Department of Neurology, The University of Arizona, USA.
Department of Neuroradiology, MD Anderson Cancer Center, USA.
Neuroradiol J. 2024 Feb;37(1):74-83. doi: 10.1177/19714009231212364. Epub 2023 Nov 3.
We aimed to use machine learning (ML) algorithms with clinical, lab, and imaging data as input to predict various outcomes in traumatic brain injury (TBI) patients.
In this retrospective study, blood samples were analyzed for glial fibrillary acidic protein (GFAP) and ubiquitin C-terminal hydrolase L1 (UCH-L1). The non-contrast head CTs were reviewed by two neuroradiologists for TBI common data elements (CDE). Three outcomes were designed to predict: discharged or admitted for further management (prediction 1), deceased or not deceased (prediction 2), and admission only, prolonged stay, or neurosurgery performed (prediction 3). Five ML models were trained. SHapley Additive exPlanations (SHAP) analyses were used to assess the relative significance of variables.
Four hundred forty patients were used to predict predictions 1 and 2, while 271 patients were used in prediction 3. Due to Prediction 3's hospitalization requirement, deceased and discharged patients could not be utilized. The Random Forest model achieved an average accuracy of 1.00 for prediction 1 and an accuracy of 0.99 for prediction 2. The Random Forest model achieved a mean accuracy of 0.93 for prediction 3. Key features were extracranial injury, hemorrhage, UCH-L1 for prediction 1; The Glasgow Coma Scale, age, GFAP for prediction 2; and GFAP, subdural hemorrhage volume, and pneumocephalus for prediction 3, per SHAP analysis.
Combining clinical and laboratory parameters with non-contrast CT CDEs allowed our ML models to accurately predict the designed outcomes of TBI patients. GFAP and UCH-L1 were among the significant predictor variables, demonstrating the importance of these biomarkers.
我们旨在使用机器学习（ML）算法以及临床、实验室和影像学数据作为输入，来预测创伤性脑损伤（TBI）患者的各种结局。
在这项回顾性研究中，分析了脑损伤患者的胶质纤维酸性蛋白（GFAP）和泛素羧基末端水解酶 L1（UCH-L1）的血液样本。两名神经放射科医生对非对比头部 CT 进行了 TBI 常见数据元素（CDE）的审查。设计了三个结局来进行预测：出院或留院进一步治疗（预测 1）、死亡或存活（预测 2），以及仅入院、延长住院时间或进行神经外科手术（预测 3）。训练了 5 种 ML 模型。使用 Shapley 加法解释（SHAP）分析来评估变量的相对重要性。
440 名患者用于预测预测 1 和 2，而 271 名患者用于预测 3。由于预测 3 需要住院治疗，因此无法使用死亡和出院的患者。随机森林模型对预测 1 的平均准确率为 1.00，对预测 2 的准确率为 0.99。随机森林模型对预测 3 的平均准确率为 0.93。根据 SHAP 分析，关键特征是预测 1 中的颅外损伤、出血、UCH-L1；预测 2 中的格拉斯哥昏迷量表、年龄、GFAP；预测 3 中的 GFAP、硬膜下血肿量、气颅。
将临床和实验室参数与非对比 CT CDE 相结合，使我们的 ML 模型能够准确预测 TBI 患者的设计结局。GFAP 和 UCH-L1 是重要的预测变量，这表明这些生物标志物的重要性。
