基于影像组学特征预测左室射血分数的可行性研究

doi:10.3969/j.issn.1674-8115.2023.09.010

上海交通大学学报（医学版） ›› 2023, Vol. 43 ›› Issue (9): 1162-1168.doi: 10.3969/j.issn.1674-8115.2023.09.010

• 论著 · 临床研究 • 上一篇

基于影像组学特征预测左室射血分数的可行性研究

刘启明(), 卢启帆, 柴烨子, 姜萌(), 卜军()

上海交通大学医学院附属仁济医院心内科，上海 200127

收稿日期:2023-03-23 接受日期:2023-08-04 出版日期:2023-09-28 发布日期:2023-09-28
通讯作者: 姜萌,卜军 E-mail:090503liu@sjtu.edu.cn;jiangmeng0919@163.com;pujun310@hotmail.com
作者简介:刘启明（1996—），男，硕士生；电子信箱：090503liu@sjtu.edu.cn。
基金资助:
国家自然科学基金(81971570);上海申康医院发展中心促进市级医院临床技能与临床创新能力三年行动计划项目(SHDC2020CR2025B);上海市科学技术委员会医学创新研究专项(20Y11910500);上海市科学技术委员会优秀技术带头人计划(21XD143210);上海交通大学医学院“双百人”项目(20172014);上海市浦东新区卫生和计划生育委员会联合攻关项目(PW2018D-03);上海交通大学“交大之星”计划医工交叉研究基金(YG2019ZDA13);上海理工大学医工交叉研究基金(10-20-302-425)

Radiomics-based left ventricular ejection fraction prediction: a feasibility study

LIU Qiming(), LU Qifan, CHAI Yezi, JIANG Meng(), PU Jun()

Department of Cardiology, Renji Hospital, Shanghai Jiao Tong University School of Medicine, Shanghai 200127, China

Received:2023-03-23 Accepted:2023-08-04 Online:2023-09-28 Published:2023-09-28
Contact: JIANG Meng,PU Jun E-mail:090503liu@sjtu.edu.cn;jiangmeng0919@163.com;pujun310@hotmail.com
Supported by:
National Natural Science Foundation of China(81971570);Shanghai Shenkang Hospital Development Center Three-year Action Plan: Promoting Clinical Skills and Innovation in Municipal Hospital(SHDC2020CR2025B);Innovation Research Project of Shanghai Science and Technology Commission(20Y11910500);Advanced Technology Leader of the Shanghai Science and Technology Commission(21XD143210);“Two-hundred Talents” Program of Shanghai Jiao Tong University School of Medicine(20172014);Shanghai Pudong Municipal Health Commission-Joint Research Project(PW2018D-03);Medical-Engineering Cross Research of Shanghai Jiao Tong University(YG2019ZDA13);Medical-Engineering Cross Research of University of Shanghai for Science and Technology(10-20-302-425)

摘要/Abstract

摘要：

目的·评估使用心脏磁共振（cardiac magnetic resonance，CMR）短轴电影图像提取的3D影像组学特征用于预测左室射血分数（left ventricular ejection fraction，LVEF）这一方法的可行性。方法·纳入2018年1月至2021年12月期间就诊于上海交通大学医学院附属仁济医院心内科的左室肥厚（left ventricular hypertrophy，LVH）患者共100例，及同时期健康对照（health control，HC）受试者100例，在心内科医师及放射科医师共同监督下完成CMR检查。随后由心内科医师手动勾画心内膜及心外膜并完成对入组人员心功能及形态学的测量与记录，主要包括LVEF、左心室舒张末期容积（left ventricular end-diastolic volume，LVEDV）、左心室舒张末期心肌质量（left ventricular end-diastolic mass，LVEDM）。通过Pyradiomics包提取CMR-cine序列心肌3D影像组学特征并使用相关系数与K-best方法进行特征筛选与排序。构建线性回归（linear regression，LR）、随机森林（random forest，RF）及梯度增强（gradient boost，Gb）3种机器学习回归模型，使用影像组学特征对LVEF进行预测，并与基于临床信息及CMR参数的LVEF预测结果进行比较。结果·在临床指标方面，LVH组与HC组之间LVEDV、LVEDM差异均有统计学意义（P<0.05）。通过提取心肌3D影像组学特征，获得10项特征并用于测试集样本LVEF预测。当所选特征为7项时，RF回归模型获得最佳表现，平均绝对误差（mean absolute error，MAE）为0.066±0.002，显著低于其他2种方法（P<0.001）；对于RF回归模型，基于临床信息与舒张末期CMR参数的LVEF预测结果显示，结合影像组学信息与CMR参数的模型（MAE=0.056±0.001）效果最佳，并显著优于单用影像组学（MAE=0.066±0.002）或单用CMR参数（MAE=0.060±0.001）的模型表现（P<0.05）。结论·影像组学特征用于LVEF预测具有一定的可行性，并且将影像组学信息与CMR参数结合，可以进一步提升模型预测准确率。

