Department of Radiology, Massachusetts General Hospital, Harvard Medical School, Boston, MA, USA.
Department of Radiology, VA Boston Healthcare System, Boston University School of Medicine, Boston, MA, USA.
Osteoarthritis Cartilage. 2024 Jan;32(1):18-27. doi: 10.1016/j.joca.2023.10.005. Epub 2023 Oct 23.
This narrative review summarizes the original research in the field of in vivo osteoarthritis (OA) imaging between 1 January 2022 and 1 April 2023.
A PubMed search was conducted using the following several terms pertaining to OA imaging, including but not limited to "Osteoarthritis / OA", "Magnetic resonance imaging / MRI", "X-ray" "Computed tomography / CT", "artificial intelligence /AI", "deep learning", "machine learning". This review is organized by topics including the anatomical structure of interest and modality, AI, challenges of OA imaging in the context of clinical trials, and imaging biomarkers in clinical trials and interventional studies. Ex vivo and animal studies were excluded from this review.
Two hundred and forty-nine publications were relevant to in vivo human OA imaging. Among the articles included, the knee joint (61%) and MRI (42%) were the predominant anatomical area and imaging modalities studied. Marked heterogeneity of structural tissue damage in OA knees was reported, a finding of potential relevance to clinical trial inclusion. The use of AI continues to rise rapidly to be applied in various aspect of OA imaging research but a lack of generalizability beyond highly standardized datasets limit interpretation and wide-spread application. No pharmacologic clinical trials using imaging data as outcome measures have been published in the period of interest.
Recent advances in OA imaging continue to heavily weigh on the use of AI. MRI remains the most important modality with a growing role in outcome prediction and classification.
本叙述性综述总结了 2022 年 1 月 1 日至 2023 年 4 月 1 日期间体内骨关节炎（OA）成像领域的原始研究。
使用与 OA 成像相关的几个术语（包括但不限于“骨关节炎/OA”、“磁共振成像/MRI”、“X 射线”、“计算机断层扫描/CT”、“人工智能/AI”、“深度学习”、“机器学习”）在 PubMed 上进行搜索。本综述按主题组织，包括感兴趣的解剖结构和模态、人工智能、临床试验中 OA 成像的挑战以及临床试验和介入研究中的成像生物标志物。本综述排除了离体和动物研究。
有 249 篇与体内人类 OA 成像相关的出版物。在纳入的文章中，膝关节（61%）和 MRI（42%）是研究最多的解剖区域和成像方式。OA 膝关节的结构组织损伤存在明显的异质性，这一发现与临床试验的纳入具有潜在相关性。人工智能的使用继续迅速增加，应用于 OA 成像研究的各个方面，但在高度标准化数据集之外缺乏通用性限制了其解释和广泛应用。在研究期间，没有使用影像学数据作为结局指标的药物临床试验被发表。
OA 成像的最新进展继续严重依赖于人工智能的使用。MRI 仍然是最重要的模态，在结局预测和分类方面发挥着越来越重要的作用。
