Quantitative bone imaging biomarkers and joint space analysis of the articular Fossa in temporomandibular joint osteoarthritis using artificial intelligence models

Tamara Mackie; Najla Al Turkestani; Jonas Bianchi; Tengfei Li; Antonio Ruellas; Marcela Gurgel; Erika Benavides; Fabiana Soki; Lucia Cevidanes

doi:10.3389/fdmed.2022.1007011

. Author manuscript; available in PMC: 2022 Nov 18.
作者手稿；可在 PubMed 中央存取：2022 年 11 月 18 日Front Dent Med. Author manuscript; available in PMC: 2022 Nov 18.

Published in final edited form as: Front Dent Med. 2022 Sep 19;3:1007011. doi: 10.3389/fdmed.2022.1007011
发表在最终编辑形式为：口腔牙医杂志。2022 年 9 月 19 日；3:1007011。doi：10.3389/fdmed.2022.1007011 Published in final edited form as: Front Dent Med. 2022 Sep 19;3:1007011. doi: 10.3389/fdmed.2022.1007011

Quantitative bone imaging biomarkers and joint space analysis of the articular Fossa in temporomandibular joint osteoarthritis using artificial intelligence models
使用人工智能模型在颞下颌关节骨关节炎中对骨影像生物标志物进行定量分析以及关节窝关节空间的分析Quantitative bone imaging biomarkers and joint space analysis of the articular Fossa in temporomandibular joint osteoarthritis using artificial intelligence models

Tamara Mackie Tamara Mackie ¹, Najla Al Turkestani 1, Najla Al Turkestani ^1,^2,^*, Jonas Bianchi 1,2,*, Jonas Bianchi ³, Tengfei Li 3, Tengfei Li ⁴, Antonio Ruellas 4, Antonio Ruellas ⁵, Marcela Gurgel 5, Marcela Gurgel ¹, Erika Benavides 1, Erika Benavides ⁶, Fabiana Soki 6, Fabiana Soki ⁶, Lucia Cevidanes 6, Lucia Cevidanes ¹¹
塔玛拉·麦基 ¹ ，纳吉拉·阿尔·图尔克斯坦 ^1, ^2, ^* ，乔纳斯·比安基 ³ ，滕飞·李 ⁴ ，安托尼奥·鲁埃拉斯 ⁵ ，玛尔切拉·古尔格尔 ¹ ，埃里卡·本纳维德斯 ⁶ ，法比安娜·索基 ⁶ ，露西亚·塞维达尼斯 ¹

PMCID: PMC9673279 NIHMSID: NIHMS1846208 PMID: 36404987 PMCID: PMC9673279 NIHMSID: NIHMS1846208 PMID: 36404987

Abstract 摘要Abstract Abstract

Temporomandibular joint osteoarthritis (TMJ OA) is a disease with a multifactorial etiology, involving many pathophysiological processes, and requiring comprehensive assessments to characterize progressive cartilage degradation, subchondral bone remodeling, and chronic pain. This study aimed to integrate quantitative biomarkers of bone texture and morphometry of the articular fossa and joint space to advance the role of imaging phenotypes for diagnosis of Temporomandibular Joint Osteoarthritis (TMJ OA) in early to moderate stages by improving the performance of machine-learning algorithms to detect TMJ OA status. Ninety-two patients were prospectively enrolled (184 h-CBCT scans of the right and left mandibular condyles), divided into two groups: 46 control and 46 TMJ OA subjects. No significant difference in the articular fossa radiomic biomarkers was found between TMJ OA and control patients. The superior condyle-to-fossa distance (p < 0.05) was significantly smaller in diseased patients. The interaction effects of the articular fossa radiomic biomarkers enhanced the performance of machine-learning algorithms to detect TMJ OA status. The LightGBM model achieved an AUC 0.842 to diagnose the TMJ OA status with Headaches and Range of Mouth Opening Without Pain ranked as top features, and top interactions of VE-cadherin in Serum and Angiogenin in Saliva, TGF-β1 in Saliva and Headaches, Gender and Muscle Soreness, PA1 in Saliva and Range of Mouth Opening Without Pain, Lateral Condyle Grey Level Non-Uniformity and Lateral Fossa Short Run Emphasis, TGF-β1 in Serum and Lateral Fossa Trabeculae number, MMP3 in Serum and VEGF in Serum, Headaches and Lateral Fossa Trabecular spacing, Headaches and PA1 in Saliva, and Headaches and BDNF in Saliva. Our preliminary results indicate that condyle imaging features may be more important in regards to main effects, but the fossa imaging features may have a larger contribution in terms of interaction effects. More studies are needed to optimize and further enhance machine-learning algorithms to detect early markers of disease, improve prediction of disease progression and severity to ultimately better serve clinical decision support systems in the treatment of patients with TMJ OA.
颞下颌关节骨关节炎（TMJ OA）是一种多因素病因学疾病，涉及多种病理生理过程，需要进行全面评估以描述渐进性软骨退化、次骨重塑和慢性疼痛。本研究旨在整合骨纹理的定量生物标志物和关节窝及关节间隙的形态学，以提高早期至中度阶段颞下颌关节骨关节炎（TMJ OA）诊断的性能，通过改进机器学习算法来检测 TMJ OA 状态。共纳入 92 名患者（184 份右侧和左侧下颌关节突的 h-CBCT 扫描），分为两组：46 名对照组和 46 名 TMJ OA 患者。在 TMJ OA 和对照患者之间，关节窝的放射学生物标志物没有显著差异。病态患者的上关节窝到关节窝的距离（p < 0.05）显著较小。关节窝的放射学生物标志物的交互作用增强了机器学习算法检测 TMJ OA 状态的性能。LightGBM 模型的 AUC 为 0。诊断颞下颌关节骨关节炎（TMJ OA）状态时，头痛和无痛开口范围排名为顶级特征，血管细胞粘附素（VE-cadherin）在血清中的顶级交互作用，血管生成素（Angiogenin）在唾液中的顶级交互作用，唾液中的转化生长因子β1（TGF-β1）和头痛，性别和肌肉疼痛，唾液中的 PA1 和无痛开口范围，髁突灰度非均匀性和外侧窝短运行强调，唾液中的 TGF-β1 和外侧窝骨小梁数量，血清中的基质金属蛋白酶 3（MMP3）和血管内皮生长因子（VEGF），头痛和外侧窝骨小梁间距，头痛和唾液中的 PA1，头痛和唾液中的脑源性神经营养因子（BDNF）。初步结果表明，髁突影像特征可能在主要效应方面更为重要，但窝影像特征在交互作用效应方面可能具有更大的贡献。需要更多的研究来优化和进一步增强机器学习算法，以检测疾病的早期标志，提高疾病进展和严重程度的预测，最终更好地服务于治疗颞下颌关节骨关节炎患者的临床决策支持系统。颞下颌关节骨关节炎（TMJ OA）是一种具有多因素病因的疾病，涉及许多病理生理过程，需要全面评估以表征进行性软骨退化、软骨下骨重塑和慢性疼痛。本研究旨在整合关节窝和关节间隙的骨骼纹理和形态特征的定量生物标志物，通过提高机器学习算法检测 TMJ OA 状态的性能，推进成像表型在诊断早期至中度颞下颌关节骨关节炎（TMJ OA）的作用。前瞻性纳入 92 例患者（左右下颌髁 184 次 h-CBCT 扫描），分为两组： 46 例对照和 46 例 TMJ OA 受试者。在 TMJ OA 患者和对照组患者之间，关节窝影像学生物标志物没有显着差异。髁突到颅窝的上距离（p< 0.05）在患病患者中显著较小。关节窝放射组学生物标志物的交互作用增强了机器学习算法检测 TMJ OA 状态的性能。LightGBM 模型在诊断 TMJ OA 状态方面达到了 AUC 0.842，其中头痛和无痛张口范围被评为主要特征，血清中 VE-钙粘蛋白和唾液中血管生成素的主要相互作用，TGF-β唾液和头痛 1，性别和肌肉酸痛，唾液和 PA1 张口范围无痛，外侧髁灰度不均匀和侧颅窝短期强调，血清和外侧窝小梁数 TGF-β 1，血清和 VEGF 中的 MMP3 和血清中 VEGF，头痛和外侧颅窝小梁间距，头痛和唾液中的 PA1，以及唾液中的头痛和 BDNF。我们的初步结果表明，髁突成像特征在主效应方面可能更重要，但窝成像特征在交互效应方面可能具有更大的贡献。需要更多的研究来优化和进一步增强机器学习算法，以检测疾病的早期标志物，改进对疾病进展和严重程度的预测，最终更好地服务于治疗 TMJ OA 患者的临床决策支持系统。

Keywords: articular fossa, imaging biomarkers, hr-CBCT, joint space, temporomandibular osteoarthritis, artificial intelligence
关键词：关节窝，影像生物标志物，hr-CBCT，关节间隙，颞下颌关节骨关节炎，人工智能关键词：关节窝，成像生物标志物， hr-CBCT，关节空间，颞下颌骨关节炎，人工智能

Introduction 简介Introduction 介绍

The Diagnostic Criteria (DC) for TemporoMandibular Joint Disorders (TMD) have recently described the condition of TemporoMandibular Joint Osteoarthritis (TMJ OA) defined by Ahmad et al 2009 (1) as Degenerative Joint Disease (2). In this study, we use the 2009 term, TMJ OA, because this is a disease with a multifactorial etiology, involving many pathophysiological processes, and requires comprehensive assessments to characterize progressive cartilage degradation, subchondral bone remodeling, and chronic pain (3–5). TMJ OA – once thought to be a condition involving “wear and tear” over time – is now classified as a “low inflammatory arthritic condition” (6) and associated with inflammatory mediators that lead to proliferative and resorptive inflammatory response with overall destructive consequences on the structural components of the TMJ, such as its cartilage, bone, and synovium (7). The progression of TMJ-OA may be slow (8), and the initial stages may be subclinical until the disease process has advanced to chronical stages (9). The TMJ provides a unique model to study early bone changes in OA, as only a thin layer of fibrocartilage covers the articular bone surface in the TMJ condyle (10, 11). Numerous animal studies indicate that the bone microarchitecture (4, 5, 12) is an important factor in the OA pathogenesis initiation, preceding articular cartilage changes (12, 13), and should be investigated in human studies for early TMJ OA detection. As treatments to reverse the chronic damage of TMJ OA are for the most part unavailable and limited (14), it is clear that early diagnosis may provide the best opportunity to prevent extensive and permanent joint damage. Current diagnosis standard protocols recommended in the DC/TMD criteria (1, 2) are based on pre-existent condylar damage, such as subcortical cysts, surface erosions, osteophytes, or generalized sclerosis.
诊断标准（DC）中描述了由 Ahmad 等人于 2009 年（1）定义的颞下颌关节骨关节炎（TMJ OA）为退行性关节病（2）。在本研究中，我们使用 2009 年的术语 TMJ OA，因为这是一种多因素病因的疾病，涉及许多病理生理过程，并需要全面评估来表征渐进性软骨退化、骨下结构重塑和慢性疼痛（3-5）。过去认为 TMJ OA 是涉及“磨损和老化”的条件，现在被归类为“低炎症性关节病”（6），并与导致增生性和吸收性炎症反应的炎症介质相关，对颞下颌关节的结构成分，如其软骨、骨和滑膜，产生总体破坏性影响（7）。TMJ-OA 的进展可能缓慢（8），初期可能无症状，直到疾病过程发展到慢性阶段（9）。 TMJ 提供了一种独特的模型来研究 OA 早期骨变化，因为在 TMJ 关节软骨下只有薄薄的一层纤维软骨覆盖关节骨表面（10, 11）。众多动物研究表明，骨微结构（4, 5, 12）是 OA 发病机制启动的重要因素，先于关节软骨变化（12, 13），并且在人类研究中应被调查以早期检测 TMJ OA。由于治疗逆转 TMJ OA 的慢性损害在很大程度上不可用且有限（14），显然早期诊断可能提供防止广泛和永久性关节损伤的最佳机会。DC/TMD 标准诊断协议（1, 2）目前基于已存在的关节软骨损伤，如皮质下囊肿、表面侵蚀、骨刺或普遍硬化。颞下颌关节疾病（TMD）的诊断标准（DC）最近描述了颞下颌关节骨关节炎（TMJ OA）的状况，由 Ahmad 等人 2009 （1）定义为退行性关节病（2）。在这项研究中，我们使用 2009 年的术语 TMJ OA，因为这是一种具有多因素病因的疾病，涉及许多病理生理过程，需要全面评估以表征进行性软骨退化、软骨下骨重塑和慢性疼痛（3-5）。TMJ OA - 曾经被认为是一种涉及 “磨损 ”的疾病 - 现在被归类为“低炎症性关节炎病症” （6），并与导致增殖和再吸收炎症反应的炎症介质有关，对 TMJ 的结构成分产生整体破坏性后果，如软骨、骨骼和滑膜（7）。TMJ-OA 的进展可能很缓慢（8），初始阶段可能是亚临床的，直到疾病过程发展到慢性阶段（9）。TMJ 提供了一个独特的模型来研究 OA 的早期骨骼变化，因为只有一层薄薄的纤维软骨覆盖了 TMJ 髁的关节骨表面（10， 11）。大量动物研究表明，骨微结构（4， 5， 12）是 OA 发病机制启动的重要因素，先于关节软骨变化（12， 13），应在人体研究中进行研究以早期 TMJ OA 检测。由于逆转 TMJ OA 慢性损伤的治疗在很大程度上是不可用且有限的（14），很明显，早期诊断可能提供防止广泛和永久性关节损伤的最佳机会。DC/TMD 标准（1， 2）中推荐的当前诊断标准方案基于预先存在的髁损伤，例如皮质下囊肿、表面侵蚀、骨赘或全身硬化。