关键词: 心脏磁共振, 影像组学, 左室肥厚, 电影序列

Abstract:

Objective ·To assess the feasibility of using 3D imaging features extracted from cardiac magnetic resonance (CMR) short-axis cine images to predict left ventricular ejection fraction (LVEF). Methods ·A total of 100 left ventricular hypertrophy (LVH) patients who visited the Department of Cardiology, Renji Hospital, Shanghai Jiao Tong University School of Medicine from January 2018 to December 2021, as well as 100 healthy control (HC) subjects during the same period, were included. All subjects completed CMR examinations under the supervision of experienced cardiologists and radiologists. The endocardial and epicardial contours were then manually delineated by cardiologists. Measurements of cardiac function and morphology were completed and data was recorded, including LVEF, left ventricular end-diastolic volume (LVEDV), and left ventricular end-diastolic mass (LVEDM). Myocardial 3D radiomic features of CMR-cine sequences were extracted by the Pyradiomics package, and selected and sorted by using correlation coefficient and K-best method. The LVEF prediction was performed with linear regression (LR), random forest (RF) and gradient boost (GB) methods. Results were also compared with LVEF prediction based on clinical information and CMR parameters. Results ·In terms of clinical indicators, there were significant differences between the LVH and HC groups, such as LVEDV and LVEDM (all P<0.05); after extracting 3D radiomic features, the top 10 features were selected for further analysis. LR regression model, GB regression model and RF regression model were constructed for predicting the LVEF, and RF regression models showed the best results with seven features, in which the mean absolute error (MAE) was 0.066±0.002. Further comparison results showed that the model using radiomic information with CMR parameters (MAE=0.056±0.001) had the best performance and it was significantly better than the model using radiomic features (MAE=0.066±0.002) or CMR parameters (MAE=0.060±0.001) alone (both P<0.05). Conclusion ·The use of radiomic features for LVEF prediction has certain feasibility, and combining radiomic features with CMR parameters can further improve the prediction accuracy of the model.

Key words: cardiac magnetic resonance (CMR), radiomics, left ventricular hypertrophy, cine sequence

中图分类号:

R542.2

刘启明, 卢启帆, 柴烨子, 姜萌, 卜军. 基于影像组学特征预测左室射血分数的可行性研究[J]. 上海交通大学学报（医学版）, 2023, 43(9): 1162-1168.

LIU Qiming, LU Qifan, CHAI Yezi, JIANG Meng, PU Jun. Radiomics-based left ventricular ejection fraction prediction: a feasibility study[J]. Journal of Shanghai Jiao Tong University (Medical Science), 2023, 43(9): 1162-1168.

