Rapid Low-Cost Quantitative 3D MRI and Gait Assessment of the Knee

快速、低成本定量 3D MRI 和膝关节步态评估

• 批准号: 10671520
• 负责人: Brian Andrew Hargreaves
• 金额: $ 65.93万
• 依托单位: STANFORD UNIVERSITY
• 依托单位国家: 美国
• 财政年份: 2020
• 资助国家: 美国
• 起止时间: 2020-08-15 至 2025-06-30
• 项目状态: 未结题
• 来源: https://reporter.nih.gov/project-details/10671520
• 关键词: 3-Dimensional; Affect; American; Anterior Cruciate Ligament; Articulation; Bilateral; Biochemical; Biomechanics; Bone Marrow; Bone Spur; Cartilage; Chronic; Clinical; Cluster Analysis; Collaborations; Connective Tissue; Cost Measures; Coupled; Data; Data Pooling; Degenerative polyarthritis; Detection; Development; Diagnosis; Diagnostic; Diffusion; Diffusion Magnetic Resonance Imaging; Disease; Enrollment; Environment; Equipment; Etiology; Evaluation; Female; Femur; Fibrocartilages; Future; Gait; Gait abnormality; Goals; Health; Human Resources; Image; Inflammatory; Injections; Injury; Intervention; Joints; Kellgren-Lawrence grade; Knee; Knee Osteoarthritis; Left; Length; Lesion; Ligaments; Magnetic Resonance Imaging; Manuals; Maps; Measurement; Measures; Meniscus structure of joint; Methods; Morphology; Motion; Motivation; Orthopedics; Output; Pain; Patient Triage; Patients; Population; Protocols documentation; Protons; Quality of life; Quantitative Evaluations; Radiation; Reader; Replacement Arthroplasty; Research; Research Personnel; Risk; Risk Factors; Roentgen Rays; Sampling; Scanning; Sex Differences; Slice; Structure; Surface; Synovitis; Techniques; Tendon structure; Testing; Thick; Time; Tissues; Traumatic Arthropathy; Visualization; analysis pipeline; anterior cruciate ligament injury; anterior cruciate ligament rupture; automated analysis; biomechanical test; bone; cohort; cost; deep learning; density; disability; disease phenotype; disorder risk; early onset; effective therapy; gait examination; high body mass index; image reconstruction; improved; kinematics; learning classifier; meniscus injury; minimally invasive; novel; primary outcome; quantitative imaging; radiological imaging; reconstruction; research study; sensor; skills; societal costs; therapy development; tissue biomarkers; ultra high resolution

Project Abstract Motivation: Osteoarthritis (OA) is a painful disease that affects tens of millions of Americans, but is poorly understood, resulting in a lack of treatments. Enabling low-cost approaches for widespread study of risk factors, onset and early progression of OA will enable better understanding of OA mechanisms, treatment development, and triage of patients to different treatments based on specific disease phenotypes. Multiple systemic factors, biochemical factors, and other risk factors are associated with OA, but causes are diffi- cult to isolate and study during slow progression. Currently OA is diagnosed as joint-space narrowing using X-ray radiography, at a stage well beyond when interventions can be effective. Magnetic resonance imaging (MRI) of- fers sensitivity to morphologic and biochemical changes, but most methods are impractical for widespread clinical or research use. Usually MRI exams study only one knee, precluding the opportunity to compare knees. Sim- ilarly, biomechanics assessment typically requires numerous tests using advanced and rarely-available equip- ment and time-intensive analysis by skilled personnel, making this a challenge for widespread use. We have shown rapid, simultaneous 3D scanning of both knees with quantitative relaxometry and diffusion map- ping of connective tissues, combined with novel visualization of longitudinal change validated in a population with anterior cruciate ligament (ACL) tears. We have developed fully-automated cartilage and meniscus seg- mentation to simplify post-processing. (Our automated cartilage segmentation variability approaches that of reader-to-reader variability.) We now propose to combine MRI acquisition, reconstruction and analysis tech- niques with simple measures of kinematics into a widely applicable low-cost imaging and biomechanical test, which we will validate in subjects with ACL-injury and subjects with varying Kellgren-Lawrence grades of OA. Approach: We will begin by developing a robust 5-to-8-minute bilateral knee MRI exam, using an efficient 3D isotropic acquisition and novel deep-learning based image reconstructions. This will be followed with automated cartilage segmentation and quantitative analysis (thickness, T2, diffusion) of all 3 knee plates and automated semiquantitative scoring approaches for synovitis, bone marrow and cartilage lesions. Inertial measurement units (IMUs) will be used to measure kinematics, and gait asymmetries. We will continue our studies in ACL pa- tients to validate techniques and to develop asymmetry analyses for both imaging and biomechanical measures. Finally, in subjects with varying OA grade, we will evaluate the potential of the overall low-cost approach to relate asymmetry and longitudinal change measures to progression and OA grade. Significance: This project will develop an acquisition and analysis pipeline to quantify knee changes and left/right asymmetries that precede OA. We will characterize methods in idiopathic OA subjects and ACL- injured subjects at risk of post-traumatic OA. The very low target cost, under $120/subject, will ultimately enable widespread study of early onset and progression of different OA types, leading to earlier and better treatments.

