Neurosurgery Clinic, Department of Neuroscience, Psychology, Pharmacology and Child Health, Careggi University Hospital and University of Florence, Largo Piero Palagi 1, 50137, Florence, Italy.
Institute of Physics, Alma Mater Studiorum, University of Bologna, Bologna, Italy.
Acta Neurochir (Wien). 2020 Dec;162(12):3093-3105. doi: 10.1007/s00701-020-04484-6. Epub 2020 Jul 8.
Shunt-dependent hydrocephalus significantly complicates subarachnoid hemorrhage (SAH), and reliable prognosis methods have been sought in recent years to reduce morbidity and costs associated with delayed treatment or neglected onset. Machine learning (ML) defines modern data analysis techniques allowing accurate subject-based risk stratifications. We aimed at developing and testing different ML models to predict shunt-dependent hydrocephalus after aneurysmal SAH.
We consulted electronic records of patients with aneurysmal SAH treated at our institution between January 2013 and March 2019. We selected variables for the models according to the results of the previous works on this topic. We trained and tested four ML algorithms on three datasets: one containing binary variables, one considering variables associated with shunt-dependency after an explorative analysis, and one including all variables. For each model, we calculated AUROC, specificity, sensitivity, accuracy, PPV, and also, on the validation set, the NPV and the Matthews correlation coefficient (ϕ).
Three hundred eighty-six patients were included. Fifty patients (12.9%) developed shunt-dependency after a mean follow-up of 19.7 (± 12.6) months. Complete information was retrieved for 32 variables, used to train the models. The best models were selected based on the performances on the validation set and were achieved with a distributed random forest model considering 21 variables, with a ϕ = 0.59, AUC = 0.88; sensitivity and specificity of 0.73 (C.I.: 0.39-0.94) and 0.92 (C.I.: 0.84-0.97), respectively; PPV = 0.59 (0.38-0.77); and NPV = 0.96 (0.90-0.98). Accuracy was 0.90 (0.82-0.95).
Machine learning prognostic models allow accurate predictions with a large number of variables and a more subject-oriented prognosis. We identified a single best distributed random forest model, with an excellent prognostic capacity (ϕ = 0.58), which could be especially helpful in identifying low-risk patients for shunt-dependency.
分流依赖性脑积水显著增加蛛网膜下腔出血（SAH）的复杂性，近年来人们一直在寻求可靠的预后方法，以降低与延迟治疗或忽视发病相关的发病率和成本。机器学习（ML）定义了现代数据分析技术，可实现基于对象的准确风险分层。我们旨在开发和测试不同的 ML 模型，以预测颅内动脉瘤性 SAH 后分流依赖性脑积水。
我们查阅了 2013 年 1 月至 2019 年 3 月在我们机构接受治疗的颅内动脉瘤性 SAH 患者的电子病历。我们根据该主题的先前研究结果选择模型的变量。我们在三个数据集上训练和测试了四种 ML 算法：一个包含二进制变量，一个考虑探索性分析后与分流依赖性相关的变量，一个包含所有变量。对于每个模型，我们计算了 AUROC、特异性、敏感性、准确性、PPV，还在验证集上计算了 NPV 和 Matthews 相关系数（ϕ）。
共纳入 386 例患者。50 例（12.9%）在平均随访 19.7（±12.6）个月后出现分流依赖性。检索到 32 个变量的完整信息，用于训练模型。基于验证集上的性能，选择了最佳模型，并通过考虑 21 个变量的分布式随机森林模型获得，该模型的 ϕ为 0.59，AUC 为 0.88；敏感性和特异性分别为 0.73（0.39-0.94）和 0.92（0.84-0.97），PPV 为 0.59（0.38-0.77）；NPV 为 0.96（0.90-0.98）。准确性为 0.90（0.82-0.95）。
机器学习预后模型可使用大量变量进行准确预测，并具有更面向对象的预后。我们确定了一个最佳的分布式随机森林模型，具有出色的预后能力（ϕ=0.58），这对于识别分流依赖性的低风险患者可能特别有帮助。