图/表 4

参考文献 31

1	HUYNH K. Heart failure: improvement of LVEF in patients with HF is linked to better prognosis[J]. Nat Rev Cardiol, 2016, 13(9): 505.
2	LEWEY J, LEVINE L D, ELOVITZ M A, et al. Importance of early diagnosis in peripartum cardiomyopathy[J]. Hypertension, 2020, 75(1): 91-97.
3	Lang RM, Badano LP, Mor-Avi V, et al. Recommendations for cardiac chamber quantification by echocardiography in adults: an update from the American Society of Echocardiography and the European Association of Cardiovascular Imaging[J]. J Am Soc Echocardiogr, 2015, 28(1): 1-39.e14.
4	OMMEN S R, MITAL S, BURKE M A, et al. 2020 AHA/ACC guideline for the diagnosis and treatment of patients with hypertrophic cardiomyopathy: executive summary: a report of the American college of cardiology/American heart association joint committee on clinical practice guidelines[J]. Circulation, 2020, 142(25): e533-e557.
5	LAVIE C J, HARMON K G. Routine ECG screening of young athletes[J]. J Am Coll Cardiol, 2016, 68(7): 712-714.
6	SHIGA T, WAJIMA Z, INOUE T, et al. Survey of observer variation in transesophageal echocardiography: comparison of anesthesiology and cardiology literature[J]. J Cardiothorac Vasc Anesth, 2003, 17(4): 430-442.
7	IBANEZ B, ALETRAS A H, ARAI A E, et al. Cardiac MRI endpoints in myocardial infarction experimental and clinical trials[J]. J Am Coll Cardiol, 2019, 74(2): 238-256.
8	KWAN A C, POURMORTEZA A, STUTMAN D, et al. Next-generation hardware advances in CT: cardiac applications[J]. Radiology, 2021, 298(1): 3-17.
9	YU X, YAO X X, WU B F, et al. Using deep learning method to identify left ventricular hypertrophy on echocardiography[J].Int J Cardiovasc Imaging, 2022, 38(4): 759-769.
10	SALIBA L J, MAFFETT S. Hypertensive heart disease and obesity: a review[J]. Heart Fail Clin, 2019, 15(4): 509-517.
11	GEORGIOPOULOS G, FIGLIOZZI S, PATERAS K, et al. Comparison of demographic, clinical, biochemical, and imaging findings in hypertrophic cardiomyopathy prognosis: a network meta-analysis[J]. JACC Heart Fail, 2023, 11(1): 30-41.
12	FUNADA A, KANZAKI H, NOGUCHI T, et al. Prognostic significance of late gadolinium enhancement quantification in cardiac magnetic resonance imaging of hypertrophic cardiomyopathy with systolic dysfunction[J]. Heart Vessels, 2016, 31(5): 758-770.
13	LYU Q, SHAN H M, XIE Y B, et al. Cine cardiac MRI motion artifact reduction using a recurrent neural network[J]. IEEE Trans Med Imaging, 2021, 40(8): 2170-2181.
14	LARKMAN D J, HERLIHY A H, COUTTS G A, et al. Elimination of magnetic field foldover artifacts in MR images[J]. J Magn Reson Imaging, 2000, 12(5): 795-797.
15	RAJIAH P, KAY F, BOLEN M, et al. Cardiac magnetic resonance in patients with cardiac implantable electronic devices: challenges and solutions[J]. J Thorac Imaging, 2020, 35(1): W1-W17.
16	FERREIRA P F, GATEHOUSE P D, MOHIADDIN R H, et al. Cardiovascular magnetic resonance artefacts[J]. J Cardiovasc Magn Reson, 2013, 15(1): 41.
17	GILLIES R J, KINAHAN P E, HRICAK H. Radiomics: images are more than pictures, they are data[J]. Radiology, 2016, 278(2): 563-577.
18	LAMBIN P, LEIJENAAR R T H, DEIST T M, et al. Radiomics: the bridge between medical imaging and personalized medicine[J]. Nat Rev Clin Oncol, 2017, 14(12): 749-762.
19	KELLER H, WANGER K C, GOEPFRICH M, et al. Morphological quantification and differentiation of left ventricular hypertrophy in hypertrophic cardiomyopathy and hypertensive heart disease. A two dimensional echocardiographic study[J]. Eur Heart J, 1990, 11(1): 65-74.
20	ZWANENBURG A, VALLIÈRES M, ABDALAH M A, et al. The image biomarker standardization initiative: standardized quantitative radiomics for high-throughput image-based phenotyping[J]. Radiology, 2020, 295(2): 328-338.
21	VAN GRIETHUYSEN J J M, FEDOROV A, PARMAR C, et al. Computational radiomics system to decode the radiographic phenotype[J]. Cancer Res, 2017, 77(21): e104-e107.
22	PEDREGOSA F, VAROQUAUX G, GRAMFORT A, et al. Scikit-learn: machine learning in python[EB/OL]. (2018-06-05) [2023-03-23]. https://arxiv.org/abs/1201.0490.
23	BUSTIN A, FUIN N, BOTNAR R M, et al. From compressed-sensing to artificial intelligence-based cardiac MRI reconstruction[J]. Front Cardiovasc Med, 2020, 7: 17.
24	VARRIANO G, GUERRIERO P, SANTONE A, et al. Explainability of radiomics through formal methods[J]. Comput Methods Programs Biomed, 2022, 220: 106824.
25	FAHMY A S, ROWIN E J, ARAFATI A, et al. Radiomics and deep learning for myocardial scar screening in hypertrophic cardiomyopathy[J]. J Cardiovasc Magn Reson, 2022, 24(1): 40.
26	RAISI-ESTABRAGH Z, JAGGI A, GKONTRA P, et al. Cardiac magnetic resonance radiomics reveal differential impact of sex, age, and vascular risk factors on cardiac structure and myocardial tissue[J]. Front Cardiovasc Med, 2021, 8: 763361.
27	WANG S, PATEL H, MILLER T, et al. AI based CMR assessment of biventricular function[J]. JACC Cardiovasc Imaging, 2022, 15(3): 413-427.
28	ZHENG X Y, YAO Z, HUANG Y N, et al. Deep learning radiomics can predict axillary lymph node status in early-stage breast cancer[J]. Nat Commun, 2020, 11(1): 1236.
29	BERNARD O, LALANDE A, ZOTTI C, et al. Deep learning techniques for automatic MRI cardiac multi-structures segmentation and diagnosis: is the problem solved?[J]. IEEE Trans Med Imaging, 2018, 37(11): 2514-2525.
30	CAMPELLO V M, MARTÍN-ISLA C, IZQUIERDO C, et al. Minimising multi-centre radiomics variability through image normalisation: a pilot study[J]. Sci Rep, 2022, 12(1): 12532.
31	PIANTADOSI G, SANSONE M, FUSCO R, et al. Multi-planar 3D breast segmentation in MRI via deep convolutional neural networks[J]. Artif Intell Med, 2020, 103: 101781.