项目摘要 动机：骨关节炎（OA）是一种痛苦的疾病，影响着数千万美国人， 了解，导致缺乏治疗。使广泛研究风险因素的低成本方法成为可能， OA的发病和早期进展将使人们能够更好地理解OA的机制、治疗开发 并根据特定疾病表型将患者分诊至不同治疗。 多种全身性因素、生化因素和其他风险因素与OA相关，但病因各异。 在缓慢进展期间进行隔离和研究。目前，OA被诊断为关节间隙狭窄使用X线 放射线照相，在一个阶段远远超过干预措施可以有效。磁共振成像（MRI）- 对形态学和生化变化敏感，但大多数方法对广泛的临床应用不切实际。 或研究用途。通常MRI检查只研究一个膝关节，排除了比较膝关节的机会。西姆- 以前，生物力学评估通常需要使用先进和罕见的设备进行大量测试， 需要技术人员进行大量的时间和时间密集型分析，这使得广泛使用面临挑战。 我们已经展示了双膝的快速，同时3D扫描与定量松弛和扩散图- 结缔组织平片，结合在人群中验证的纵向变化的新可视化 前交叉韧带撕裂我们开发了全自动软骨和半月板分离器- 简化后处理。(Our自动软骨分割可变性接近 阅读者之间的可变性。）我们现在建议将联合收割机MRI采集、重建和分析技术结合起来， niques与简单的运动学措施到一个广泛适用的低成本成像和生物力学测试， 我们将在ACL损伤受试者和具有不同Kelling-Lawrence OA等级的受试者中验证。 方法：我们将开始，使用有效的3D成像技术，进行5-8分钟的双侧膝关节MRI检查。 各向同性采集和基于深度学习的新型图像重建。这将是继自动 所有3个膝关节板的软骨分割和定量分析（厚度、T2、扩散），并自动 滑膜炎、骨髓和软骨病变的半定量评分方法。惯性测量 单位（伊穆斯）将用于测量运动学和步态不对称性。我们将继续在ACL PA学习- 旨在验证技术，并为成像和生物力学测量开发不对称分析。 最后，在不同OA等级的受试者中，我们将评估整体低成本方法的潜力， 进展和OA分级的不对称性和纵向变化测量。 意义：该项目将开发一个采集和分析管道，以量化膝关节变化， 左/右不对称性先于OA。我们将描述特发性OA受试者和ACL的方法- 有创伤后OA风险的受伤受试者。非常低的目标成本（低于120美元/受试者）将最终使 广泛研究不同类型OA的早期发病和进展，从而获得更早和更好的治疗。

项目成果

Detection of knee synovitis using non-contrast-enhanced qDESS compared with contrast-enhanced MRI.

• DOI: 10.1186/s13075-021-02436-8
• 发表时间: 2021-02-13
• 期刊: Arthritis research & therapy
• 影响因子: 4.9
• 作者: de Vries BA;Breda SJ;Sveinsson B;McWalter EJ;Meuffels DE;Krestin GP;Hargreaves BA;Gold GE;Oei EHG
• 通讯作者: Oei EHG

Improving Data-Efficiency and Robustness of Medical Imaging Segmentation Using Inpainting-Based Self-Supervised Learning.

使用基于介入的自我监督学习来改善医学成像分割的数据效率和鲁棒性。

• DOI: 10.3390/bioengineering10020207
• 发表时间: 2023-02-04
• 期刊: Bioengineering (Basel, Switzerland)

Prospective Deployment of Deep Learning in MRI: A Framework for Important Considerations, Challenges, and Recommendations for Best Practices.

• DOI: 10.1002/jmri.27331
• 发表时间: 2021-08
• 期刊: Journal of magnetic resonance imaging : JMRI
• 作者: Chaudhari AS;Sandino CM;Cole EK;Larson DB;Gold GE;Vasanawala SS;Lungren MP;Hargreaves BA;Langlotz CP
• 通讯作者: Langlotz CP

Adapted large language models can outperform medical experts in clinical text summarization.

• DOI: 10.1038/s41591-024-02855-5
• 发表时间: 2023-09
• 期刊: Nature medicine
• 影响因子: 82.9
• 作者: Dave Van Veen;Cara Van Uden;Louis Blankemeier;Jean-Benoit Delbrouck;Asad Aali;Christian Blüthgen;A. Pareek;Malgorzata Polacin;William Collins;Neera Ahuja;C. Langlotz;Jason Hom;S. Gatidis;John M. Pauly;Akshay S. Chaudhari

Gadolinium-free assessment of synovitis using diffusion tensor imaging.

• DOI: 10.1002/nbm.4614
• 发表时间: 2022-01
• 期刊: NMR in biomedicine
• 影响因子: 2.9
• 作者: Sandford HJC;MacKay JW;Watkins LE;Gold GE;Kogan F;Mazzoli V
• 通讯作者: Mazzoli V