Virtual growing child 5-dimensional functional models for treating respiratory anomalies

用于治疗呼吸异常的虚拟成长儿童 5 维功能模型

• 批准号: 10532776
• 负责人: Drew Torigian
• 金额: $ 74.26万
• 依托单位: UNIVERSITY OF PENNSYLVANIA
• 依托单位国家: 美国
• 财政年份: 2020
• 资助国家: 美国
• 起止时间: 2020-02-01 至 2024-11-30
• 项目状态: 已结题
• 来源: https://reporter.nih.gov/project-details/10532776
• 关键词: 3-Dimensional; Address; Adult; Aftercare; Age; Algorithms; Anatomic Models; Anatomy; Applications Grants; Architecture; Birth; Chest; Chest imaging; Child; Childhood; Clinical; Clinical Data; Complex; Data; Databases; Deformity; Development; Dimensions; Disease; Face; Gender; Goals; Growth; Health Care Costs; Image; Incidence; Joints; Life; Lung; MRI Scans; Magnetic Resonance Imaging; Measurement; Methodology; Methods; Modeling; Names; Normalcy; Operative Surgical Procedures; Orthopedic Procedures; Outcome; Outcome Measure; Patient Selection; Patients; Population; Population Group; Postoperative Period; Procedures; Pulmonary function tests; Registries; Respiration; Rod; Scanning; Specialist; Specificity; Spinal; Spinal Fusion; Spirometry; Structure; Syndrome; Techniques; Technology; Testing; Thoracic Diseases; Tidal Volume; Time; Translating; Vertebral column; Vital capacity; age group; clinical outcome measures; cohort; deep learning; deep learning algorithm; health assessment; innovation; novel; prospective; pulmonary function; radiological imaging; respiratory; surgery outcome; treatment effect; ventilation; virtual

Thoracic Insufficiency Syndrome (TIS) is a group of serious disorders of the pediatric thorax resulting in an inability of the thorax to support respiration or lung growth. TIS is associated with at least 28 pediatric syndromes, with an estimated health care cost per patient that can easily exceed a million dollars. In TIS, three-dimensional deformity of the thoracic components anatomically and functionally reduces the volume available for ventilation. Pediatric specialists dealing with TIS currently face several serious challenges: (a) The complex interplay among dynamic and growing thoracic structures and its influence on thoracic function and growth are not understood at present. (b) The prime outcome measure for the corrective procedures has remained the radiographic Cobb angle of the spine, a 60-year old metric with poor correlation with lung dynamic function and limited true health assessment value. (c) A normative imaging database with functional metrics describing dynamics and growth of the thoracic structures of the normal pediatric population does not exist. Due to these hurdles, innovations in growth-modulating surgical techniques are difficult to achieve. Supported by extensive preliminary results based on dynamic MRI (dMRI) of patients and normal subjects, the overarching goal of this proposal is to develop novel dynamic functional metrics for TIS by establishing a normative database of dMRI images and anatomic and functional models and metrics, and to translate these to develop markers of TIS and of its corrective-surgery outcomes. The project has three aims. Aim 1: To develop a new methodology called The Virtual Growing Child (VGC) consisting of 4 key components: a) To build a normative database of dMRI images prospectively gathered from 200 normal children divided into 10 groups. b) To build population anatomic models involving key thoraco-abdominal objects following an established automatic anatomy recognition (AAR) technology and deep learning (DL) techniques. c) To develop and validate joint AAR-DL algorithms to segment these objects in dMRI images of TIS patients. d) To build a normative database of measurements derived from dMRI images describing normal thoracic architecture, dynamic function, and growth. The database will also include a full battery of Pulmonary Function Testing data and anthropometric measurements. Aim 2: To test retrospectively the utility of the VGC ensemble in deriving markers of TIS and its surgical treatment effects on a cohort of 100 TIS patients. Aim 3: To retrospectively test the utility of the VGC approach for planning surgery in 30 TIS patients by comparing VGC-guided surgical planning to the current planning method. The post-operative key dMRI parameters of patients whose surgical plan would have changed due to VGC data will be compared to those of patients whose plan did not change. Expected outcomes: (i) A unique registry of thoracic dMRI of 200 normal pediatric subjects, segmented objects, and the associated anatomic, dynamic, and developmental parameters. (ii) A validated VGC approach for studying TIS which can also be utilized for studying other pediatric and adult thoracic disorders.