Indicator	Total (n=200)	HC group (n=100)	LVH group (n=100)	P value
Demographic information
Age/year	53 (36, 62)	53 (36, 62)	53 (36, 63)	0.562
Male/n(%)	143 (71.5)	71 (71.0)	72 (72.0)	0.876
Weight/kg	72 (63, 80)	70±14	75 (65, 81)	0.014
Height/cm	170 (163, 175)	169±10	169±8	0.793
BMI/(kg·m^-2)	24.7 (22.3, 27.1)	24.2±3.5	26.0±4.0	0.000
BSA/m²	1.80±0.22	1.77±0.22	1.84 (1.70, 1.94)	0.051
CMR parameter
LVEF/%	65.1 (59.1, 70.2)	65.3±7.0	65.3 (55.2, 71.0)	0.336
LVEDV/mL	127.3 (107.1, 158.2)	125.7±34.7	134.8 (117.0, 167.8)	0.000
LVEDV/BSA/(mL·m^-2)	101.7 (80.1, 135.5)	70.2±14.2	75.9 (65.2, 91.8)	0.000
LVEDM/g	72.2 (62.7, 83.7)	83.7±24.0	134.1 (105.5, 175.6)	0.000
LVEDM/BSA/(g·m^-2)	55.5 (46.7, 71.9)	46.6±9.6	72.0 (59.5, 89.2)	0.000

Indicator	Total (n=200)	HC group (n=100)	LVH group (n=100)	P value
Demographic information
Age/year	53 (36, 62)	53 (36, 62)	53 (36, 63)	0.562
Male/n(%)	143 (71.5)	71 (71.0)	72 (72.0)	0.876
Weight/kg	72 (63, 80)	70±14	75 (65, 81)	0.014
Height/cm	170 (163, 175)	169±10	169±8	0.793
BMI/(kg·m^-2)	24.7 (22.3, 27.1)	24.2±3.5	26.0±4.0	0.000
BSA/m²	1.80±0.22	1.77±0.22	1.84 (1.70, 1.94)	0.051
CMR parameter
LVEF/%	65.1 (59.1, 70.2)	65.3±7.0	65.3 (55.2, 71.0)	0.336
LVEDV/mL	127.3 (107.1, 158.2)	125.7±34.7	134.8 (117.0, 167.8)	0.000
LVEDV/BSA/(mL·m^-2)	101.7 (80.1, 135.5)	70.2±14.2	75.9 (65.2, 91.8)	0.000
LVEDM/g	72.2 (62.7, 83.7)	83.7±24.0	134.1 (105.5, 175.6)	0.000
LVEDM/BSA/(g·m^-2)	55.5 (46.7, 71.9)	46.6±9.6	72.0 (59.5, 89.2)	0.000

Model	MAE	P value^①
LR	0.075±0.000	0.000
GB	0.072±0.000	0.000
RF	0.066±0.002	-

Model	MAE	P value^①
LR	0.075±0.000	0.000
GB	0.072±0.000	0.000
RF	0.066±0.002	-

Feature group^①	MAE	P value^②
Basic information^③	0.071±0.002	0.000
Radiomics	0.066±0.002	0.000
CMR^④	0.060±0.001	0.001
CMR+radiomics	0.056±0.001	NA

基于影像组学特征预测左室射血分数的可行性研究

Radiomics-based left ventricular ejection fraction prediction: a feasibility study

RichHTML

PDF (PC)

可视化

摘要/Abstract

引用本文

使用本文

图/表 4

参考文献 31

相关文章 15

编辑推荐

Metrics

[1]	马奔, 赵成, 束翌俊, 董平. CT影像组学在胃肠道间质瘤中的应用进展[J]. 上海交通大学学报（医学版）, 2023, 43(7): 923-930.
[2]	程然, 胡佳佳, 李彪. ¹⁸F-FDG PET/CT影像组学应用于淋巴瘤诊疗及预后预测的研究进展[J]. 上海交通大学学报（医学版）, 2023, 43(6): 781-787.
[3]	何俊林, 路青, 徐昕, 胡曙东. CT影像组学特征预测甲状腺乳头状癌颈部淋巴结转移的价值研究[J]. 上海交通大学学报(医学版), 2021, 41(9): 1233-1239.
[4]	胡培堃, 何杰, 吴连明, 葛恒, 许建荣, 卜军. ST段抬高型心肌梗死患者微血管阻塞对左室功能及预后的影响[J]. 上海交通大学学报(医学版), 2021, 41(2): 173-179.
[5]	冯泽豪, 柴烨子, 苏璇, 孙宝航行, 刘启明, 姜萌, 卜军. 体质量指数对系统性红斑狼疮患者心肌累及的影响[J]. 上海交通大学学报(医学版), 2021, 41(2): 180-186.
[6]	董建勋, 魏莱, 何杰, 孔令璁, 葛恒, 卜军. 心脏磁共振评估左心室机械不同步的研究进展[J]. 上海交通大学学报(医学版), 2021, 41(12): 1698-1702.
[7]	高亚洁, 马文坤, 高程洁, 周翌, 潘静薇. 心肌应变对急性ST段抬高型心肌梗死后心室重构的预测价值探讨[J]. 上海交通大学学报(医学版), 2021, 41(11): 1478-1484.
[8]	冯泽豪1，张清1，柴烨子1，苏璇1，孙宝航行1，刘启明1，严福华2，姜萌1#，卜军1#. 吸烟对急性ST段抬高型心肌梗死急性期心肌损伤及预后的影响[J]. 上海交通大学学报(医学版), 2020, 40(5): 573-582.
[9]	王玮，赵航，葛恒，丁嵩，沈学东，卜军. 二维斑点追踪超声心动图在评价急性心肌梗死后存活心肌及预测左心室重构中的价值[J]. 上海交通大学学报（医学版）, 2018, 38(12): 1447-.
[10]	吴昊，万青，高程洁，陶逸菁，夏智丽，魏盟，潘静薇 . 探讨梗阻性与非梗阻性肥厚型心肌病左心室应变力的差异[J]. 上海交通大学学报（医学版）, 2017, 37(5): 637-.
[11]	顾伟峰,卜军 . 心脏磁共振成像技术在冠状动脉粥样硬化性心脏病诊疗中的应用进展[J]. 上海交通大学学报（医学版）, 2017, 37(11): 1558-.
[12]	安东敖蕾,吴连明,葛恒,何奔,路青,胡伽尼，丁海燕,许建荣. T2 mapping在心肌梗死后心肌水肿检测中的应用以及与血清肌酸激酶间的相关性分析[J]. 上海交通大学学报（医学版）, 2016, 36(4): 532-.
[13]	黄斯怡，潘静薇，魏盟. 心脏磁共振在冠心病诊断及评估预后中的应用[J]. 上海交通大学学报（医学版）, 2015, 35(1): 132-.
[14]	宾锋利，王高峰，胡红杰. 双源CT心肌灌注成像临床应用的研究进展[J]. 上海交通大学学报（医学版）, 2013, 33(8): 1159-.
[15]	吴卫华, 黄艳, 陆静, 等. 二尖瓣环位移对肥厚性重构患者左室收缩功能的评估作用[J]. , 2011, 31(3): 313-.