm cyst xiaofu ( OR=5.918, 95% CI 1.104-24.948, P=0.015), wide carotid aneurysm ( OR=4.381, 95% CI 1.029-18.645, P=0.046) and treatment Strategy ( OR=4.887, 95% CI 1.235-19.329, P=0.024) is an independent risk factor for nerve-related complications after RIAs interventional therapy.The predictive model of nomogram showed that Hunt-Hess classification (grade IV, V) was 100, aneurysm bleb (with) 98, treatment strategy (stent implantation) 95, wide-necked aneurysm (yes) 92 and Fisher grade (grade III, IV) 81; the C-index of the predictive model was 0.871; the nomogram relied on ROC curve AUC 0.871, and the treatment strategy (stent implantation) was 95; the Fisher grade (grade III, IV) was 81; the C-index of the predictive model was 0.871.The sensitivity and specificity were 85.71%(24/28) and 77.05%(47/61) respectively. Conclusion:Hunt Hess classification, Fisher classification, aneurysmal sac caruncle, wide necked aneurysms and treatment strategies will affect the occurrence of neurological complications after RIAS interventional therapy.The nomogram established by this method can provide intuitive and reliable reference for clinical practice.