Radiomics is the conversion of digital medical images into mineable high-dimensional data (15) – it refers to the extraction and analysis of advanced quantitative imaging from medical images to diagnose and/or predict diseases. This process is motivated by the concept that biomedical images contain information that reflect underlying pathophysiology and that these relationships can be revealed via quantitative image analyses (15). With high-throughput computing, it is now possible to promptly obtain countless quantitative features from relatively new high resolution low radiation CBCT (hr-CBCT) (16), and new software applications, with a user-friendly interface, can now easily extract large amounts of quantitative features from hr-CBCT greyscale images (17, 18). A study conducted by Bianchi et al. using quantitative bone imaging biomarkers for diagnosis of TMJ OA from hr-CBCT scans of mandibular condyles showed differences in subchondral bone microstructure between control and TMJ OA groups, and that they provided an acceptable diagnostic performance for the diagnosis of TMJ-OA. This opens up the notion that these biomarkers could be clinically significant in recognizing early onset of TMJ OA and enabling early, conservative therapy (19).
放射组学是将数字医疗图像转换为可挖掘的高维数据（15）——它指的是从医疗图像中提取和分析高级定量成像，用于诊断和/或预测疾病。这一过程受到以下概念的推动：生物医学图像包含反映潜在病理生理学的信息，而这些关系可以通过定量图像分析揭示（15）。随着高性能计算的发展，现在有可能从相对新的高分辨率低辐射 CBCT（hr-CBCT）（16）中迅速获得无数定量特征，并且新的软件应用程序，具有用户友好的界面，现在可以轻松从 hr-CBCT 灰度图像中提取大量定量特征（17, 18）。Bianchi 等人进行的一项研究使用定量骨成像生物标志物，从下颌关节突的 hr-CBCT 扫描中诊断 TMJ OA，显示了控制组和 TMJ OA 组之间软骨下骨微结构的差异，并且它们提供了诊断 TMJ-OA 的可接受诊断性能。这开辟了这些生物标志物在识别颞下颌关节 OA 早期 onset 和启用早期、保守治疗的可能性（19）。影像组学是将数字医学图像转换为可挖掘的高维数据（15） – 它是指从医学图像中提取和分析高级定量成像以诊断和/或预测疾病。这个过程的动机是生物医学图像包含反映潜在病理生理学的信息，并且这些关系可以通过定量图像分析来揭示（15）。通过高通量计算，现在可以从相对较新的高分辨率低辐射 CBCT （hr-CBCT）（16）中迅速获得无数定量特征，并且具有用户友好界面的新软件应用程序现在可以轻松地从 hr-CBCT 灰度图像中提取大量定量特征（17， 18).Bianchi 等人使用定量骨成像生物标志物从下颌髁的 hr-CBCT 扫描中诊断 TMJ OA 的研究显示，对照组和 TMJ OA 组之间的软骨下骨微结构存在差异，它们为诊断 TMJ-OA 提供了可接受的诊断性能。这开启了这样一种概念，即这些生物标志物在识别 TMJ OA 的早期发作和实现早期保守治疗方面可能具有临床意义（19）。

This study seeks to investigate whether the inclusion of articular fossa data improves the performance of machine-learning algorithms to detect TMJ OA status. Dislocation of the mandibular condyle from the articular fossa (mimicking the absence of the condyle) results in the arrested development of the fossa (20). This suggests that normal fossa development depends on normal condyle development, and the fossa bony microstructure may show signs of TMJ OA comparable to the condyle. Literature on changes related to the articular fossa in patients with TMJ OA is limited to roof thickness and joint space narrowing (21–22). The present study aims specifically to evaluate whether the integration of condyle-to-fossa distances and quantitative bone texture and morphometry imaging biomarkers in the articular fossa improve the performance of machine-learning algorithms for the diagnosis of TMJ OA in early to moderate stages.
本研究旨在探讨关节窝数据的加入是否能提高机器学习算法检测下颌关节 OA 状态的性能。下颌关节头从关节窝脱位（模拟关节头的缺失）导致关节窝的发育停滞（20）。这表明正常的关节窝发育依赖于正常的关节头发育，关节窝的骨微结构可能显示出与关节头相似的下颌关节 OA 迹象。关于与关节窝变化相关的下颌关节 OA 患者文献仅限于顶壁厚度和关节间隙变窄（21-22）。本研究的目的是特别评估将关节头到关节窝的距离和关节窝内定量骨纹理和形态学成像生物标志物的整合是否能提高早期至中度阶段下颌关节 OA 诊断的机器学习算法的性能。本研究旨在调查纳入关节窝数据是否提高了机器学习算法检测 TMJ OA 状态的性能。下颌骨髁与关节窝脱位（模仿髁的缺失）导致窝发育停滞（20）。这表明正常的颅窝发育取决于正常的髁突发育，颅窝骨微观结构可能显示出与髁突相当的 TMJ OA 迹象。关于 TMJ OA 患者关节窝相关变化的文献仅限于屋顶厚度和关节间隙狭窄（21-22）。本研究旨在专门评估关节窝中髁突到颅窝距离与定量骨纹理和形态测量成像生物标志物的整合是否能提高机器学习算法在早期至中度诊断 TMJ OA 的性能。

Materials and methodsMaterials and methods 材料和方法

This study followed the STROBE guidelines for observational studies. This cross-sectional study was approved by the Institutional Review Board of the University of Michigan (HUM00113199). All patients signed an informed consent and agreed to participate.本研究遵循 STROBE 观察性研究指南。这项横断面研究得到了密歇根大学机构审查委员会（HUM00113199）的批准。所有患者均签署知情同意书并同意参加。

Study design and participantsStudy design and participants 研究设计和受试者

The following inclusion criteria were applied for all patients: age between 21 and 70 years, no history of systemic disease, no history of TMJ trauma, surgery, or recent TMJ injections, no current pregnancy, and no congenital bone or cartilage disease. The control subjects were recruited by advertisements placed in the University Of Michigan School Of Dentistry and at The University of Michigan Dentistry Hospital; potential participants were first screened by telephone interview. The TMJ OA patients were recruited at their appointment with the TMD specialist from the University of Michigan. A total of 92 patients were selected, for a total 184 h-CBCT scans of the mandibular condyles. All subjects were clinically evaluated by the same TMD specialist using the Diagnostic Criteria for Temporomandibular Disorders (DC/TMD) (1–2) They were then divided into two groups: a control group (n = 46 patients, 46 condyles) and a TMJ OA group (n = 46 patients, 46 condyles). The inclusion criteria for control subjects were no history of clinical signs/symptoms of TMD. The inclusion criteria for the TMJ OA group were the presence of TMJ pain for less than 10 years, with clinical signs and symptoms evaluated using the DC/TMD: TMJ noise during movement or function in the last 30 days and crepitus detected during mandibular excursive movements. The radiographic CBCT interpretation was conducted by two oral and maxillofacial radiologists to confirm the presence of TMJ OA and was positive for at least one of the following: subchondral cyst, erosion, generalized sclerosis, and/or osteophytes (1). The exclusion criteria for the TMJ OA group were subjects with more than 10 years since the diagnosis of TMJ OA, or condyles with severe stages of bone destruction, subchondral cyst, erosion, generalized sclerosis, and/or osteophytes. The subjects were age and sex matched, with a mean 36 ± 11.4 years for control subjects and 40.2 ± 13.1 years for TMJ OA patients; with 4 control and 4 TMJ OA male subjects. The majority of female subjects than male subjects corroborates the sex distribution reported in the literature (22, 23). This study data included 3 sources of diagnostic features: clinical, biomolecular (levels of proteins in serum and saliva), and imaging features.
所有患者的纳入标准如下：年龄在 21 至 70 岁之间，无系统性疾病史，无颞下颌关节创伤、手术或近期颞下颌关节注射史，无当前怀孕，无先天性骨骼或软骨疾病。对照组通过在密歇根大学牙科学院和密歇根大学牙科医院发布广告招募，潜在参与者首先通过电话访谈筛选。密歇根大学的 TMD 专家在患者预约时招募了颞下颌关节 OA 患者。总共选择了 92 名患者，共进行了 184 次下颌关节突的高分辨 CT 扫描。所有受试者均由同一 TMD 专家使用颞下颌关节障碍诊断标准（DC/TMD）进行临床评估（1-2）。然后将他们分为两组：对照组（46 名患者，46 个关节突）和颞下颌关节 OA 组（46 名患者，46 个关节突）。对照组的纳入标准是没有临床表现/症状的 TMD 病史。颞下颌关节骨关节炎组的纳入标准为，颞下颌关节疼痛持续时间少于 10 年，临床体征和症状评估使用 DC/TMD：在过去的 30 天内有运动或功能时的颞下颌关节噪音，以及在下颌运动时检测到的摩擦声。由两位口腔颌面放射科医生进行的放射学 CBCT 解读，以确认颞下颌关节骨关节炎的存在，并至少符合以下一项：骨下囊肿、侵蚀、广泛硬化和/或骨刺（1）。颞下颌关节骨关节炎组的排除标准为，诊断颞下颌关节骨关节炎超过 10 年，或髁突有严重的骨破坏阶段，骨下囊肿、侵蚀、广泛硬化和/或骨刺。受试者在年龄和性别上匹配，对照组平均年龄为 36±11.4 岁，颞下颌关节骨关节炎患者为 40.2±13.1 岁；有 4 名对照组和 4 名颞下颌关节骨关节炎男性受试者。女性受试者多于男性受试者，这与文献报道的性别分布相符（22, 23）。本研究数据包含 3 种诊断特征来源：临床特征、生物分子特征（血清和唾液中蛋白质水平）以及影像特征。所有患者均适用以下纳入标准：年龄在 21 至 70 岁之间，无全身性疾病史，无 TMJ 外伤、手术或近期 TMJ 注射史，目前无妊娠，无先天性骨或软骨疾病。对照受试者是通过在密歇根大学牙科学院和密歇根大学牙科医院投放的广告招募的;首先通过电话访谈筛选潜在参与者。TMJ OA 患者是在与密歇根大学的 TMD 专家预约时招募的。共选择 92 例患者，对下颌髁进行共 184 次 h-CBCT 扫描。所有受试者均由同一位 TMD 专家使用颞下颌关节疾病诊断标准（DC/TMD）（1-2）进行临床评估，然后将他们分为两组：对照组（n = 46 名患者，46 个髁突）和 TMJ OA 组（n = 46 名患者，46 个髁突）。对照受试者的纳入标准是无 TMD 临床体征/症状史。TMJ OA 组的纳入标准是 TMJ 疼痛存在时间少于 10 年，使用 DC/TMD 评估临床体征和症状：过去 30 天内运动或功能期间的 TMJ 噪声，以及在下颌骨展开运动期间检测到的捻发音。影像学 CBCT 解释由两名口腔颌面放射科医生进行，以确认 TMJ OA 的存在，并且至少对以下一项呈阳性：软骨下囊肿、糜烂、全身硬化和/或骨赘（1）。TMJ OA 组的排除标准是诊断为 TMJ OA 后 10 年以上的受试者，或髁突具有严重骨破坏、软骨下囊肿、糜烂、全身硬化和/或骨赘的受试者。受试者年龄和性别匹配，对照组平均 36 ± 11.4 岁，TMJ OA 患者平均 40.2 ± 13.1 岁;有 4 名对照和 4 名 TMJ OA 男性受试者。大多数女性受试者比男性受试者证实了文献中报道的性别分布（22， 23）。该研究数据包括 3 个诊断特征来源：临床、生物分子（血清和唾液中的蛋白质水平）和影像学特征。

Clinical signs and symptomsClinical signs and symptoms 临床体征和症状

The same investigator collected and measured the clinical signs and symptoms of the participants based on the DC/TMD criteria (1, 2). The variables measured and selected for further statistical analysis were: Age, Pain began in years - TMJ OA group only, Current Facial Pain -TMJ OA group only, Worst Facial Pain in last 6 months -TMJ OA group only, Average Pain -TMJ OA group only, Last 6 Months Distressed by Headaches, Last 6 Months Distressed by Muscle Soreness, Vertical Range Unassisted Mouth Opening Without Pain (mm), Vertical Range Unassisted Maximum (mm), Vertical Range Assisted Maximum (mm).
研究者根据 DC/TMD 标准收集并测量了参与者的临床症状和体征（1, 2）。进一步统计分析的变量包括：年龄，开始疼痛的年份 - 只限于颞下颌关节 OA 组，当前面部疼痛 - 只限于颞下颌关节 OA 组，过去 6 个月内最严重的面部疼痛 - 只限于颞下颌关节 OA 组，平均疼痛 - 只限于颞下颌关节 OA 组，过去 6 个月内因头痛困扰 - 只限于颞下颌关节 OA 组，过去 6 个月内因肌肉酸痛困扰 - 只限于颞下颌关节 OA 组，无痛状态下自然开口范围（mm），无痛状态下最大开口范围（mm），有辅助下的最大开口范围（mm）。同一研究者根据 DC/TMD 标准（1， 2）收集并测量了参与者的临床体征和症状。测量和选择用于进一步统计分析的变量是：年龄、几年开始的疼痛 - 仅限 TMJ OA 组，当前面部疼痛 - 仅限 TMJ OA 组，过去 6 个月内最严重的面部疼痛 - 仅限 TMJ OA 组，平均疼痛 - 仅限 TMJ OA 组，最近 6 个月因头痛而困扰，最近 6 个月因肌肉酸痛而困扰，垂直范围无痛无辅助张口（mm）、垂直范围无辅助最大（mm）、垂直范围辅助最大（mm）。

