Asadi Hamed, Kok Hong Kuan, Looby Seamus, Brennan Paul, O'Hare Alan, Thornton John
Neurointerventional Service, Department of Radiology, Beaumont Hospital, Dublin, Ireland; School of Medicine, Faculty of Health, Deakin University, Waurn Ponds, Australia.
Interventional Radiology Service, Department of Radiology, Beaumont Hospital, Dublin, Ireland.
World Neurosurg. 2016 Dec;96:562-569.e1. doi: 10.1016/j.wneu.2016.09.086. Epub 2016 Sep 28.
To identify factors influencing outcome in brain arteriovenous malformations (BAVM) treated with endovascular embolization. We also assessed the feasibility of using machine learning techniques to prognosticate and predict outcome and compared this to conventional statistical analyses.
A retrospective study of patients undergoing endovascular treatment of BAVM during a 22-year period in a national neuroscience center was performed. Clinical presentation, imaging, procedural details, complications, and outcome were recorded. The data was analyzed with artificial intelligence techniques to identify predictors of outcome and assess accuracy in predicting clinical outcome at final follow-up.
One-hundred ninety-nine patients underwent treatment for BAVM with a mean follow-up duration of 63 months. The commonest clinical presentation was intracranial hemorrhage (56%). During the follow-up period, there were 51 further hemorrhagic events, comprising spontaneous hemorrhage (n = 27) and procedural related hemorrhage (n = 24). All spontaneous events occurred in previously embolized BAVMs remote from the procedure. Complications included ischemic stroke in 10%, symptomatic hemorrhage in 9.8%, and mortality rate of 4.7%. Standard regression analysis model had an accuracy of 43% in predicting final outcome (mortality), with the type of treatment complication identified as the most important predictor. The machine learning model showed superior accuracy of 97.5% in predicting outcome and identified the presence or absence of nidal fistulae as the most important factor.
BAVMs can be treated successfully by endovascular techniques or combined with surgery and radiosurgery with an acceptable risk profile. Machine learning techniques can predict final outcome with greater accuracy and may help individualize treatment based on key predicting factors.
确定影响脑动静脉畸形（BAVM）血管内栓塞治疗效果的因素。我们还评估了使用机器学习技术预测治疗效果的可行性，并将其与传统统计分析方法进行比较。
对一家国家级神经科学中心22年间接受BAVM血管内治疗的患者进行回顾性研究。记录临床表现、影像学检查、手术细节、并发症及治疗效果。采用人工智能技术分析数据，以确定治疗效果的预测因素，并评估最终随访时预测临床疗效的准确性。
199例患者接受了BAVM治疗，平均随访时间为63个月。最常见的临床表现是颅内出血（56%）。随访期间，又发生了51次出血事件，包括自发性出血（n = 27）和与手术相关的出血（n = 24）。所有自发性事件均发生在先前栓塞的远离手术部位的BAVM中。并发症包括10%的缺血性卒中、9.8%的症状性出血和4.7%的死亡率。标准回归分析模型预测最终结局（死亡率）的准确率为43%，其中治疗并发症类型被确定为最重要的预测因素。机器学习模型在预测结局方面显示出97.5%的更高准确率，并将有无瘤巢瘘管确定为最重要的因素。
BAVM可通过血管内技术成功治疗，或与手术及放射外科联合治疗，风险可接受。机器学习技术能更准确地预测最终结局，并可能有助于根据关键预测因素实现个体化治疗。