胸腔功能不全综合征（TIS）是一组严重的小儿胸部疾病，导致 胸部无法支持呼吸或肺部生长。 Tis至少与28个小儿有关 综合征，每名患者的医疗保健费用估计很容易超过一百万美元。在tis中 胸部成分的三维畸形在解剖学上且功能减少体积 可供通风。目前正在处理TI的小儿专家面临几个严重的挑战：（a） 动态和成长的胸部结构之间的复杂相互作用及其对胸部功能的影响 目前尚不清楚增长。 （b）纠正程序的主要结果度量 仍然是脊柱的X光摄影角角，这是60岁的度量指标，与肺部相关性较差 动态功能和有限的真实健康评估价值。 （c）具有功能性的规范成像数据库 描述正常小儿种群胸部结构的动态和生长的指标不属于 存在。由于这些障碍，很难实现生长调节手术技术的创新。 基于患者和正常受试者的动态MRI（DMRI）的广泛初步结果支持， 该建议的总体目标是通过建立一个新的动态功能指标来建立一个 DMRI图像以及解剖和功能模型和指标的规范数据库，并将其转换为 开发TIS及其矫正手术结果的标记。该项目有三个目标。目标1：发展 一种称为虚拟成长儿童（VGC）的新方法，该方法由4个关键组成部分组成：a）建造一个 DMRI图像的规范数据库前瞻性地收集了200个正常儿童，分为10组。 b）建立涉及关键胸腔腹对物体的人口解剖模型 自动解剖学识别（AAR）技术和深度学习（DL）技术。 c）发展和 验证联合AAR-DL算法在TIS患者的DMRI图像中分割这些对象。 d）建造一个 源自DMRI图像的测量数据库，描述了正常的胸腔体系结构， 动态功能和增长。该数据库还将包含一系列肺功能测试数据 和人体测量值。目标2：回顾性测试VGC合奏的实用性 TIS的标记及其手术治疗对100例TIS患者的队列的影响。目标3：回顾性测试 VGC方法在30名TIS患者中进行计划手术的实用性通过比较VGC指导的手术 规划当前的计划方法。手术后的术后键DMRI参数 计划将由于VGC数据与计划不变的患者的数据进行比较。 预期结果：（i）胸腔DMRI的独特注册表，由200个普通小儿受试者进行分割 对象以及关联的解剖，动态和发展参数。 （ii）经过验证的VGC方法 用于研究TI，也可以用于研究其他儿科和成人胸部疾病。

项目成果

Automatic lung segmentation in dynamic thoracic MRI using two-stage deep convolutional neural networks.

使用两级深度卷积神经网络在动态胸部 MRI 中自动进行肺部分割。

• DOI: 10.1117/12.2612558
• 发表时间: 2022
• 期刊: Proceedings of SPIE--the International Society for Optical Engineering
• 作者: Xie,Lipeng;Udupa,JayaramK;Tong,Yubing;McDonough,JosephM;Wu,Caiyun;Lott,Carina;Anari,JasonB;Cahill,PatrickJ;Torigian,DrewA
• 通讯作者: Torigian,DrewA

Regional diaphragm motion analysis via dynamic MRI.

通过动态 MRI 进行区域膈肌运动分析。

• DOI: 10.1117/12.2611951
• 发表时间: 2022
• 期刊: Proceedings of SPIE--the International Society for Optical Engineering
• 作者: Hao,You;Udupa,JayaramK;Tong,Yubing;Wu,Caiyun;McDonough,JosephM;Lott,Carina;Clark,Abigail;Anari,JasonB;Cahill,PatrickJ;Torigian,DrewA
• 通讯作者: Torigian,DrewA

Automatic labeling of respiratory phases and detection of abnormal respiratory signals in free-breathing thoracic dynamic MR image acquisitions based on deep learning.

基于深度学习的自由呼吸胸部动态MR图像采集中呼吸相位的自动标记和异常呼吸信号的检测。

• DOI: 10.1117/12.2549983
• 发表时间: 2020
• 期刊: Proceedings of SPIE--the International Society for Optical Engineering
• 作者: Sun,Changjian;Udupa,JayaramK;Tong,Yubing;Wu,Caiyun;McDonough,JosephM;Qiu,Catherine;Lott,Carina;Anari,JasonB;Torigian,DrewA;Cahill,PatrickJ
• 通讯作者: Cahill,PatrickJ

QdMRI: A system for comprehensive analysis of thoracic dynamics via dynamic MRI.

QdMRI：通过动态 MRI 综合分析胸部动力学的系统。

• DOI: 10.1117/12.2612117
• 发表时间: 2022
• 期刊: Proceedings of SPIE--the International Society for Optical Engineering
• 作者: Tong,Yubing;Udupa,JayaramK;Hao,You;Xie,Lipeng;McDonough,JosephM;Wu,Caiyun;Lott,Carina;Clark,Abigail;Anari,JasonB;Torigian,DrewA;Cahill,PatrickJ
• 通讯作者: Cahill,PatrickJ

Estimation of the dynamic volume of each lung via rapid limited-slice dynamic MRI.

通过快速有限切片动态 MRI 估计每个肺的动态体积。

• DOI: 10.1117/12.2581952
• 发表时间: 2021
• 期刊: Proceedings of SPIE--the International Society for Optical Engineering
• 作者: Hao,You;Udupa,JayaramK;Tong,Yubing;Wu,Caiyun;Li,Hua;McDonough,JosephM;Lott,Carina;Anari,JasonB;Torigian,DrewA;Cahill,PatrickJ
• 通讯作者: Cahill,PatrickJ