Biomolecular data 生物分子数据Biomolecular data 生物分子数据

The participants had 5 ml of venous blood collected by a trained nurse at the University of Michigan. The blood was centrifuged for 20 min at 1,000 RPM to separate only the serum that was then aliquoted in 2 ml Eppendorf tubes and stored at −8 °C. For the saliva collection, the participants received a 14 ml sterile test tube with a funnel inserted; they were instructed to tilt their head forward and drip the saliva off into the tube until 2 ml was collected. They were informed to not spit, talk, or swallow during this process. We evaluated 14 proteins (10) in serum and saliva associated with nociception, inflammation, angiogenesis and bone resorption: 6ckine, Angiogenin, BDNF, CXCL16, ENA-78, MMP-3, MMP-7, OPG, PAI-1, TGFb1, TIMP-1, TRANCE, VE-Cadherin and VEGF. However, the expression of 6ckine was below the limit of detection in the serum and saliva samples in this study, and MMP-3 was not expressed in saliva. Those proteins were selected in a previous study that detected these markers in the TMJ synovial fluid and saliva of OA patients, showing correlations with bone surface changes (24). Custom human quantibody protein microarrays obtained from RayBiotech, Inc. Norcross, GA, was used to quantitatively assess the saliva and serum samples for the 14 specific biomarkers. Each participant had duplicates run for the saliva and serum samples.参与者由密歇根大学训练有素的护士收集了 5 ml 静脉血。将血液以 1,000 RPM 离心 20 分钟，仅分离血清，然后将其分装在 2 ml Eppendorf 管中并储存在 -8 °C 下。对于唾液采集，参与者收到了一个 14 毫升的无菌试管，并插入了漏斗;他们被指示将头向前倾斜并将唾液滴入试管中，直到收集到 2 mL。他们被告知在此过程中不要吐痰、说话或吞咽。我们评估了血清和唾液中与伤害感受、炎症、血管生成和骨吸收相关的 14 种蛋白质（10）：6ckine、血管生成素、BDNF、CXCL16、ENA-78、MMP-3、MMP-7、OPG、PAI-1、TGFb1、TIMP-1、TRANCE、VE-Cadherin 和 VEGF。然而，在本研究中，6ckine 在血清和唾液样本中的表达低于检测限，MMP-3 在唾液中不表达。这些蛋白是在之前的一项研究中选择的，该研究在 OA 患者的 TMJ 滑液和唾液中检测到这些标志物，显示与骨表面变化相关（24）。使用从佐治亚州诺克罗斯的 RayBiotech， Inc. 获得的定制人 quantibody 蛋白微阵列用于定量评估唾液和血清样品中 14 种特异性生物标志物的检测。每个参与者都对唾液和血清样本进行了重复运行。

ImagingImaging 成像

All small field of view 0.08 mm isotropic voxel CBCT scans were acquired using a 3D Accuitomo scanner (J. Morita Mfg. Corp., Tokyo, Japan). The TMJ acquisition protocol was as follows: field of view (FOV) of 40 × 40 mm, 90 kVp, 5 mAs and scanning time of 30.8s. The limitation of the exposure to the smallest FOV possible is in accordance to the ALARA (as low as reasonably achievable) principle, and this radiation reduction to the patient, maintaining or even improving the level of precision and accuracy in the diagnosis, supports the concept “as low as diagnostically acceptable” (ALADA) (25). Imaging features of one condyle per patient were included to reduce possible bias due to non-specific side data in systemic biological samples and comorbidities, technical problems in the hr-CBCT image acquisition, and presence of unilateral TMJ OA. The detailed image analysis protocol provided applied to the articular fossa region is shown in Figure 1 and all imaging features included have previously been validated for the mandibular condyle by Bianchi, et al (2021) (19) using 3D Slicer (26) and ITK-SNAP (27) open-source software. Anterolateral and articular eminence volume of interest (VOI) articular fossa regions in the FOV (Figure 2A) were selected and extracted using the “crop-volume” module in 3D Slicer (Figure 2B) with 30 × 30 × 30 slices. The posterior regions of the articular fossa were not included due to the presence of air cells in the temporal bone samples and difficulty distinguishing from trabecular bone. A total of 23 surrogate imaging biomarkers were evaluated (19, 28–30), as described in the Table 1. The BoneTexture module in 3D Slicer was used to compute the bone imaging biomarkers and obtain the subchondral bone microstructure values. The software computation parameters were chosen based on the pilot calibration studies from Bianchi et al (2021). The following computational software parameters were selected: (1) for GLCM: mask “inside” value = 1; number of bins = 10; voxel intensity range min = −1,000, max = 2,500; neighborhood radius = 4; (2) for GLRLM: mask “inside” value = 1; number of bins = 10; voxel intensity range min = −1,000, max = 2,500; distance range min = 0, max = 1; neighborhood radius = 4. For bone morphometry (BM), the software parameters were threshold = 250 and neighborhood radius = 4. Five measurements of joint space (Figure 2C) were measured as condylar-to-fossa distances (anterior, anterolateral, medial, superior and posterior). The statistical analysis of the imaging protocol was performed using IBM SPSS Statistics version 27.0 (IBM Corp., Armonk, NY). With an interval of 2 weeks between repeated measures, intra-class correlation coefficients (ICC) were used to assess the study error of the method in the selection of VOIs and computation of radiomic and bone morphometry features, as well as repeatability and interobserver reproducibility of joint space measures. The t-test for independent samples was used to compare the TMJ OA and Control groups with Levene’s Test for Equality of Variances to determine the assumption for homogeneity of variance.所有小视场 0.08 毫米各向同性体素 CBCT 扫描均使用 3D Accuitomo 扫描仪（J. Morita Mfg. Corp.，日本东京）获取。TMJ 采集方案如下：40 × 40 mm 的视野（FOV），90 kVp，5 mAs，扫描时间为 30.8s。将暴露于尽可能小的 FOV 的限制符合 ALARA（合理实现的尽可能低）原则，并且这种对患者的辐射减少，保持甚至提高诊断的精度和准确性水平，支持“低至诊断可接受”（ALADA）的概念（25）。包括每位患者一个髁突的影像学特征，以减少由于全身生物样本和合并症中的非特异性侧面数据、hr-CBCT 图像采集中的技术问题以及单侧 TMJ OA 的存在而导致的可能偏倚。提供给关节窝区域的详细图像分析方案如图 1 所示，并且包含的所有成像特征之前已由 Bianchi 等人（2021）（19）使用 3D Slicer （26）和 ITK-SNAP （27）开源软件验证了下颌骨髁。使用 3D 切片器（图 2B）中的“裁剪体积”模块选择和提取 FOV 中的前外侧和关节隆起感兴趣体积（VOI）关节窝区域（图 2A），具有 30 × 30 × 30 个切片。由于颞骨样本中存在气室并且难以与小梁区分，因此不包括关节窝的后部区域。共评估了 23 种替代成像生物标志物（19， 28-30），如表 1 所述。使用 3D Slicer 中的 BoneTexture 模块计算骨成像生物标志物并获得软骨下骨微观结构值。软件计算参数是根据 Bianchi 等人（2021）的试点校准研究选择的。选择了以下计算软件参数：（1）对于 GLCM：掩码“内部”值 = 1;箱数 = 10;体素强度范围最小值 = −1,000，最大值 = 2,500;邻域半径 = 4;（2）对于 GLRLM：掩码 “inside” 值 = 1;箱数 = 10;体素强度范围最小值 = −1,000，最大值 = 2,500;距离范围 min = 0， max = 1;邻域半径 = 4。对于骨形态测量（BM），软件参数为阈值 = 250 和邻域半径 = 4。关节间隙的 5 次测量（图 2C）测量为髁突到颅窝的距离（前、前外侧、内侧、上和后）。使用 IBM SPSS Statistics 版本 27.0 （IBM Corp.， Armonk， NY）对成像方案进行统计分析。重复测量之间间隔 2 周，使用类内相关系数（ICC）评估该方法在选择 VOI 和放射组学和骨形态学特征计算方面的研究误差，以及关节间隙测量的可重复性和观察者间再现性。独立样本的 t 检验用于将 TMJ OA 和对照组与 Levene 方差相等检验进行比较，以确定方差同质性的假设。

Image processing workflow adapted from bianchi et al. (2021) using 3D Slicer and ITK-SNAP open source software. (A) hr-CBCT files were anonymized and compressed. (B) The condyle and articular fossa were segmented. (C) 3D Slicer was used to convert the segmented articular fossa volume to a 3D surface. (D) Using the “transform” module, a standardized spatial orientation for each 3D TMJ bones model was made. Left TMJ scans were mirrored to the right side. (E) The spatial orientation matrix created in the last step was applied to the TMJ scan. (F) Using the “crop-volume” tool, two regions of the articular fossa (anterolateral, and articular eminence) were selected. (G) Using the “BoneTexture” module, all of the radiomic variables were computed.
图像处理工作流程改编自 bianchi 等人（2021 年）使用 3D Slicer 和 ITK-SNAP 开源软件。（A）hr-CBCT 文件进行了匿名化和压缩。（B）髁突和关节窝被分割。（C）使用 3D Slicer 将分割的关节窝体积转换为 3D 表面。（D）使用“变换”模块，为每个 3D TMJ 骨骼模型创建标准化空间定位。左 TMJ 扫描被镜像到右侧。（E）在上一步中创建的空间定位矩阵应用于 TMJ 扫描。（F）使用“裁剪体积”工具，选择了关节窝的两个区域（前外侧，关节隆起）。（G）使用“BoneTexture”模块，计算了所有放射学变量。图像处理工作流程改编自 bianchi et al. （2021）使用 3D Slicer 和 ITK-SNAP 开源软件。（A） hr-CBCT 文件经过匿名化和压缩处理。（B）髁突和关节窝分段。（C）使用 3D 切片器将分割的关节窝体积转换为 3D 表面。（D）使用 “transform” 模块，为每个 3D TMJ 骨骼模型制作了一个标准化的空间方向。左侧 TMJ 扫描被镜像到右侧。（E）将上一步中创建的空间方向矩阵应用于 TMJ 扫描。（F）使用“裁剪体积”工具，选择了关节窝的两个区域（前外侧和关节隆起）。（G）使用 “BoneTexture” 模块，计算了所有的放射组学变量。

TMJ imaging protocol. (A) Small FOV hr-CBCT scans for TMJ imaging features analysis. Note that marked bone destruction is seen in the condyle and the articular fossa also shows erosion. (B) Volumes of interest (VOIs) in the lateral portion of the articular fossa and in the articular eminence regions. (C) Superior condylar-to-fossa distance as indicated by the blue axis line seen in the coronal and sagittal views.TMJ 成像协议。（A）用于 TMJ 成像特征分析的小 FOV hr-CBCT 扫描。请注意，在髁突中可以看到明显的骨破坏，关节窝也显示侵蚀。（B）关节窝外侧部分和关节隆起区域的感兴趣体积（VOI）。（C）髁突到颅窝的上距离，如冠状位和矢状位视图中的蓝色轴线所示。

TABLE 1.TABLE 1. 表 1.

Description of the variables for radiomic and bone morphometry features.放射组学和骨形态测量学特征的变量描述。

FeaturesFeatures 特征	VariablesVariables 变量	DefinitionsDefinitions 定义
Grey-Level Co-occurrence Matrix (GLCM)Grey-Level Co-occurrence Matrix (GLCM)灰度共现矩阵（GLCM）	EnergyEnergy 能源	Uniformity of the grey-level textural organization.灰度纹理组织的统一性。
	EntropyEntropy 熵	Randomization of the grey-level distribution.灰度分布的随机化。
	CorrelationCorrelation 相关	Grey-level linear dependence among the pixels.像素之间的灰度线性依赖性。
	Inverse Difference MomentInverse Difference Moment 反差矩	Local homogeneity of the grey-level distribution.灰度级分布的局部均匀性。
	InertiaInertia 惯性	Contrast between a pixel and its neighbor.像素与其相邻像素之间的对比度。
	Cluster ShadeCluster Shade Cluster Shade（群集阴影）	Skewness and uniformity of the grey-level distribution.灰度级分布的偏度和均匀性。
	Cluster ProminenceCluster Prominence 群集突出度	Skewness and asymmetry of the grey-level distribution.灰度级分布的偏度和不对称性。
	Haralick CorrelationHaralick Correlation Haralick 相关性	Linear dependence between the pixels.Linear dependence between the pixels.像素之间的线性依赖性。
Grey-Level Run Length Matrix (GLRLM)Grey-Level Run Length Matrix (GLRLM)灰度游程矩阵（GLRLM）	Short Run EmphasisShort Run Emphasis 短期强调	Distribution of short run lengths.Distribution of short run lengths. 短运行长度的分布。
	Long Run EmphasisLong Run Emphasis 长期强调	Distribution of long run lengths.Distribution of long run lengths. 长运行长度的分布。
	Grey Level Non UniformityGrey Level Non Uniformity 灰度级不均匀性	Variability of the grey-level intensity.灰度强度的变化。
	Run Length Non UniformityRun Length Non Uniformity 运行长度不均匀	Similarity of run lengths in the image.Similarity of run lengths in the image.图像中游程长度的相似性。
	Low Grey Level Run EmphasisLow Grey Level Run Emphasis 低灰度运行强调	Distribution of the lower grey-level values.较低灰度值的分布。
	High Grey Level Run EmphasisHigh Grey Level Run Emphasis 高灰度运行强调	Distribution of the higher grey-level values.较高灰度值的分布。
	Short Run Low GreyShort Run Low Grey Short Run 低灰	Joint distribution of shorter runJoint distribution of shorter run 短途联合分配
	Level Run EmphasisLevel Run Emphasis Level Run Emphasis	lengths with lower grey-level values.lengths with lower grey-level values.灰度值较低的长度。
	Short Run High GreyShort Run High Grey Short Run High Grey	Joint distribution of shorter runJoint distribution of shorter run 短途联合分配
	Level Run EmphasisLevel Run Emphasis Level Run Emphasis	lengths with higher grey-level values.lengths with higher grey-level values.具有较高灰度值的长度。
	Long Run Low GreyLong Run Low Grey Long Run Low 灰色	Joint distribution of long runJoint distribution of long run 长期联合分配
	Level Run EmphasisLevel Run Emphasis Level Run Emphasis	lengths with lower grey-level values.lengths with lower grey-level values.灰度值较低的长度。
	Long Run High Grey 长跑高灰Long Run High Grey Long Run High Grey	Joint distribution of long runJoint distribution of long run 长期联合分配
	Level Run EmphasisLevel Run Emphasis Level Run Emphasis	lengths with higher grey-level values.lengths with higher grey-level values.具有较高灰度值的长度。
Bone MorphometryBone Morphometry 骨形态测量	BV/TVBV/TV BV/电视	Ratio between bone volume and total volume.骨体积与总体积之间的比率。
	Tb.ThTb.Th Tb.Th	Trabecular thickness.Trabecular thickness. 小梁厚度。
	Tb.SpTb.Sp 结核病 SP	Trabecular separation.Trabecular separation. 小梁分离。
	Tb.NTb.N 结核病	Trabecular number.Trabecular number. 小梁数。
	BS/BVBS/BV BS/BV （英语） BS/BVBS/BV（英文）	Ratio between bone surface and bone volume.骨表面积和骨体积之间的比率。

Open in a new tab 在新选项卡中打开

Diagnostic performance of the markers in machine learning algorithms机器学习算法中标记物的诊断性能

The data from this study was incorporated into two artificial intelligence-based tools – TMJOAI (TMJ Osteoarthritis Artificial Intelligence) tool (31) that integrates biological, clinical and imaging data; and the TMJPI (TMJ Privileged Information) tool (32). The Learning Using Privileged Information (LUPI) implemented in the TMJPI tool uses biological data to train the machine learning model but classifies new patients based on clinical and imaging data only, which is the current standard of care. These tools are available in an open-source web system DSCI (Data Storage for Computation and Integration) used for data management with storage and integration of patient information from multiple sources (33).这项研究的数据被整合到两个基于人工智能的工具中——TMJOAI（TMJ 骨关节炎人工智能）工具（31），它集成了生物、临床和成像数据;和 TMJPI（TMJ 特权信息）工具（32）。在 TMJPI 工具中实现的使用特权信息学习（LUPI）使用生物数据来训练机器学习模型，但仅根据临床和成像数据对新患者进行分类，这是当前的护理标准。这些工具在开源 Web 系统 DSCI（用于计算和集成的数据存储）中可用，用于数据管理，存储和集成来自多个来源的患者信息（33）。

The TMJOAI tool approach included feature normalization, selection, and model evaluation. We normalized all features to have zero mean and one standard deviation. Next, we calculated the AUC (Area Under the Curve), p-value and q-value from a two-sample Mann-Whitney U test to evaluate the significance of each feature. Then, we performed cross-validation (CV) to avoid overfitting – 100 times five-fold CV – resulting in 500 models in total. Each subject was predicted by the ensemble (averaging) of 100 models whose training set did not include that subject. Top main effect features and interactions, filtered with AUC > 0.7 and AUC > 0.65, respectively, calculated from the training subjects were then fed into models to make diagnostic predictions. We trained Extreme Gradient Boosting (XGBoost) (34) and Light Gradient Boosting Machine (LightGBM) (35) machine learning models. For both XGBoost and LightGBM models, we fixed the depth D = 1, and tuned the iteration steps by further splitting the training subjects into training and validation subjects. The following metrics were calculated to evaluate the performances of the model: accuracy, precision, recall, F1-score, and AUC, where AUC was chosen as the evaluation criterion to measure the test’s discriminative ability, i.e., how good is the test in a given clinical situation, with an AUC > 0.7–0.8 as fair, 0.81–0.9 as good and 0.91–1 as very good (36).
TMJOAI 工具方法包括特征规范化、选择和模型评估。我们将所有特征规范化为均值为零和标准差为一。接下来，我们从两样本 Mann-Whitney U 测试计算 AUC（曲线下的面积）、p 值和 q 值，以评估每个特征的显著性。然后，我们进行了交叉验证（CV）以避免过拟合——100 次五折 CV，总共产生了 500 个模型。每个受试者通过 100 个模型的集合（平均值）进行预测，这些模型的训练集不包括该受试者。从训练受试者中计算出 AUC 大于 0.7 和 AUC 大于 0.65 的主效应特征和交互作用，然后将这些特征输入模型进行诊断预测。我们训练了极端梯度提升（XGBoost）（34）和轻量级梯度提升机（LightGBM）（35）机器学习模型。对于 XGBoost 和 LightGBM 模型，我们将深度 D 固定为 1，并通过进一步将训练受试者分为训练和验证受试者来调整迭代步骤。以下指标用于评估模型性能：准确率、精确度、召回率、F1 分数和 AUC，其中选择 AUC 作为评估标准来衡量测试的区分能力，即在特定临床情况下，测试的性能如何，AUC 大于 0.7-0.8 为公平，0.81-0.9 为良好，0.91-1 为非常良好（36）。TMJOAI 工具方法包括特征归一化、选择和模型评估。我们将所有特征标准化为具有零均值和一个标准差。接下来，我们从双样本 Mann-Whitney U 检验中计算了 AUC（曲线下面积）、p 值和 q 值，以评估每个特征的显著性。然后，我们进行了交叉验证（CV）以避免过拟合 – CV 的 100 倍 5 倍 – 总共有 500 个模型。每个主题都由 100 个模型的集合（平均）预测，这些模型的训练集不包括该主题。然后将分别用 AUC > 0.7 和 AUC > 0.65 过滤的主要主效应特征和交互作用从训练对象中计算出来，然后输入模型以进行诊断预测。我们训练了 Extreme Gradient Boosting （XGBoost）（34）和 Light Gradient Boosting Machine （LightGBM）（35）机器学习模型。对于 XGBoost 和 LightGBM 模型，我们固定了深度 D = 1，并通过进一步将训练对象分为训练对象和验证对象来调整迭代步骤。计算以下指标以评估模型的性能：准确度、精密度、召回率、F1 分数和 AUC，其中 AUC 被选为衡量测试鉴别能力的评价标准，即在给定临床情况下测试的好坏，AUC > 0.7-0.8 为一般，0.81-0.9 为好，0.91-1 为非常好（36）。

The TMJPI tool approach tested the performance of RVFL and KRVFL+ models using biological data as privileged information (32). Considering that biological data is not routinely acquired for TMJ OA patients, we performed five-fold cross-validation and hyper-parameter tuning using a grid-search approach, utilized feature selection approaches such as normalized mutual information feature selection (NMIFS), MRMR (maximum relevancy minimum redundancy) and calculated Shapley Additive explanations values to rank features by their importance (37). We tested the performance of the TMJPI model using AUC, F1-score, sensitivity, specificity, precision, accuracy.TMJPI 工具方法使用生物数据作为特权信息来测试 RVFL 和 KRVFL+ 模型的性能（32）。考虑到 TMJ OA 患者不是常规获取生物数据，我们使用网格搜索方法进行了五重交叉验证和超参数调整，利用特征选择方法，如归一化互信息特征选择（NMIFS）、MRMR（最大相关性最小冗余）和计算的 Shapley 加法解释值来按特征的重要性对特征进行排名（37).我们使用 AUC 、 F1 评分、灵敏度、特异性、精度、准确性测试了 TMJPI 模型的性能。

ResultsResults 结果

In the articular eminence and anterolateral VOIs, 22 of the 23 proposed markers had an ICC value of greater than 0.8, indicating good repeatability of these values. In the articular eminence, the ICC value for Cluster Shade was 0.549, and in the anterolateral region, the ICC value for Correlation was 0.539, and these values were excluded from the machine learning models. The ICC values for all five distances in the 3D measurement were greater than 0.8, indicating good repeatability and reproducibility. Statistical significance was detected between patients exhibiting early to moderate stages of TMJ OA and control patients in the 3D measurement of the superior condyle-to-fossa distance (p = 0.013) with diseased patients exhibiting a smaller superior condyle-to-fossa distance.在关节隆起和前外侧 VOI 中，23 个拟议标志物中有 22 个的 ICC 值大于 0.8，表明这些值具有良好的可重复性。在关节隆起处，Cluster Shade 的 ICC 值为 0.549，在前外侧区域，Correlation 的 ICC 值为 0.539，这些值被排除在机器学习模型中。3D 测量中所有五个距离的 ICC 值均大于 0.8，表明具有良好的可重复性和再现性。在髁上颅窝距离（p = 0.013）的 3D 测量中，表现出早期至中度 TMJ OA 的患者与对照患者之间检测到统计学意义，患病患者表现出较小的髁上颅窝距离。

Using the TMJOAI tool, we found that articular fossa radiomics, bone morphometry and joint space data improved the performance of machine learning models in detecting TMJ OA status mainly through interaction effects among the integrated features. The best performing machine learning model was LightGBM model, even better than XGBoost + LightGBM combined, with the highest AUCs and F1- scores. Our results in Table 2 show that the LightGBM model now implemented in the TMJOAI with these features and interactions achieves the accuracy of 0.804, AUC 0.842, and F1-score 0.804 to diagnose the TMJ OA status with 3,081 features interactions.
使用 TMJOAI 工具，我们发现关节窝影像组学、骨形态测量学和关节间隙数据通过集成特征之间的交互作用，提高了检测 TMJ OA 状态的机器学习模型的性能。表现最佳的机器学习模型是 LightGBM 模型，甚至优于 XGBoost + LightGBM 的组合，具有最高的 AUCs 和 F1-分数。表 2 中的结果表明，使用这些特征和交互作用的 LightGBM 模型现在在 TMJOAI 中实现，诊断 TMJ OA 状态的准确率为 0.804，AUC 为 0.842，F1 分数为 0.804，共有 3,081 个特征交互。使用 TMJOAI 工具，我们发现关节窝影像组学、骨形态测量和关节间隙数据主要通过集成特征之间的交互效应提高了机器学习模型检测 TMJ OA 状态的性能。表现最好的机器学习模型是 LightGBM 模型，甚至优于 XGBoost + LightGBM 的总和，具有最高的 AUC 和 F1- 分数。我们在表 2 中的结果表明，现在在 TMJOAI 中实现的具有这些特征和交互的 LightGBM 模型实现了 0.804、AUC 0.842 和 F1 分数 0.804 的准确率，以 3,081 个特征交互来诊断 TMJ OA 状态。

TABLE 2.TABLE 2. 表 2.

Comparison metrics for the final model in TMJOAI and TMJPI. Multi-class precision and recall for the TMJOAI models are shown as Precision1 and Recall1 for OA and Precision0 and Recall0 for control groups, respectively, and the final F1-score was calculated by taking the macro average of the two classes’ F-1 scores (38, 39).TMJOAI 和 TMJPI 中最终模型的比较指标。TMJOAI 模型的多类精度和召回率分别显示为 OA 的 Precision1 和 Recall1 以及对照组的 Precision0 和 Recall0，最终的 F1 分数是通过取两个类的 F-1 分数的宏平均值（38， 39）来计算的。

TMJOAI ModelsTMJOAI Models TMJOAI 模型	AUCAUC AUC	AccuracyAccuracy 准确性	Precision1Precision1 精度1	Precision0Precision0 精度0	Recall1Recall1 召回1	Recall0Recall0 召回0	F1 ScoreF1 Score F1 分数
XGBoostXGBoost XGBoost	0.8290.829	0.7720.772	0.7780.778	0.7660.766	0.7610.761	0.7830.783	0.7720.772
LightGBMLightGBM LightGBM	0.8420.842	0.8040.804	0.8040.804	0.8040.804	0.8040.804	0.8040.804	0.8040.804
XGBoost + LightGBMXGBoost + LightGBM XGBoost + LightGBM	0.8370.837	0.7830.783	0.7830.783	0.7830.783	0.7830.783	0.7830.783	0.7830.783
TMJPI ModelTMJPI Model TMJPI 模型	AUCAUC AUC	AccuracyAccuracy 准确性	PrecisionPrecision 精度	SensitivitySensitivity 敏感性	SpecificitySpecificity 特异性	F1 scoreF1 score F1 分数
KRVFL+KRVFL+ KRVFL+	0.8090.809	0.7090.709	0.7740.774	0.6270.627	0.7910.791	0.6610.661

Open in a new tab 在新选项卡中打开

The values for the AUC, p-value, and q-value for all features are shown in Figures 3A, 4A. Figure 3A shows the AUC (upper plot), p-value (middle plot) and q-value (lower plot) for each category of variables (biological, clinical, condylar radiomics, articular fossa radiomics, and joint space). Figure 5 shows the 12 features with >90% top contributions sum: Headaches, VE-cadherin in Serum and Angiogenin in Saliva, TGF-β1 in Saliva and Headaches, Gender and Muscle Soreness, PA1 in Saliva and Range of mouth opening without pain, Lateral Condyle Grey Level Non-Uniformity and Lateral Fossa Short Run Emphasis, Range of mouth opening without pain, TGF-β1 in Serum and Lateral Fossa Trabeculae number, MMP3 in Serum and VEGF in Serum, Headaches and Lateral Fossa Trabecular spacing, Headaches and PA1 in Saliva, and Headaches and BDNF in Saliva. Most of the features with significant AUC values are clinical or condylar radiomics; no fossa radiomic or joint space features are detected with AUC > 0.65 (Figure 3A). The highest AUC value for a main effect fossa radiomic or joint space features was the superior joint space distance (Figure 3A); the interaction of Headaches and Lateral Fossa Trabecular spacing, Lateral Condyle Grey Level Non-Uniformity and Lateral Fossa Short Run Emphasis, and TGF-β1 in Serum and Lateral Fossa Trabeculae number were found to significantly contribute to the prediction of TMJ OA status (Figures 3B, 4B, 6A).所有特征的 AUC 、 p 值和 q 值的值如图 3A 、 4A 所示。图 3A 显示了每类变量（生物、临床、髁突影像组学、关节窝影像组学和关节间隙）的 AUC（上图）、p 值（中间图）和 q 值（下图）。图 5 显示了 12 个特征，其中 >90% 的主要贡献总和：头痛、血清中的 VE-钙粘蛋白和唾液中的血管生成素、唾液和头痛中的 TGF-β1、性别和肌肉酸痛、唾液中的 PA1 和无痛的张口范围、外侧髁灰度不均匀和外侧窝短期强调、无痛的张口范围、血清和侧颅窝小梁数中的 TGF-β 1，血清中血清和 VEGF 中的 MMP3，头痛和外侧窝小梁间距，唾液中的头痛和 PA1，以及唾液中的头痛和 BDNF。大多数具有显著 AUC 值的特征是临床或髁突影像组学;AUC > 0.65 时未检测到窝放射组学或关节间隙特征（图 3A）。主效应窝放射组学或关节间隙特征的最高 AUC 值是上关节间隙距离（图 3A）;发现头痛和外侧窝小梁间距、外侧髁灰度水平不均匀性和外侧窝短期强调以及血清和外侧窝小梁数中的 TGF-β 1 的相互作用对 TMJ OA 状态的预测有显着贡献（图 3B、4B、6A）。

General association analysis of risk factors for 79 features (A) and 66 top interactions (B). (**A,B**) The outer circle shows the AUC, middle circle shows the p-values, and the inner circle shows the q-values for each single feature.79 个特征（A）和 66 个主要交互作用（B）的风险因素的一般关联分析。（**A，B**）外圈显示 AUC，中间圈显示 p 值，内圈显示每个特征的 q 值。

Graphic displays of 79 features (A) and 3,081 interactions (B). (**A,B**) The upper graphic shows the AUC, the middle graph shows the p-values, and the lower category shows the q-values for each category of features.79 个特征（A）和 3,081 个交互（B）的图形显示。（**A，B**）上图显示 AUC，中间图形显示 p 值，下部类别显示每个特征类别的 q 值。

ROC curves of diagnostic sensitivity and specificity for individual features with top mean importance and the mean prediction of XGBoost and LightGBM.具有最高平均重要性的单个特征的诊断敏感性和特异性的 ROC 曲线以及 XGBoost 和 LightGBM 的平均预测。

Top 12 features in the LightGBM prediction model. (A) Mean contribution (according to feature importance) greater than 90% for 100 times 5-fold CV. (B) Boxplots of normalized features to diagnose disease status.LightGBM 预测模型中的前 12 个特征。（A） 100 乘以 5 倍 CV 的平均贡献（根据特征重要性）大于 90。（B）用于诊断疾病状态的标准化特征的箱线图。

For articular fossa markers, prediction models show that the interaction between Lateral Condyle Grey Level Non-Uniformity and Lateral Fossa Short Run Emphasis, TGF-β1 in Serum and Lateral Fossa Trabeculae number, and Headaches and Lateral Fossa Trabecular spacing, are found to be top features for accurate diagnosis of early stages of this clinical condition (Figure 6A). After the selection of the best features and interactions (Figure 6A), Figure 6B displays the boxplots for comparison between OA and control groups with corresponding AUCs, further demonstrating performance in diagnosis of TMJ OA status. Figure 5B shows the ROC curves of diagnostic sensitivity and specificity for individual features with top mean importance and the mean prediction of XGBoost, LightGBM and their ensemble with LightGBM demonstrating the largest ROC curve and highest discriminative ability of the models and features (Table 2).对于关节窝标志物，预测模型显示外侧髁灰度不均匀性与外侧窝短期强调、血清和外侧窝小梁数中的 TGF-β 1 以及头痛和外侧窝小梁间距之间的相互作用，被发现是准确诊断这种临床状况早期阶段的主要特征（图 6A）。在选择最佳特征和交互（图 6A）之后，图 6B 显示了 OA 组和对照组之间用于比较的箱线图以及相应的 AUC，进一步展示了 TMJ OA 状态诊断的性能。图 5B 显示了具有最高平均重要性的单个特征的诊断敏感性和特异性的 ROC 曲线，以及 XGBoost、LightGBM 及其集成的平均预测，其中 LightGBM 展示了模型和特征的最大 ROC 曲线和最高的区分能力（表 2）。

Using TMJPI, when combining clinical, radiomic (condyle and fossa), and 3D joint space features, LUPI-based models with additional biological features significantly enhanced the model performance on clinical, joint space measurement, and condyle datasets. The best clinical performance was obtained with the KRVFL+ model, keeping all clinical criteria and applying feature selection on the condyle and joint space features (Evaluation metrics shown in Table 2). The Shapley ranking of features based on their importance indicated 12 top features: 3 Clinical (Headaches, Muscle Soreness, Vertical Range Unassisted Mouth Opening Without Pain), 7 condylar radiomics and morphometry (Trabecular thickness, ShortRunHighGreyLevelEmphasis, Cluster Prominence, Entropy, Correlation, InverseDifferenceMoment and Energy) and 2 joint space features (Superior and Medial).使用 TMJPI，当结合临床、放射组学（髁突和窝）和 3D 关节间隙特征时，具有额外生物学特征的基于 LUPI 的模型显着增强了模型在临床、关节间隙测量和髁突数据集上的性能。使用 KRVFL+ 模型获得最佳临床性能，保留所有临床标准并对髁突和关节间隙特征进行特征选择（评估指标如表 2 所示）。根据重要性对特征的 Shapley 排名表明了 12 个主要特征：3 个临床特征（头痛、肌肉酸痛、垂直范围无痛的无辅助张口）、7 个髁突影像组学和形态测量学（小梁厚度、ShortRunHighGreyLevelEmphasis、簇突出、熵、相关性、逆差矩和能量）和 2 个关节间隙特征（上部和内侧）。

DiscussionDiscussion 讨论

This study demonstrates the diagnostic performance of joint space distances and radiomic biomarkers of the subchondral bone in hr-CBCT scans of TMJ OA patients in the articular fossa region. Surrogate articular fossa bone morphometry and textural features were not significantly different between TMJ OA patients and controls, whereas the superior joint space was significantly smaller in TMJ OA patients. This may suggest that joint space narrowing in the superior region may serve as an early sign of TMJ OA as found in previous studies (22).本研究证明了关节间隙距离和软骨下骨放射组学生物标志物在关节窝区域 TMJ OA 患者的 hr-CBCT 扫描中的诊断性能。TMJ OA 患者和对照组的替代关节窝骨形态和纹理特征没有显著差异，而 TMJ OA 患者的上关节间隙显著较小。这可能表明上部区域的关节间隙变窄可能是 TMJ OA 的早期体征，如以前的研究所示（22）。

The inclusion of quantitative articular fossa radiomics and joint space to machine-learning algorithms proved to be useful in enhancing the performance of TMJ OA classifiers. While articular fossa imaging biomarkers alone may not be diagnostic of early disease stages, through interactions with condylar, clinical and biological changes, fossa features may serve to strengthen the performance of machine-learning algorithms. Headaches and Range of mouth opening without pain and interactions of VE-cadherin in Serum and Angiogenin in Saliva, TGF-β1 in Saliva and Headaches, Gender and Muscle Soreness; PA1 in Saliva and Range of mouth opening without pain, Lateral Condyle Grey Level Non-Uniformity and Lateral Fossa Short Run Emphasis, TGF-β1 in Serum and Lateral Fossa Trabeculae number, MMP3 in Serum and VEGF in Serum, Headaches and Lateral Fossa Trabecular spacing, Headaches and PA1 in Saliva, and Headaches and BDNF in Saliva were the top features/interactions to accurately diagnose early stages of this clinical condition. Three of these interactions include fossa components showing that the assessment of fossa markers proves useful in diagnosis, as shown in our results with TMJOAI interaction effects. Therefore, while the articular fossa markers alone are not ranked among the features with highest AUC (Figures 3A, 4A), many articular fossa feature interactions present higher AUC (Figures 3B, 4B). The prediction model shows that the interaction between Lateral Condyle Grey Level Non-Uniformity and Lateral Fossa Short Run Emphasis, TGF-β1 in Serum and Lateral Fossa Trabeculae number, Headaches and Lateral Fossa Trabecular spacing are found to be top contributing features for accurate diagnosis of early stages of this clinical condition (Figure 6). This finding is similar to that of Bianchi et al. (2020) (10) prediction models that showed that the interaction between VE-cadherin in Serum and Angiogenin in Saliva, Headaches and PA1 in Saliva, TGF-β1 in Saliva and Headaches, VE-cad_Sal*Headaches (AUC= 0.698), TGF-β1 in Saliva and Headaches, and PA1 in Saliva and Range of mouth opening without pain are top features with mean >80% contribution to the information gain in the XGBoost and LightGBM predictive models. Therefore, our preliminary results suggest that while the condyle imaging features may be more important in regard to main effect (Figure 3A), the fossa features may have a larger contributing factor in terms of interaction effects (Figure 3B) – though future studies with larger sample sizes are needed.
包括定量关节窝放射组学和关节间隙在内的量化关节窝放射组学和关节间隙的纳入机器学习算法中，证明了在增强 TMJ OA 分类器性能方面的有用性。虽然仅通过关节窝影像生物标志物可能无法诊断早期疾病阶段，但通过与髁突、临床和生物变化的相互作用，窝特征可能有助于增强机器学习算法的性能。头痛、无痛开口范围以及血清中 VE-cadherin、唾液中 Angiogenin、唾液中 TGF-β1、头痛、性别和肌肉疼痛；唾液中 PA1、无痛开口范围、外侧髁灰色等级非均匀性和外侧窝短运行强调、唾液中 TGF-β1、外侧窝骨小梁数量、血清中 MMP3 和血清中 VEGF、头痛、唾液中 Lateral Fossa Trabecular spacing、头痛、唾液中 PA1、头痛和唾液中 BDNF 是准确诊断此临床状况早期阶段的顶级特征/相互作用。这些交互作用包括腔室组件，表明腔室标记的评估在诊断中证明是有用的，如我们在 TMJOAI 交互作用效果中所展示的结果所示。因此，虽然仅腔室标记在最高 AUC 特征（图 3A、4A）中未排名，但许多腔室特征交互作用表现出更高的 AUC（图 3B、4B）。预测模型显示，外侧髁灰度非均匀性和外侧腔室短运行强调、血清中的 TGF-β1、外侧腔室骨小梁数量、头痛和外侧腔室骨小梁间距是准确诊断此临床状况早期阶段的顶级贡献特征（图 6）。这一发现与 Bianchi 等人（2020）（10）的预测模型相似，该模型表明 VE-cadherin 在血清中的交互作用、唾液中的 Angiogenin、头痛和唾液中的 PA1、血清中的 TGF-β1 和头痛、VE-cad_Sal*Headaches（AUC= 0.698), 唾液中的 TGF-β1 和头痛，以及唾液中的 PA1 和无痛开口范围是顶级特征，其对信息增益的平均贡献率超过 80%。在 XGBoost 和 LightGBM 预测模型中。因此，初步结果显示，尽管关节窝成像特征可能在主要效应方面更为重要（图 3A），但窝特征在交互作用效应方面可能具有更大的贡献因素（图 3B） - 但需要未来研究使用更大的样本量来验证。事实证明，将定量关节窝影像组学和关节间隙纳入机器学习算法有助于提高 TMJ OA 分类器的性能。虽然单独的关节窝成像生物标志物可能无法诊断早期疾病阶段，但通过与髁突、临床和生物变化的相互作用，窝特征可能有助于增强机器学习算法的性能。头痛和无痛的张口范围以及血清中 VE-钙粘蛋白和唾液中血管生成素的相互作用，唾液和头痛中的 TGF-β1，性别和肌肉酸痛;唾液中的 PA1 和无痛的张口范围、外侧髁灰度不均匀和外侧窝短期强调、血清和外侧窝小梁数中的 TGF-β 1、血清和 VEGF 中的 MMP3、头痛和外侧窝小梁间距、唾液中的头痛和 PA1 以及唾液中的头痛和 BDNF 是准确诊断这种临床状况早期阶段的主要特征/相互作用。其中三种相互作用包括窝成分，表明窝标志物的评估被证明对诊断有用，正如我们与 TMJOAI 交互效应的结果所示。因此，虽然单独的关节窝标志物没有被列为具有最高 AUC 的特征（图 3A、4A），但许多关节窝特征相互作用呈现更高的 AUC（图 3B、4B）。预测模型显示，外侧髁灰度不均匀性与外侧窝短期强调、血清中 TGF-β 1 和外侧窝小梁数量、头痛和外侧窝小梁间距之间的相互作用被发现是准确诊断这种临床状况早期阶段的主要贡献特征（图6）。这一发现与 Bianchi 等人（2020）（10）预测模型相似，该模型显示血清中的 VE-钙粘蛋白与唾液、头痛和 PA1 中的血管生成素、唾液和 PA1、唾液和头痛中的 TGF-β 1、VE-cad_Sal*头痛（AUC= 0.698）、唾液和头痛中的 TGF-β 1 以及唾液中的 PA1 和无痛张口范围之间的相互作用是主要特征，平均对信息增益的贡献为 >80%XGBoost 和 LightGBM 预测模型。因此，我们的初步结果表明，虽然髁突成像特征在主效应方面可能更重要（图 3A），但窝特征在相互作用效应方面可能具有更大的贡献因素（图 3B）——尽管未来需要更大样本量的研究。

In TMJPI, the machine learning models only tested the original features and the main effect of each feature in overall TMJ OA status, whereas in TMJOAI, the machine models also tested interactions between features. This could explain why model performance decreased in TMJPI with the inclusion of all radiomic features, as TMJOAI models showed that the radiomic contribution is predominantly through interaction effects. Currently, the TMJPI tool is limited in computational approaches for testing features interactions due to the fact that it is supercomputing intensive and it takes a long time to train the model with interactions built in. As our baseline sample recruitment continues, larger sample sizes will allow further training of non-LUPI and LUPI-based algorithms on the TMJ OA datasets using grid search and 5-fold cross-validation (CV) on the training set to determine the optimal hyperparameters and features for each algorithm.
在 TMJPI 中，机器学习模型仅测试了原始特征和每个特征在整体 TMJ OA 状态中的主要效应，而在 TMJOAI 中，机器模型还测试了特征之间的交互作用。这可以解释为什么在包括所有放射学特征后，TMJPI 模型的性能下降，因为 TMJOAI 模型表明放射学贡献主要通过交互效应。目前，TMJPI 工具在计算方法上受限于测试特征交互，因为它需要超级计算，并且在训练模型时构建交互需要很长时间。随着我们基础样本招募的继续，更大的样本量将允许在 TMJ OA 数据集上对非 LUPI 和 LUPI 为基础的算法进行进一步的训练，使用网格搜索和训练集上的 5 折交叉验证（CV）来确定每个算法的最佳超参数和特征。在 TMJPI 中，机器学习模型只测试了原始特征和每个特征在整体 TMJ OA 状态中的主效应，而在 TMJOAI 中，机器模型还测试了特征之间的交互。这可以解释为什么在包含所有放射组学特征的情况下，TMJPI 中的模型性能下降，因为 TMJOAI 模型显示放射组学贡献主要是通过交互效应。目前，TMJPI 工具在测试特征交互的计算方法方面受到限制，因为它是超级计算密集型的，并且需要很长时间来训练内置交互的模型。随着我们的基线样本招募继续进行，更大的样本量将允许在 TMJ OA 数据集上使用网格搜索和训练集上的 5 倍交叉验证（CV）进一步训练非 LUPI 和基于 LUPI 的算法，以确定每种算法的最佳超参数和特征。

A limitation of the study similar to that of Bianchi et al. (2021) (10) was the use of the DC/TMD (1, 2) imaging criteria to confirm the diagnosis of the TMJ OA; however, the hr-CBCT used has a voxel size of 0.08 mm³, showing higher resolution and details than described in the DC/TMD imaging data, which uses CT scans with 0.7–1 mm slice thickness. Even with the addition of radiographic criteria to the DC/TMD – the standardized and widely used protocols for TMJ OA assessment – there is still a reliance on subjective radiological interpretation of pre-existing bone changes and clinical symptoms (1, 2). Furthermore, the cross-sectional study design does not allow assessment of the disease progression and how different disease stages affect the proposed biomarkers. This study was conducted only at baseline – providing another classification of disease vs. control that is already available with imaging and clinical symptoms. However, the ultimate goal of this work is the longitudinal assessments that will follow that test the potential of these baseline predictor values to also be predictive of risk of disease progression. This is valuable in determining which subjects are at greater risk of worsening over time, or which subjects would respond better to conservative approaches such as a mouthguard or splint therapy. Therefore, these initial markers detected in this study can serve as surrogate markers to be tested in future studies of risk of disease progression. Future studies using the proposed machine learning models and longitudinal data will provide better information on the feature’s behavior and disease progression. However, a drawback currently is that feature extraction from Cone-Beam Computed Tomography (CBCT) images remains time consuming before this integrative model can be applied in larger scale studies. Automatization of image processing steps and further refinements in machine-learning algorithms to detect early markers of disease have the potential to improve prediction of disease progression and severity to ultimately better serve and treat patients with TMJ OA.
研究的局限性类似于 Bianchi 等人（2021 年）（10）的研究，使用 DC/TMD（1,2）成像标准来确认颞下颌关节 OA 的诊断；然而，所使用的 hr-CBCT 的体素大小为 0.08 mm ³ ，显示了比 DC/TMD 成像数据中使用的 CT 扫描更高的分辨率和细节，CT 扫描的切片厚度为 0.7–1 mm。即使在 DC/TMD 成像标准中增加了放射学标准，仍然依赖于对现有骨变化和临床症状的主观放射学解释（1,2）。此外，横断面研究设计无法评估疾病进展以及不同疾病阶段对提出的生物标志器的影响。这项研究仅在基线时进行，提供了与影像学和临床症状已有的疾病分类。然而，这项工作的最终目标是后续的纵向评估，以测试这些基线预测值是否也能预测疾病进展的风险。这些发现对于确定哪些受试者在时间上恶化风险更高，或者哪些受试者对保守方法，如牙套或夹板治疗反应更好，非常有价值。因此，本研究中检测到的初始标记可以作为未来研究疾病进展风险的替代标记进行测试。使用提议的机器学习模型和纵向数据的未来研究将提供关于特征行为和疾病进展的更好信息。然而，目前的一个缺点是，在应用此综合模型进行大规模研究之前，从锥形束计算机断层成像（CBCT）图像中提取特征仍然耗时。图像处理步骤的自动化和机器学习算法进一步改进以检测疾病早期标记，有可能提高疾病进展和严重程度的预测，最终更好地服务于和治疗颞下颌关节骨关节炎患者。与 Bianchi 等人（2021）（10）类似的研究的一个局限性是使用 DC/TMD （1， 2）成像标准来确认 TMJ OA 的诊断;然而，所使用的 hr-CBCT 的体素尺寸为 0.08 mm3，显示出比 DC/TMD 成像数据中描述的更高的分辨率和细节，后者使用 0.7-1 mm 切片厚度的 CT 扫描。即使在 DC/TMD 中增加了放射学标准（TMJ OA 评估的标准化和广泛使用的方案），仍然依赖于对预先存在的骨骼变化和临床症状的主观放射学解释（1， 2）。此外，横断面研究设计不允许评估疾病进展以及不同的疾病阶段如何影响拟议的生物标志物。这项研究仅在基线时进行——提供了另一种疾病与对照组的分类，该分类已经可用于影像学和临床症状。然而，这项工作的最终目标是随后的纵向评估，以测试这些基线预测值也可以预测疾病进展风险的潜力。这对于确定哪些受试者随着时间的推移恶化的风险更大，或者哪些受试者对保守方法（如护齿器或夹板疗法）的反应更好很有价值。因此，本研究中检测到的这些初始标志物可以作为替代标志物，在未来疾病进展风险研究中进行测试。使用拟议的机器学习模型和纵向数据的未来研究将提供有关该特征行为和疾病进展的更好信息。然而，目前的一个缺点是，在将这种综合模型应用于更大规模的研究之前，从锥形束计算机断层扫描（CBCT）图像中提取特征仍然很耗时。图像处理步骤的自动化和机器学习算法的进一步改进以检测疾病的早期标志物，有可能改善对疾病进展和严重程度的预测，最终更好地服务和治疗 TMJ OA 患者。

Conclusion 结论Conclusion 结论

Our results indicate that the condyle imaging features may be more important in regard to main effect; whereas, for interaction effect, the fossa features may play a crucial role in the diagnosis of TMJ OA. Narrowing of the superior joint space was observed in TMJ OA patients. We developed a methodology for extraction of articular fossa radiomics and joint space distances utilizing machine learning for a comprehensive integration and management of data from various sources to improve articular joint health and predict patient-specific TMJ OA status.我们的结果表明，髁突成像特征在主效应方面可能更重要;而对于交互效应，窝特征可能在 TMJ OA 的诊断中起关键作用。在 TMJ OA 患者中观察到上关节间隙变窄。我们开发了一种提取关节窝影像组学和关节间隙距离的方法，利用机器学习全面整合和管理来自各种来源的数据，以改善关节健康并预测患者特定的 TMJ OA 状态。

FundingFunding 资金

This work was supported by NIDCR R01DE024450.这项工作得到了 NIDCR R01DE024450 的支持。

FootnotesFootnotes 脚注

Ethics statementEthics statement 道德声明

The studies involving human participants were reviewed and approved by University of Michigan Institutional Review Board HUM00113199. The patients/participants provided their written informed consent to participate in this study.涉及人类参与者的研究由密歇根大学机构审查委员会 HUM00113199 审查和批准。患者/参与者提供了参与本研究的书面知情同意书。

Conflict of interestConflict of interest 利益冲突

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.作者声明，该研究是在没有任何可能被解释为潜在利益冲突的商业或财务关系的情况下进行的。

Data availability statementData availability statement 数据可用性声明

The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.支持本文结论的原始数据将由作者提供，不得无故保留。

