中国组织工程研究 ›› 2022, Vol. 26 ›› Issue (33): 5323-5328.doi: 10.12307/2022.724

• 骨与关节图像与影像 bone and joint imaging • 上一篇    下一篇

MRI组学特征构建机器学习模型预测胸腰段再骨折

刘  进1,2,印宏坤3,陈  果2,张  宇2,顾祖超2,唐  静4   

  1. 1成都市第七人民医院骨科,四川省成都市   610041;2成都市第一人民医院骨科,四川省成都市   610041;3推想医疗科技股份有限公司,北京市   100080;4四川大学华西医院放射科,四川省成都市   610041
  • 收稿日期:2021-09-06 接受日期:2021-10-28 出版日期:2022-11-28 发布日期:2022-03-31
  • 通讯作者: 唐静,博士,副主任医师,四川大学华西医院放射科,四川省成都市 610041
  • 作者简介:刘进,男,1985年生,四川省都江堰市人,汉族,2021年四川大学毕业,博士,主治医师,主要从事脊柱外科、骨质疏松的诊断与治疗相关研究。
  • 基金资助:
    四川省卫生健康委员会科研项目(20PJ194),项目负责人:刘进;成都市卫生健康委员会科研项目(2020133),项目负责人:刘进

A machine learning prediction model based on MRI radiomics for refracture of thoracolumbar segments

Liu Jin1, 2, Yin Hongkun3, Chen Guo2, Zhang Yu2, Gu Zuchao2, Tang Jing4   

  1. 1Department of Orthopedics, Chengdu Seventh People’s Hospital, Chengdu 610041, Sichuan Province, China; 2Department of Orthopedics, Chengdu First People’s Hospital, Chengdu 610041, Sichuan Province, China; 3Beijing Infervision Technology Co., Ltd., Beijing 100080, China; 4Department of Radiology, West China Hospital, Sichuan University, Chengdu 610041, Sichuan Province, China
  • Received:2021-09-06 Accepted:2021-10-28 Online:2022-11-28 Published:2022-03-31
  • Contact: Tang Jing, MD, Associate chief physician, Department of Radiology, West China Hospital, Sichuan University, Chengdu 610041, Sichuan Province, China
  • About author:Liu Jin, MD, Attending physician, Department of Orthopedics, Chengdu Seventh People’s Hospital, Chengdu 610041, Sichuan Province, China; Department of Orthopedics, Chengdu First People’s Hospital, Chengdu 610041, Sichuan Province, China
  • Supported by:
    Scientific Research Project of Sichuan Provincial Health Commission, No. 20PJ194 (to LJ); Scientific Research Project of Chengdu Municipal Health Commission, No. 2020133 (to LJ)

摘要:

文题释义:
影像组学:由荷兰学者LAMBIN等于2012年首先提出,即指通过计算机软件从医学影像中提取影像定量特征,借助大数据手段解析其蕴含的临床信息以指导临床决策的研究方法。该方法可对人眼观察不到的图像变化引起的异质性进行量化,主要流程包括:原始影像获取、感兴趣区分割、特征提取和分析、模型构建及临床信息解读等,已被广泛应用于肿瘤等领域研究中,而在非肿瘤研究领域尚处于起步阶段。
机器学习:为人工智能领域的重要研究方法,是通过计算机基于数据构建概率统计模型并运用模型对数据进行预测与分析的一门技术,其算法模型主要包括决策树、随机森林、人工神经网络、贝叶斯学习等。

背景:影像组学能够对图像异质性进行量化,能否从骨质疏松椎体MRI影像中筛选出类似指纹等具有特征性的影像差异,用于预测再骨折的发生仍有待研究。
目的:探讨通过MRI组学特征联合临床信息构建椎体强化后胸腰段再骨折机器学习预测模型的可行性。
方法:回顾性收集成都市第一人民医院2014年5月至2019年4月由MRI确诊并行椎体强化治疗的骨质疏松性椎体压缩骨折患者资料,使用PyRadiomics工具提取强化前T11-L2节段非骨折椎体MRI T1序列影像组学特征。所有模型在训练集中构建,并在验证集中进行预测效能评估,采用最小绝对收缩和选择算子对组学数据进行降维,采用逻辑回归、随机森林和自适应提升算法针对临床信息、组学特征和二者结合构建相应的再骨折预测模型,采用受试者工作特征曲线对模型的诊断效能进行评估,采用决策分析曲线比较各模型的临床价值。
结果与结论:①共纳入135例患者的336个椎体,其中67个椎体发生再骨折,每个椎体分别提取到1 746个组学特征,经降维共获得13个重要特征;②在3种计算方法方案下,综合模型在训练集与验证集中的AUC均显著高于临床模型(P < 0.05),决策分析曲线同样显示综合模型预测胸腰段再骨折的净收益在大部分阈值区间内均高于临床模型;③结果表明,采用MRI T1序列影像组学特征联合临床信息构建再骨折预测模型具有可行性,有助于早期识别出具有高度再骨折风险的椎体。

https://orcid.org/0000-0002-5623-6106 (刘进) 

中国组织工程研究杂志出版内容重点:人工关节;骨植入物;脊柱;骨折;内固定;数字化骨科;组织工程

关键词: 影像组学, 骨质疏松, 椎体压缩骨折, 经皮椎体成形, 再骨折, 机器学习, 预测模型

Abstract: BACKGROUND: Radiomics can be used to quantify image heterogeneity. Whether radiomics can be used to screen out features such as fingerprint from MRI images of osteoporotic vertebral bodies to predict the occurrence of new fracture is worth studying.  
OBJECTIVE: To explore the feasibility of constructing a machine learning prediction model for thoracolumbar refracture after vertebral augmentation through combining MRI radiomics features and clinical information.
METHODS:  This study retrospectively collected the data of patients who were diagnosed with osteoporotic vertebral compression fracture by MRI and treated with percutaneous vertebral augmentation in Chengdu First People’s Hospital from May 2014 to April 2019. PyRadiomics was used to extract the imaging features of T1 sequences of vertebral MRI at the T11-L2 segments before percutaneous vertebral augmentation. All models were constructed in the training set, and prediction performance evaluation was performed in the validation set. Feature dimension reduction was conducted by applying least absolute shrinkage and selection operator regression. The corresponding refracture prediction models were constructed by multivariate logistic regression, random forest and adaptive lifting algorithm analysis using clinical parameters, selected features or the integrating of both. The diagnostic efficacy of the model was evaluated using the receiver operating characteristic curve. The decision analysis curve was used to compare the clinical value of each model.  
RESULTS AND CONCLUSION: (1) A total of 336 vertebrae were included in 135 patients, of which 67 vertebrae had refractures. 1 746 features were extracted from each vertebra, and 13 important features were obtained through dimension reduction. (2) Among the three models, area under curve of the combined model in the training set and validation set was significantly higher than that of the clinical model (P < 0.05), and the decision analysis curve also showed that the net benefit of the combined model in predicting thoracolumbar refracture was higher than that of the clinical model in most threshold intervals. (3) The results indicated that it was feasible to construct a refracture prediction model based on MRI T1 sequence imaging and clinical information, which could help to identify the vertebrae with high risk of refracture at early stage.

Key words: radiomics, osteoporotic, vertebral compression fracture, percutaneous vertebroplasty, refracture, machine learning, prediction model

中图分类号: