Research on constructing prognosis prediction model of intracerebral hemorrhage using multimodal medical data and machine learning

doi:10.3877/cma.j.issn.2095-123X.2023.04.001

Abstract

Abstract:

Objective

To construct a predictive model of cerebral hemorrhage prognosis prognosis using multimodal medical data and machine learning, and to explore its predictive value.

Methods

A total of 400 cerebral hemorrhage data was retrospective analyzed, included 98 cerebral hemorrhage patients in the Neurology Department of Peking Union Medical College Shenzhen Hospital, Huazhong University of Science and Technology from January to December 2020 and 302 cerebral hemorrhage patients from the National Multicenter Intracranial Hemorrhage Database established by Peking Union Medical College Hospital from January to December 2020 to build an imaging database of cerebral hemorrhage, extract the labels of imaging omics and clinical related factors, and build a model to predict patients' prognosis. Another 100 patients with intracerebral hemorrhage in the Neurology Department of Peking Union Medical College Shenzhen Hospital, Huazhong University of Science and Technology from January to December 2021 were selected for prospective verification of the model.

Results

The incidence of poor prognosis in 400 patients was 19.00%. The results of univariate analysis and multivariate logistic regression analysis showed that GCS score, systolic blood pressure, diastolic blood pressure, blood glucose, hematoma volume, peripheral edema volume, fibrinogen were all the influencing factors of poor prognosis (P<0.05). The sensitivity, specificity, accuracy and area under curve (AUC) [95% confidence interval (95%CI)] of self coding image feature clinical data model for predicting prognosis in patients with cerebral hemorrhage were 100.00%, 99.38%, 99.50% and 0.994 (0.935-0.998) respectively, which were higher than those of self coding image feature model and traditional model (P<0.05), and those of the self coding image feature model were higher than those of traditional model (P<0.05). After verification, the sensitivity, specificity, accuracy and AUC (95%CI) of the self coded image feature clinical data model for predicting prognosis in patients with cerebral hemorrhage were 100.00%, 97.47%, 98.00% and 0.974 (0.922-0.996) respectively.

Conclusion

The self coding image feature clinical data model based on multi-modal medical data and machine learning has high efficiency in predicting cerebral hemorrhage prognosis.

Key words: Multimodal medical data, Machine learning, Cerebral hemorrhage, Quality of life

Xianjin Chen, Qinqin Wu, Changchun He, Qinghua Zhang. Research on constructing prognosis prediction model of intracerebral hemorrhage using multimodal medical data and machine learning[J]. Chinese Journal of Brain Diseases and Rehabilitation(Electronic Edition), 2023, 13(04): 193-198.

/ / Recommend

Add to citation manager EndNote|Ris|BibTeX

URL: https://zhnkjbykfzz.cma-cmc.com.cn/EN/10.3877/cma.j.issn.2095-123X.2023.04.001