ReferencesReferences 引用

1.Ahmad M, Hollender L, Anderson Q, Kartha K, Ohrbach R, Truelove E, et al. Research diagnostic criteria for temporomandibular disorders (RDC/TMD): development of image analysis criteria and examiner reliability for image analysis. Oral Surg Oral Med Oral Pathol Oral Radiol Endod. (2009) 107 (6):844–60. doi: 10.1016/j.tripleo.2009.02.023 [DOI] [PMC free article] [PubMed] [Google Scholar]
1. 阿哈迈德 M, 洛兰德 L, 安德森 Q, 卡尔塔 K, 奥尔巴赫 R, 特鲁洛夫 E, 等。颞下颌关节障碍（TMD）的研究诊断标准（RDC/TMD）：开发图像分析标准和图像分析的检查员可靠性。口腔外科，口腔医学，口腔病理学，口腔放射学，牙周病学。2009 年，107 卷（6）：844–60。doi: 10.1016/j.tripleo.2009.02.023 [DOI] [PMC 免费文章] [PubMed] [谷歌学者]1. [DOI] [PMC 免费文章] [PubMed] [Google Scholar]
2.Schiffman E, Ohrbach R, Truelove E, Look J, Anderson G, Goulet JP, et al. Diagnostic criteria for temporomandibular disorders (DC/TMD) for clinical and research applications: recommendations of the international RDC/TMD consortium network* and orofacial pain special interest group†. J Oral Facial Pain Headache. (2014) 28(1):6–27. doi: 10.11607/jop.1151 [DOI] [PMC free article] [PubMed] [Google Scholar]
2. Schiffman E, Ohrbach R, Truelove E, Look J, Anderson G, Goulet JP, 等. 临床上和研究应用的颞下颌关节障碍诊断标准（DC/TMD）：国际 RDC/TMD 联盟网络*和口腔面部疼痛特别兴趣小组†的建议. J Oral Facial Pain Headache. (2014) 28(1):6–27. doi: 10.11607/jop.1151 [ DOI] [ PMC 免费文章] [ PubMed] [ Google Scholar]2. [DOI] [PMC 免费文章] [PubMed] [Google Scholar]
3.Wang X, Zhang J, Gan Y, Zhou Y. Current understanding of pathogenesis and treatment of TMJ osteoarthritis. J Dent Res. (2015) 94(5):666–73. doi: 10.1177/0022034515574770 [DOI] [PubMed] [Google Scholar]
3. 王 X，张 J，甘 Y，周 Y. 对颞下颌关节骨关节炎病理解剖和治疗的当前理解. J Dent Res. (2015) 94(5):666–73. doi: 10.1177/0022034515574770 [ DOI] [ PubMed] [ Google Scholar] 3. [DOI] [PubMed] [Google Scholar]3. [DOI] [PubMed] [Google 学术搜索]
4.Embree M, Ono M, Kilts T, Walker D, Langguth J, Mao J, et al. Role of subchondral bone during early-stage experimental TMJ osteoarthritis. J Dent Res. (2011) 90(11):1331–8. doi: 10.1177/0022034511421930 [DOI] [PMC free article] [PubMed] [Google Scholar]4. [DOI] [PMC 免费文章] [PubMed] [Google 学术搜索]
5.Jiao K, Niu L, Wang M, Dai J, Yu S, Liu X, Wang J. Subchondral bone loss following orthodontically induced cartilage degradation in the mandibular condyles of rats. Bone. (2011) 48(2):362–71. doi: 10.1016/j.bone.2010.09.010 [DOI] [PubMed] [Google Scholar] 5. [DOI] [PubMed] [Google Scholar]5. [DOI] [PubMed] [谷歌学术搜索]
6.de Souza R, Lovato da Silva C, Nasser M, Fedorowicz Z, Al-Muharraqi M. Interventions for the management of temporomandibular joint osteoarthritis. Cochrane Database Syst Rev. (2012) 2012(4):CD007261. doi: 10.1002/14651858.cd007261.pub2 [DOI] [PMC free article] [PubMed] [Google Scholar]6. [DOI] [PMC 免费文章] [PubMed] [Google 学术搜索]
7.Berenbaum F. Osteoarthritis as an inflammatory disease (osteoarthritis is not osteoarthrosis!). Osteoarthritis Cartilage. (2013) 21(1):16–21. doi: 10.1016/j.joca.2012.11.012 [DOI] [PubMed] [Google Scholar] 7. [DOI] [PubMed] [Google Scholar]7. [DOI] [PubMed] [谷歌学术搜索]
8.Rousseau JC, Delmas PD. Biological markers in osteoarthritis. Nat Clin Pract Rheumatol. (2007) 3(6):346–56. doi: 10.1038/ncprheum0508 [DOI] [PubMed] [Google Scholar] 8. [DOI] [PubMed] [Google Scholar]8. [DOI] [PubMed] [谷歌学术搜索]
9.Su N, Liu Y, Yang X, Luo Z, Shi Z. Correlation between bony changes measured with cone beam computed tomography and clinical dysfunction index in patients with temporomandibular joint osteoarthritis. J Craniomaxillofac Surg. (2014) 42(7):1402–7. doi: 10.1016/j.jcms.2014.04.001 [DOI] [PubMed] [Google Scholar] 9. [DOI] [PubMed] [Google Scholar]9. [DOI] [PubMed] [谷歌学术搜索]
10.Bianchi J, de Oliveira Ruellas A, Gonçalves J, Paniagua B, Prieto J, Styner M, et al. Osteoarthritis of the temporomandibular joint can be diagnosed earlier using biomarkers and machine learning. Sci Rep. (2020) 10(1):8012. doi: 10.1038/s41598-020-64942-0 [DOI] [PMC free article] [PubMed] [Google Scholar]10. [DOI] [PMC 免费文章] [PubMed] [Google 学术搜索]
11.Alomar X, Medrano J, Cabratosa J, Clavero JA, Lorente M, Serra I, et al. Anatomy of the temporomandibular joint. Semin Ultrasound CT MR. (2007) 28(3):170–83. doi: 10.1053/j.sult.2007.02.002 [DOI] [PubMed] [Google Scholar]
11. 阿洛马尔 X, Medrano J, Cabratosa J, Clavero JA, Lorente M, Serra I, 等. 下颌关节解剖学. 《超声学、CT 和 MRI 研讨会》. (2007) 28(3):170–83. doi: 10.1053/j.sult.2007.02.002 [DOI] [PubMed] [谷歌学者] 11. [DOI] [PubMed] [Google Scholar]11. [DOI] [PubMed] [谷歌学术搜索]
12.Li G, Yin J, Gao J, Cheng T, Pavlos N, Zhang C, Zheng M. Subchondral bone in osteoarthritis: insight into risk factors and microstructural changes. Arthritis Res Ther. (2013) 15(6):223. doi: 10.1186/ar4405 [DOI] [PMC free article] [PubMed] [Google Scholar]
12. 李 G, 阴 J, 高 J, 成 T, Pavlos N, 张 C, 郑 M. 骨关节炎中的骨下骨：风险因素和微结构变化的见解. 关节炎研究与治疗. (2013) 15(6):223. doi: 10.1186/ar4405 [ DOI] [ PMC 免费文章] [ PubMed] [ Google Scholar]12. [DOI] [PMC 免费文章] [PubMed] [Google Scholar]
13.Brandt K. Yet more evidence that osteoarthritis is not a cartilage disease. Ann Rheum Dis. (2006) 65:1261–4. doi: 10.1136/ard.2006.058347 [DOI] [PMC free article] [PubMed] [Google Scholar]13. [DOI] [PMC 免费文章] [PubMed] [Google Scholar]
14.Tanaka E, Detamore MS, Mercuri LG. Degenerative disorders of the temporomandibular joint: etiology, diagnosis, and treatment. J Dent Res. (2008) 87(4):296–307. doi: 10.1177/154405910808700406 [DOI] [PubMed] [Google Scholar] 14. [DOI] [PubMed] [Google Scholar]14. [DOI] [PubMed] [谷歌学术搜索]
15.Gillies RJ, Kinahan PE, Hricak H. Radiomics: images are more than pictures, they are data. Radiology. (2016) 278(2):563–77. doi: 10.1148/radiol.2015151169 [DOI] [PMC free article] [PubMed] [Google Scholar]15. [DOI] [PMC 免费文章] [PubMed] [Google Scholar]
16.Brahim A, Jennane R, Riad R, Janvier T, Khedher L, Toumi H, Lespessailles E. A decision support tool for early detection of knee OsteoArthritis using x-ray imaging and machine learning: data from the OsteoArthritis initiative. Comput Med Imaging Graph. (2019) 73:11–8. doi: 10.1016/j.compmedimag.2019.01.007 [DOI] [PubMed] [Google Scholar] 16. [DOI] [PubMed] [Google Scholar]16. [DOI] [PubMed] [谷歌学术搜索]
17.Paniagua B, Ruellas AC, Benavides E, Marron S, Woldford L, Cevidanes L. Validation of CBCT for the computation of textural biomarkers. Proc SPIE Int Soc Opt Eng. (2015) 9417:94171B. doi: 10.1117/12.2081859 [DOI] [PMC free article] [PubMed] [Google Scholar]17. [DOI] [PMC 免费文章] [PubMed] [Google Scholar]
18.Bianchi J, Gonçalves J, Ruellas A, Vimort J, Yatabe M, Paniagua B, et al. Software comparison to analyze bone radiomics from high resolution CBCT scans of mandibular condyles. Dentomaxillofac Radiol. (2019) 48(6):20190049. doi: 10.1259/dmfr.20190049 [DOI] [PMC free article] [PubMed] [Google Scholar]18. [DOI] [PMC 免费文章] [PubMed] [Google 学术搜索]
19.Bianchi J, Gonçalves JR, de Oliveira Ruellas AC, Ashman LM, Vimort JB, Yatabe M, et al. Quantitative bone imaging biomarkers to diagnose temporomandibular joint osteoarthritis. Int J Oral Maxillofac Surg. (2021) 50 (2):227–35. doi: 10.1016/j.ijom.2020.04.018. [DOI] [PMC free article] [PubMed] [Google Scholar]19. [DOI] [PMC 免费文章] [PubMed] [Google 学术搜索]
20.Wang Y, Liu C, Rohr J, Liu H, He F, Yu J, et al. Tissue interaction is required for glenoid fossa development during temporomandibular joint formation. Dev Dyn. (2011) 240:2466–73. doi: 10.1002/dvdy.22748 [DOI] [PMC free article] [PubMed] [Google Scholar]
20. 王 Y, 刘 C, Rohr J, 刘 H, 何 F, 余 J, 等. 发育动力学. (2011) 240:2466–73. doi: 10.1002/dvdy.22748 [DOI] [PMC 免费文章] [PubMed] [谷歌学者]20. [DOI] [PMC 免费文章] [PubMed] [Google 学术搜索]
21.Soydan D, Doğan S, Canger E, Coşgunarslan A, Akgün I, Kış H. Effect of internal derangements and degenerative bone changes on the minimum thickness of the roof of the glenoid fossa in temporomandibular joint. Oral Radiol. (2020) 36(1):25–31. doi: 10.1007/s11282-019-00373-6 [DOI] [PubMed] [Google Scholar]
21. 草丹 D, 動 S, 江 E, 玄 A, 雲 I, 冬 H. 内部紊乱和退行性骨变化对颞下颌关节的骨窝顶最小厚度的影响. 口腔放射学. (2020) 36(1):25–31. doi: 10.1007/s11282-019-00373-6 [ DOI] [ PubMed] [ Google 学者] 21. [DOI] [PubMed] [Google Scholar]21. [DOI] [PubMed] [谷歌学术搜索]
22.Massilla Mani F, Sivasubramanian SS. A study of temporomandibular joint osteoarthritis using computed tomographic imaging. Biomed J. (2016) 39 (3):201–6. doi: 10.1016/j.bj.2016.06.003. [DOI] [PMC free article] [PubMed] [Google Scholar]22. [DOI] [PMC 免费文章] [PubMed] [Google 学术搜索]
23.Rando C, Waldron T. TMJ Osteoarthritis: a new approach to diagnosis. Am J Phys Anthropol. (2012) 148(1):45–53. doi: 10.1002/ajpa.22039 [DOI] [PubMed] [Google Scholar] 23. [DOI] [PubMed] [Google Scholar]23. [DOI] [PubMed] [谷歌学术搜索]
24.Cevidanes LH, Walker D, Schilling J, Sugai J, Giannobile W, Paniagua B, et al. 3D Osteoarthritic changes in TMJ condylar morphology correlates with specific systemic and local biomarkers of disease. Osteoarthritis Cartilage. (2014) 22(10):1657–67. doi: 10.1016/j.joca.2014.06.014 [DOI] [PMC free article] [PubMed] [Google Scholar]24. [DOI] [PMC 免费文章] [PubMed] [Google Scholar]
25.Kofler B, Jenetten L, Runge A, Degenhart G, Fischer N, Hörmann R, et al. ALADA dose optimization in the computed tomography of the temporal bone: the diagnostic potential of different low-dose CT protocols. Diagnostics (Basel). (2021) 11(10):1894. doi: 10.3390/diagnostics11101894 [DOI] [PMC free article] [PubMed] [Google Scholar]25. [DOI] [PMC 免费文章] [PubMed] [Google 学术搜索]
26.3D Slicer. https://www.slicer.org, accessed. July 23 2022.
26.3D Slicer. https://www.slicer.org, 访问日期. 2022 年 7 月 23 日. 26.
27.ITK-SNAP. http://www.itksnap.org, accessed. July 23 2022.
27.ITK-SNAP. http://www.itksnap.org, 2022 年 7 月 23 日访问. 27.
28.Shoukri B, Prieto J, Ruellas A, Yatabe M, Sugai J, Styner M, et al. Minimally invasive approach for diagnosing TMJ osteoarthritis. J Dent Res. (2019) 98 (10):1103–11. doi: 10.1177/0022034519865187 [DOI] [PMC free article] [PubMed] [Google Scholar]28. [DOI] [PMC 免费文章] [PubMed] [Google 学术搜索]
29.Nieminen M, Casula V, Nevalainen M, Saarakkala S. Osteoarthritis year in review 2018: imaging. Osteoarthritis Cartilage. (2019) 27(3):401–11. doi: 10.1016/j.joca.2018.12.009 [DOI] [PubMed] [Google Scholar] 29. [DOI] [PubMed] [Google Scholar]29. [DOI] [PubMed] [谷歌学术搜索]
30.Ebrahim F, Ruellas A, Paniagua B, Benavides E, Jepsen K, Wolford L, et al. Accuracy of biomarkers obtained from cone beam computed tomography in assessing the internal trabecular structure of the mandibular condyle. Oral Surg Oral Med Oral Pathol Oral Radiol, 2017. 124(6): p. 588–99. doi: 10.1016/j.oooo.2017.08.013 [DOI] [PMC free article] [PubMed] [Google Scholar]30. [DOI] [PMC 免费文章] [PubMed] [Google Scholar]
31.Le C, Deleat-Besson R, Turkestani NA, Cevidanes L, Bianchi J, Zhang W, et al. TMJOAI: An artificial web-based intelligence tool for early diagnosis of the temporomandibular joint osteoarthritis. Clin Image Based Proced Distrib Collab Learn Artif Intell Combat COVID 19 Secur Priv Preserv Mach Learn, 2021 Sep-Oct;12969: (2021). pp. 78–87. doi: 10.1007/978-3-030-90874-4_8. [DOI] [PMC free article] [PubMed] [Google Scholar]31. [DOI] [PMC 免费文章] [PubMed] [Google Scholar]
32.Zhang W, Bianchi J, Turkestani NA, Le C, Deleat-Besson R, Ruellas A, et al. Temporomandibular joint osteoarthritis diagnosis using privileged learning of protein markers. Annu Int Conf IEEE Eng Med Biol Soc. (2021) 2021:1810–3. doi: 10.1109/EMBC46164.2021.9629990 [DOI] [PMC free article] [PubMed] [Google Scholar]32. [DOI] [PMC 免费文章] [PubMed] [Google 学术搜索]
33.Brosset S, Dumont M, Cevidanes L, Soroushmehr R, Bianchi J, Gurgel M, et al. Web infrastructure for data management, storage and computation. Proc SPIE Int Soc Opt Eng. (2021) 11600:116001N. doi: 10.1117/12.2582283 [DOI] [PMC free article] [PubMed] [Google Scholar]33. [DOI] [PMC 免费文章] [PubMed] [Google Scholar]
34.Chen T, Guestrin C. Xgboost: A scalable tree boosting system. In Proceedings of the 22nd Acm Sigkdd International Conference on Knowledge Discovery and Data Mining (ACM, 2016). pp. 785–794. [Google Scholar] 34. [Google Scholar] 34. [谷歌学术搜索]
35.Ke G, Meng Q, Finley T, Wang T, Chen W, Ma W, et al. Lightgbm: a highly efficient gradient boosting decision tree. Adv Neural Inf Process Syst. (2017):3146–54. doi: 10.5555/3294996.3295074 [DOI] [Google Scholar] 35. [DOI] [Google Scholar] 35. [DOI] [谷歌学术搜索]
36.Garrett E, Laky F, Meier K. User protocol for evaluation of qualitative test performance; approved guideline 2nd edn. 28 No. 3. CLSI/NCCLS document EP12-A2. Wayne, PA: Clinical and Laboratory Standards Institute; (2008) [Google Scholar]
36.加里特 E, Laky F, Meier K. 用户协议用于评估定性测试性能；批准的指南第 2 版。28 期第 3 号。CLSI/NCCLS 文件 EP12-A2。宾夕法尼亚州韦恩：临床与实验室标准研究所；（2008 年） [Google Scholar] 36. [Google Scholar] 36. [谷歌学术搜索]
37.Lundberg SM, Lee SI. A unified approach to interpreting model predictions. 31st Conference on Neural Information Processing Systems (NIPS; 2017), Long Beach, CA, USA. p. 30. [Google Scholar]
37.伦德伯格 SM，李 SI。对模型预测的统一解释方法。第 31 届神经信息处理系统会议（NIPS；2017 年），美国长滩，美国。第 30 页。[谷歌学者] 37. [Google Scholar] 37. [谷歌学术搜索]
38.Lipton ZC, Elkan C, Narayanaswamy B. Thresholding classifiers to maximize F1 score. arXiv preprint arXiv:1402.1892 (2014). [DOI] [PMC free article] [PubMed] [Google Scholar]38. [DOI] [PMC 免费文章] [PubMed] [Google 学术搜索]
39.Micro, Macro & Weighted Averages of F1 Score, Clearly Explained, Available at: https://towardsdatascience.com/micro-macro-weighted-averages-of-f1-score-clearly-explained-b603420b292f#989c, (accessed July 23 2022).
39. 微调、宏调及 F1 分数的加权平均，清晰解释，可访问：https://towardsdatascience.com/micro-macro-weighted-averages-of-f1-score-clearly-explained-b603420b292f#989c，访问日期：2022 年 7 月 23 日。 39.

Associated DataAssociated Data 关联数据

This section collects any data citations, data availability statements, or supplementary materials included in this article.本节收集本文中包含的任何数据引用、数据可用性声明或补充材料。

Data Availability StatementData Availability Statement 数据可用性声明

The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.支持本文结论的原始数据将由作者提供，不得无故保留。

[R1] 1.Ahmad M, Hollender L, Anderson Q, Kartha K, Ohrbach R, Truelove E, et al. Research diagnostic criteria for temporomandibular disorders (RDC/TMD): development of image analysis criteria and examiner reliability for image analysis. Oral Surg Oral Med Oral Pathol Oral Radiol Endod. (2009) 107 (6):844–60. doi: 10.1016/j.tripleo.2009.02.023 [DOI] [PMC free article] [PubMed] [Google Scholar]
1. 阿哈迈德 M, 洛兰德 L, 安德森 Q, 卡尔塔 K, 奥尔巴赫 R, 特鲁洛夫 E, 等。颞下颌关节障碍（TMD）的研究诊断标准（RDC/TMD）：开发图像分析标准和图像分析的检查员可靠性。口腔外科，口腔医学，口腔病理学，口腔放射学，牙周病学。2009 年，107 卷（6）：844–60。doi: 10.1016/j.tripleo.2009.02.023 [DOI] [PMC 免费文章] [PubMed] [谷歌学者]1. [DOI] [PMC 免费文章] [PubMed] [Google Scholar]

[R2] 2.Schiffman E, Ohrbach R, Truelove E, Look J, Anderson G, Goulet JP, et al. Diagnostic criteria for temporomandibular disorders (DC/TMD) for clinical and research applications: recommendations of the international RDC/TMD consortium network* and orofacial pain special interest group†. J Oral Facial Pain Headache. (2014) 28(1):6–27. doi: 10.11607/jop.1151 [DOI] [PMC free article] [PubMed] [Google Scholar]
2. Schiffman E, Ohrbach R, Truelove E, Look J, Anderson G, Goulet JP, 等. 临床上和研究应用的颞下颌关节障碍诊断标准（DC/TMD）：国际 RDC/TMD 联盟网络*和口腔面部疼痛特别兴趣小组†的建议. J Oral Facial Pain Headache. (2014) 28(1):6–27. doi: 10.11607/jop.1151 [ DOI] [ PMC 免费文章] [ PubMed] [ Google Scholar]2. [DOI] [PMC 免费文章] [PubMed] [Google Scholar]

[R3] 3.Wang X, Zhang J, Gan Y, Zhou Y. Current understanding of pathogenesis and treatment of TMJ osteoarthritis. J Dent Res. (2015) 94(5):666–73. doi: 10.1177/0022034515574770 [DOI] [PubMed] [Google Scholar]
3. 王 X，张 J，甘 Y，周 Y. 对颞下颌关节骨关节炎病理解剖和治疗的当前理解. J Dent Res. (2015) 94(5):666–73. doi: 10.1177/0022034515574770 [ DOI] [ PubMed] [ Google Scholar] 3. [DOI] [PubMed] [Google Scholar]3. [DOI] [PubMed] [Google 学术搜索]

[R4] 4.Embree M, Ono M, Kilts T, Walker D, Langguth J, Mao J, et al. Role of subchondral bone during early-stage experimental TMJ osteoarthritis. J Dent Res. (2011) 90(11):1331–8. doi: 10.1177/0022034511421930 [DOI] [PMC free article] [PubMed] [Google Scholar]4. [DOI] [PMC 免费文章] [PubMed] [Google 学术搜索]

[R5] 5.Jiao K, Niu L, Wang M, Dai J, Yu S, Liu X, Wang J. Subchondral bone loss following orthodontically induced cartilage degradation in the mandibular condyles of rats. Bone. (2011) 48(2):362–71. doi: 10.1016/j.bone.2010.09.010 [DOI] [PubMed] [Google Scholar] 5. [DOI] [PubMed] [Google Scholar]5. [DOI] [PubMed] [谷歌学术搜索]

[R6] 6.de Souza R, Lovato da Silva C, Nasser M, Fedorowicz Z, Al-Muharraqi M. Interventions for the management of temporomandibular joint osteoarthritis. Cochrane Database Syst Rev. (2012) 2012(4):CD007261. doi: 10.1002/14651858.cd007261.pub2 [DOI] [PMC free article] [PubMed] [Google Scholar]6. [DOI] [PMC 免费文章] [PubMed] [Google 学术搜索]

[R7] 7.Berenbaum F. Osteoarthritis as an inflammatory disease (osteoarthritis is not osteoarthrosis!). Osteoarthritis Cartilage. (2013) 21(1):16–21. doi: 10.1016/j.joca.2012.11.012 [DOI] [PubMed] [Google Scholar] 7. [DOI] [PubMed] [Google Scholar]7. [DOI] [PubMed] [谷歌学术搜索]

[R8] 8.Rousseau JC, Delmas PD. Biological markers in osteoarthritis. Nat Clin Pract Rheumatol. (2007) 3(6):346–56. doi: 10.1038/ncprheum0508 [DOI] [PubMed] [Google Scholar] 8. [DOI] [PubMed] [Google Scholar]8. [DOI] [PubMed] [谷歌学术搜索]

[R9] 9.Su N, Liu Y, Yang X, Luo Z, Shi Z. Correlation between bony changes measured with cone beam computed tomography and clinical dysfunction index in patients with temporomandibular joint osteoarthritis. J Craniomaxillofac Surg. (2014) 42(7):1402–7. doi: 10.1016/j.jcms.2014.04.001 [DOI] [PubMed] [Google Scholar] 9. [DOI] [PubMed] [Google Scholar]9. [DOI] [PubMed] [谷歌学术搜索]

[R10] 10.Bianchi J, de Oliveira Ruellas A, Gonçalves J, Paniagua B, Prieto J, Styner M, et al. Osteoarthritis of the temporomandibular joint can be diagnosed earlier using biomarkers and machine learning. Sci Rep. (2020) 10(1):8012. doi: 10.1038/s41598-020-64942-0 [DOI] [PMC free article] [PubMed] [Google Scholar]10. [DOI] [PMC 免费文章] [PubMed] [Google 学术搜索]

[R11] 11.Alomar X, Medrano J, Cabratosa J, Clavero JA, Lorente M, Serra I, et al. Anatomy of the temporomandibular joint. Semin Ultrasound CT MR. (2007) 28(3):170–83. doi: 10.1053/j.sult.2007.02.002 [DOI] [PubMed] [Google Scholar]
11. 阿洛马尔 X, Medrano J, Cabratosa J, Clavero JA, Lorente M, Serra I, 等. 下颌关节解剖学. 《超声学、CT 和 MRI 研讨会》. (2007) 28(3):170–83. doi: 10.1053/j.sult.2007.02.002 [DOI] [PubMed] [谷歌学者] 11. [DOI] [PubMed] [Google Scholar]11. [DOI] [PubMed] [谷歌学术搜索]

[R12] 12.Li G, Yin J, Gao J, Cheng T, Pavlos N, Zhang C, Zheng M. Subchondral bone in osteoarthritis: insight into risk factors and microstructural changes. Arthritis Res Ther. (2013) 15(6):223. doi: 10.1186/ar4405 [DOI] [PMC free article] [PubMed] [Google Scholar]
12. 李 G, 阴 J, 高 J, 成 T, Pavlos N, 张 C, 郑 M. 骨关节炎中的骨下骨：风险因素和微结构变化的见解. 关节炎研究与治疗. (2013) 15(6):223. doi: 10.1186/ar4405 [ DOI] [ PMC 免费文章] [ PubMed] [ Google Scholar]12. [DOI] [PMC 免费文章] [PubMed] [Google Scholar]

[R13] 13.Brandt K. Yet more evidence that osteoarthritis is not a cartilage disease. Ann Rheum Dis. (2006) 65:1261–4. doi: 10.1136/ard.2006.058347 [DOI] [PMC free article] [PubMed] [Google Scholar]13. [DOI] [PMC 免费文章] [PubMed] [Google Scholar]

[R14] 14.Tanaka E, Detamore MS, Mercuri LG. Degenerative disorders of the temporomandibular joint: etiology, diagnosis, and treatment. J Dent Res. (2008) 87(4):296–307. doi: 10.1177/154405910808700406 [DOI] [PubMed] [Google Scholar] 14. [DOI] [PubMed] [Google Scholar]14. [DOI] [PubMed] [谷歌学术搜索]

[R15] 15.Gillies RJ, Kinahan PE, Hricak H. Radiomics: images are more than pictures, they are data. Radiology. (2016) 278(2):563–77. doi: 10.1148/radiol.2015151169 [DOI] [PMC free article] [PubMed] [Google Scholar]15. [DOI] [PMC 免费文章] [PubMed] [Google Scholar]

[R16] 16.Brahim A, Jennane R, Riad R, Janvier T, Khedher L, Toumi H, Lespessailles E. A decision support tool for early detection of knee OsteoArthritis using x-ray imaging and machine learning: data from the OsteoArthritis initiative. Comput Med Imaging Graph. (2019) 73:11–8. doi: 10.1016/j.compmedimag.2019.01.007 [DOI] [PubMed] [Google Scholar] 16. [DOI] [PubMed] [Google Scholar]16. [DOI] [PubMed] [谷歌学术搜索]

[R17] 17.Paniagua B, Ruellas AC, Benavides E, Marron S, Woldford L, Cevidanes L. Validation of CBCT for the computation of textural biomarkers. Proc SPIE Int Soc Opt Eng. (2015) 9417:94171B. doi: 10.1117/12.2081859 [DOI] [PMC free article] [PubMed] [Google Scholar]17. [DOI] [PMC 免费文章] [PubMed] [Google Scholar]

[R18] 18.Bianchi J, Gonçalves J, Ruellas A, Vimort J, Yatabe M, Paniagua B, et al. Software comparison to analyze bone radiomics from high resolution CBCT scans of mandibular condyles. Dentomaxillofac Radiol. (2019) 48(6):20190049. doi: 10.1259/dmfr.20190049 [DOI] [PMC free article] [PubMed] [Google Scholar]18. [DOI] [PMC 免费文章] [PubMed] [Google 学术搜索]

[R19] 19.Bianchi J, Gonçalves JR, de Oliveira Ruellas AC, Ashman LM, Vimort JB, Yatabe M, et al. Quantitative bone imaging biomarkers to diagnose temporomandibular joint osteoarthritis. Int J Oral Maxillofac Surg. (2021) 50 (2):227–35. doi: 10.1016/j.ijom.2020.04.018. [DOI] [PMC free article] [PubMed] [Google Scholar]19. [DOI] [PMC 免费文章] [PubMed] [Google 学术搜索]

[R20] 20.Wang Y, Liu C, Rohr J, Liu H, He F, Yu J, et al. Tissue interaction is required for glenoid fossa development during temporomandibular joint formation. Dev Dyn. (2011) 240:2466–73. doi: 10.1002/dvdy.22748 [DOI] [PMC free article] [PubMed] [Google Scholar]
20. 王 Y, 刘 C, Rohr J, 刘 H, 何 F, 余 J, 等. 发育动力学. (2011) 240:2466–73. doi: 10.1002/dvdy.22748 [DOI] [PMC 免费文章] [PubMed] [谷歌学者]20. [DOI] [PMC 免费文章] [PubMed] [Google 学术搜索]

[R21] 21.Soydan D, Doğan S, Canger E, Coşgunarslan A, Akgün I, Kış H. