Comparative study of nine machine learning models for predicting early hematoma expansion and poor prognosis of deep supratentorial spontaneous intracerebral hemorrhage
Chen Kai1, She Hualong2, Wu Tao2, Li Tao2, Yang Liu3, Liu Fei4, Jiang Yasi5, Zhang Fanjuan6
1Department of Medical Imaging, Shenzhen Samii Medical Center (the Fourth People's Hospital of Shenzhen), Shenzhen 518118, China; 2Department of Radiology, Affiliated Hospital of Xiangnan University, Chenzhou 423000, China; 3Department of Neurology, the First Affiliated Hospital of Jinan University, Guangzhou 510630, China; 4Department of Radiology, Suxian Hospital Affiliated to Xiangnan University, Chenzhou 423000, China; 5Department of Radiology, the Second Affiliated Hospital of Xiangnan University, Chenzhou 423000, China; 6Department of Emergency, Shenzhen Samii Medical Center (the Fourth People's Hospital of Shenzhen), Shenzhen 518118, China
Abstract：Objective This study aimed to compare the predictive performance of nine machine learning models for early hematoma expansion (HE) and poor outcomes in patients with supratentorial deep intracerebral hemorrhage (SICH). Methods In this study, a retrospective study design was used. A total of 420 patients with SICH in four hospitals from January 2015 to May 2019 were included, 275 male and 145 female with a mean age of (61.0±12.9) years. The included patients were divided into 294 in the training set and 126 in the validation set in a 7∶3 ratio using randomization. In reviewed CT within 72 h, the hematoma volume increased by >6 mL or >33%, which was identified as HE. The prognosis was evaluated using the modified Rankin scale (mRS), and a score of mRS >3 was considered as a poor outcome. The baseline characteristics of the training and test sets were compared, and then nine machine learning algorithms including random forest, extreme gradient boosting (XGboost), gradient boosting decision tree, adaptive boosting, Naive Bayes, logistic regression, support vector machines, K-nearest neighbor, multilayer perceptron were used to construct prediction models for early HE and poor outcomes. In the training set, subject operating characteristic curves (ROC) were plotted on the basis of the sensitivity and specificity of each model, and the area under the curve (AUC) of threefold cross-validation was used to compare the predictive performance, which was investigated in the test set. Results The differences in baseline characteristics between the training and test sets were not statistically significant (all P values>0.05). Among the 420 patients, early HE was observed in 117 patients (27.86%). Follow-up results were obtained in 399 patients, and 210 patients (52.63%) had poor outcomes. In the training set, under threefold cross-validation, AUCs of nine machine learning models for predicting early HE ranged from 0.590 to 0.685, and the XGboost model was the highest at 0.685±0.024. In the validation set, the AUC was 0.686, and the 95% confidence interval (CI) was 0.578-0.721. In predicting poor outcomes, AUCs of the nine machine learning models ranged from 0.703 to 0.852, and the logistic regression model was the highest at 0.852±0.041. In the validation set, the AUC of the logistic regression model was 0.894 (95%CI 0.862-0.912). Conclusion Among the nine machine learning models, XGboost has the best predictive performance for early HE of deep supratentorial SICH, whereas the logistic regression model has the highest predictive performance for poor outcomes. Furthermore, in predicting different clinical outcomes, an appropriate machine learning model should be selected.
陈凯, 佘华龙, 吴涛, 李涛, 杨柳, 刘飞, 蒋亚思, 张帆娟. 9种机器学习模型预测幕上深部自发性脑出血早期血肿扩张及预后不良的比较[J]. 中华解剖与临床杂志, 2022, 27(9): 601-607.
Chen Kai, She Hualong, Wu Tao, Li Tao, Yang Liu, Liu Fei, Jiang Yasi, Zhang Fanjuan. Comparative study of nine machine learning models for predicting early hematoma expansion and poor prognosis of deep supratentorial spontaneous intracerebral hemorrhage. Chinese Journal of Anatomy and Clinics, 2022, 27(9): 601-607.
中华医学会神经病学分会, 中华医学会神经病学分会脑血管病学组. 中国脑出血诊治指南(2019)[J]. 中华神经科杂志, 2019, 52(12): 994-1005. DOI: 10.3760/cma.j.issn.1006-7876.2019.12.003.Chinese Society of Neurology, Chinese Stroke Society. Chinese guidelines for diagnosis and treatment of acute intracerebral hemorrhage 2019[J]. Chin J Neurol, 2019, 52(12): 994-1005. DOI: 10.3760/cma.j.issn.1006-7876.2019.12.003.
[2]
Li Q, Warren AD, Qureshi AI, et al.Ultra-early blood pressure reduction attenuates hematoma growth and improves outcome in intracerebral hemorrhage[J]. Ann Neurol, 2020, 88(2): 388-395. DOI: 10.1002/ana.25793.
[3]
李庆勇, 林茹, 张东明, 等. 影响自发性脑出血患者预后的相关因素分析[J]. 中华保健医学杂志, 2019, 21(6): 525-527. DOI: 10.3969/j.issn.1674-3245.2019.06.010.Li QY,Lin R,Zhang DM,et al.Related risk factors for the prognosis of patients with spontaneous cerebral hemorrhage[J]. Chin J Health Care Med, 2019, 21(6): 525-527. DOI: 10.3969/j.issn.1674-3245.2019.06.010.
[4]
Neisewander BL, Hu K, Tan Z, et al.Location of thalamic hemorrhage impacts prognosis[J]. World Neurosurg, 2018,116: e525-e533. DOI: 10.1016/j.wneu.2018.05.026.
[5]
Roh D, Boehme A, Young C, et al.Hematoma expansion is more frequent in deep than lobar intracerebral hemorrhage[J]. Neurology, 2020, 95(24): e3386-e3393. DOI: 10.1212/WNL.0000000000010990.
[6]
Hatwell J, Gaber MM, Atif Azad RM.Ada-WHIPS: explaining AdaBoost classification with applications in the health sciences[J]. BMC Med Inform Decis Mak, 2020, 20(1): 250. DOI: 10.1186/s12911-020-01201-2.
[7]
Sheridan RP, Wang WM, Liaw A, et al.Extreme gradient boosting as a method for quantitative structure-activity relationships[J]. J Chem Inf Model, 2016, 56(12): 2353-2360. DOI: 10.1021/acs.jcim.6b00591.
[8]
王业青, 时代, 陆宽, 等. 预测脑出血血肿扩大的诺模图模型建立与多角度评价[J]. 中华医学杂志, 2021, 101(31): 2471-2477. DOI: 10.3760/cma.j.cn112137-20210118-00161.Wang YQ, Shi D, Lu K, et al.Establishment of a nomogram model for predicting hematoma expansion in intracerebral hemorrhage and its multidimensional evaluation[J]. Natl Med J China, 2021, 101(31): 2471-2477. DOI: 10.3760/cma.j.cn112137-20210118-00161.
[9]
Yogendrakumar V, Moores M, Sikora L, et al.Evaluating hematoma expansion scores in acute spontaneous intracerebral hemorrhage: a systematic scoping review[J]. Stroke, 2020, 51(4): 1305-1308. DOI: 10.1161/STROKEAHA.119.028574.
[10]
Chen Q, Zhu D, Liu J, et al.Clinical-radiomics nomogram for risk estimation of early hematoma expansion after acute intracerebral hemorrhage[J]. Acad Radiol, 2021, 28(3): 307-317. DOI: 10.1016/j.acra.2020.02.021.
[11]
Song Z, Guo D, Tang Z, et al.Noncontrast computed tomography-based radiomics analysis in discriminating early hematoma expansion after spontaneous intracerebral hemorrhage[J]. Korean J Radiol, 2021, 22(3): 415-424. DOI: 10.3348/kjr.2020.0254.
[12]
Chen K, Deng L, Li Q, et al.Are computed-tomography-based hematoma radiomics features reproducible and predictive of intracerebral hemorrhage expansion? An in vitro experiment and clinical study[J]. Br J Radiol, 2021, 94(1121): 20200724. DOI: 10.1259/bjr.20200724.
[13]
Ironside N, Chen CJ, Dreyer V, et al.Location-specific differences in hematoma volume predict outcomes in patients with spontaneous intracerebral hemorrhage[J]. Int J Stroke, 2020, 15(1): 90-102. DOI: 10.1177/1747493019830589.
[14]
李青润, 韩雷, 陈红日, 等. 影像组学模型对自发性脑出血早期血肿扩大的预测及与常规影像征象的比较[J]. 中国医学计算机成像杂志, 2021, 27(2): 91-96. DOI: 10.19627/j.cnki.cn31-1700/th.2021.02.002.Li QR, Han L, Chen HR, et al.Radiomics for predicting hematoma expansion in early stage of spontaneous intracerebral hemorrhage and comparison with conventional radiological predictors[J] Chinese Computed Medical Imaging, 2021, 27(2): 91-96. DOI: 10.19627/j.cnki.cn31-1700/th.2021.02.002.
[15]
Yu Z, Ma L, Zheng J, et al.Comparison of hematoma density heterogeneity and ultraearly hematoma growth in predicting hematoma expansion in patients with spontaneous intracerebral hemorrhage[J]. J Neurol Sci, 2017, 379: 44-48. DOI: 10.1016/j.jns.2017.05.049.
[16]
唐卉, 辜蕊, 周嫱, 等. 自发性脑出血长期预后因素分析以及预测评分量表的建立[J]. 四川医学, 2021, 42(9): 913-918. DOI: 10.16252/j.cnki.issn1004-0501-2021.09.011.Tang H, Gu R, Zhou Q, et al.Analysis of Long-term prognostic factors of spontaneous intracerebral hemorrhage and establishment of predictive score scale[J]. Sichuan Medical Journal, 2021, 42(9): 913-918. DOI: 10.16252/j.cnki.issn1004-0501-2021.09.011.
[17]
Sporns PB, Kemmling A, Schwake M, et al . Triage of 5 Noncontrast computed tomography markers and spot sign for outcome prediction after intracerebral hemorrhage[J]. Stroke, 2018,49(10):2317-2322. DOI:10.1161/STROKEAHA.118.021625.
李彦斌, 曲鑫, 王尚武, 等. 神经内镜与微创穿刺手术治疗高血压脑出血的近期效果分析[J]. 中国脑血管病杂志, 2020, 17(3): 135-139. DOI: 10.3969/j.issn.1672-5921.2020.03.004.Li YB, Qu X, Wang SW, et al.Analysis of short-term effect of neuroendoscopy and minimally invasive puncture on hypertensive intracerebral hemorrhage[J]. Chin J Cerebrovasc Dis, 2020, 17(3): 135-139. DOI: 10.3969/j.issn.1672-5921.2020.03.004.
[20]
梅苑, 王建村, 胡观成. 立体定向联合尿激酶治疗脑出血的疗效分析[J]. 国际神经病学神经外科学杂志, 2020, 47(2): 181-185. DOI: 10.16636/j.cnki.jinn.2020.02.015.Mei Y, Wang JC, Hu GC.Clinical effect of stereotactic therapy combined with urokinase in treatment of cerebral hemorrhage[J]. Journal of International Neurology and Neurosurgery, 2020, 47(2): 181-185. DOI: 10.16636/j.cnki.jinn.2020.02.015.
[21]
Delcourt C, Sato S, Zhang S, et al.Intracerebral hemorrhage location and outcome among INTERACT2 participants[J]. Neurology, 2017, 88(15): 1408-1414. DOI: 10.1212/WNL.0000000000003771.
2022, Vol. 27 