https://zhnkjbykfzz.cma-cmc.com.cn/EN/Y2023/V13/I04/193

Figures/Tables 6

References 17

[1]	杨帆,杨国军,杨哲.北京地区3139例首发脑出血患者流行特征及预后情况分析[J].华南预防医学, 2022, 48(1): 46-54; 46-49, 54. DOI: 10.12183/j.scjpm.2022.0046.
[2]	中华医学会神经外科学分会,中国医师协会急诊医师分会,中华医学会神经病学分会脑血管病学组,等.高血压性脑出血中国多学科诊治指南[J].中国急救医学, 2020, 40(8): 689-702. DOI: 10.3969/j.issn.1002-1949.2020.08.001.
[3]	车鹏,黄可,胡俊,等. S-100β蛋白和NES对急诊脑出血患者预后的预测价值[J].重庆医学, 2021, 50(5): 757-761. DOI: 10.3969/j.issn.1671-8348.2021.05.009.
[4]	Rao B, Zohrabian V, Cedeno P, et al. Utility of artificial intelligence tool as a prospective radiology peer reviewer - detection of unreported intracranial hemorrhage[J]. Acad Radiol, 2021, 28(1): 85-93. DOI: 10.1016/j.acra.2020.01.035.
[5]	赖建东,王宁,罗坤,等.基于深度学习计算机辅助诊断系统测量脑出血量[J].中国医学影像技术, 2020, 36(12): 1781-1785. DOI: 10.13929/j.issn.1003-3289.2020.12.005.
[6]	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.
[7]	Boase K, Machamer J, Temkin NR, et al. Central curation of Glasgow outcome scale-extended data: lessons learned from TRACK-TBI[J]. J Neurotrauma, 2021, 38(17): 2419-2434. DOI: 10.1089/neu.2020.7528.
[8]	Abulhasan YB, Ortiz Jimenez J, Teitelbaum J, et al. Milrinone for refractory cerebral vasospasm with delayed cerebral ischemia[J]. J Neurosurg, 2020, 134(3): 971-982. DOI: 10.3171/2020.1.JNS193107.
[9]	Kase CS, Hanley DF. Intracerebral hemorrhage: advances in emergency care[J]. Neurol Clin, 2021, 39(2): 405-418. DOI: 10.1016/j.ncl.2021.02.002.
[10]	Derraz I, Cagnazzo F, Gaillard N, et al. Microbleeds, cerebral hemorrhage, and functional outcome after endovascular thrombectomy[J]. Neurology, 2021, 96(13): e1724-e1731. DOI: 10.1212/WNL.0000000000011566.
[11]	滕兆平,刘卫国,杨小旺,等.中性粒细胞与淋巴细胞比值对急性脑出血病人短期预后不良的预测价值[J].中西医结合心脑血管病杂志, 2020, 18(14): 2320-2324. DOI: 10.12102/j.issn.1672-1349.2020.14.033.
[12]	Park HK, Lee JS, Kim BJ, et al. Cilostazol versus aspirin in ischemic stroke with cerebral microbleeds versus prior intracerebral hemorrhage[J]. Int J Stroke, 2021, 16(9): 1019-1030. DOI: 10.1177/1747493020941273.
[13]	Ma L, Zhang S, Li Z, et al. Morbidity after symptomatic hemorrhage of cerebral cavernous malformation: a nomogram approach to risk assessment[J]. Stroke, 2020, 51(10): 2997-3006. DOI: 10.1161/STROKEAHA.120.029942.
[14]	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.
[15]	Pyun JM, Ryoo N, Park YH, et al. Fibrinogen levels and cognitive profile differences in patients with mild cognitive impairment[J]. Dement Geriatr Cogn Disord, 2020, 49(5): 489-496. DOI: 10.1159/000510420.
[16]	吴蓉蓉,鲁珊珊,张久楼,等.基于机器学习预测超时间窗急性前循环大血管闭塞患者机械取栓预后的研究[J].临床放射学杂志, 2022, 41(3): 404-409. DOI: 10.13437/j.cnki.jcr.2022.03.005.
[17]	姜荣涛,戚世乐,吴静,等.基于多模态脑影像和机器学习算法的个体行为预测研究现状及未来趋势[J].中华诊断学电子杂志, 2021, 9(3): 145-148. DOI: 10.3877/cma.j.issn.2095-655X.2021.03.001.

资料	预后不良（n=76）	预后良好（n=324）	χ²/t值	P值
性别[例（%）]			0.256	0.613
男	47（61.84）	196（60.49）
女	29（38.16）	128（39.51）
年龄（岁，Mean±SD）	58.65±11.20	57.32±10.83	0.960	0.338
出血部位[例（%）]			0.226	0.635
脑叶	5（6.58）	20（6.17）
基底节	56（73.68）	231（71.30）
丘脑	7（9.21）	30（9.26）
其他	8（10.53）	41（12.65）
GCS评分（分，Mean±SD）	6.69±1.35	5.12±1.06	11.040	<0.001
收缩压（mmHg，Mean±SD）	149.54±20.07	121.07±17.15	12.640	<0.001
舒张压（mmHg，Mean±SD）	98.58±13.44	71.45±10.09	19.795	<0.001
血糖（mmol/L，Mean±SD）	9.15±1.63	7.12±0.63	17.722	<0.001
血肿体积（mL，Mean±SD）	52.15±8.17	38.79±6.24	15.839	<0.001
周围水肿体积（mL，Mean±SD）	20.50±4.18	16.35±3.09	9.843	<0.001
FIB（g/L，Mean±SD）	1.18±0.24	3.07±0.52	30.923	<0.001
血小板计数（×10⁹/L，Mean±SD）	215.16±27.88	213.47±26.95	0.490	0.624

资料	预后不良（n=76）	预后良好（n=324）	χ²/t值	P值
性别[例（%）]			0.256	0.613
男	47（61.84）	196（60.49）
女	29（38.16）	128（39.51）
年龄（岁，Mean±SD）	58.65±11.20	57.32±10.83	0.960	0.338
出血部位[例（%）]			0.226	0.635
脑叶	5（6.58）	20（6.17）
基底节	56（73.68）	231（71.30）
丘脑	7（9.21）	30（9.26）
其他	8（10.53）	41（12.65）
GCS评分（分，Mean±SD）	6.69±1.35	5.12±1.06	11.040	<0.001
收缩压（mmHg，Mean±SD）	149.54±20.07	121.07±17.15	12.640	<0.001
舒张压（mmHg，Mean±SD）	98.58±13.44	71.45±10.09	19.795	<0.001
血糖（mmol/L，Mean±SD）	9.15±1.63	7.12±0.63	17.722	<0.001
血肿体积（mL，Mean±SD）	52.15±8.17	38.79±6.24	15.839	<0.001
周围水肿体积（mL，Mean±SD）	20.50±4.18	16.35±3.09	9.843	<0.001
FIB（g/L，Mean±SD）	1.18±0.24	3.07±0.52	30.923	<0.001
血小板计数（×10⁹/L，Mean±SD）	215.16±27.88	213.47±26.95	0.490	0.624

影响因素	β值	SE值	Wald χ²值	P值	OR值	95%CI
GCS评分	1.593	0.543	8.607	<0.001	4.918	3.887~5.012
收缩压	1.421	0.552	6.627	0.001	4.141	3.476~4.533
舒张压	1.438	0.567	6.432	0.003	4.212	3.860~4.437
血糖	1.575	0.638	6.094	0.005	4.831	4.158~5.026
血肿体积	1.983	0.882	5.055	0.012	7.265	6.895~7.474
周围水肿体积	1.802	0.803	5.036	0.012	6.062	5.738~6.204
FIB	-0.618	0.314	3.874	0.036	0.539	0.432~0.667

影响因素	β值	SE值	Wald χ²值	P值	OR值	95%CI
GCS评分	1.593	0.543	8.607	<0.001	4.918	3.887~5.012
收缩压	1.421	0.552	6.627	0.001	4.141	3.476~4.533
舒张压	1.438	0.567	6.432	0.003	4.212	3.860~4.437
血糖	1.575	0.638	6.094	0.005	4.831	4.158~5.026
血肿体积	1.983	0.882	5.055	0.012	7.265	6.895~7.474
周围水肿体积	1.802	0.803	5.036	0.012	6.062	5.738~6.204
FIB	-0.618	0.314	3.874	0.036	0.539	0.432~0.667

预测模型		灵敏度（%）	特异度（%）	准确度（%）	AUC	95%CI
新建模型	自编码影像特征模型	90.79	96.60	95.50	0.877	0.812~0.942
新建模型	自编码影像特征-临床数据模型	100.00	99.38	99.50	0.994	0.935~0.998
传统模型	临床数据模型	80.26	90.43	88.50	0.836	0.781~0.891
	影像组学模型	78.95	91.05	88.75	0.719	0.632~0.806
	影像组学-临床数据模型	81.58	94.14	91.50	0.766	0.682~0.850

Please choose a citation manager

Content to export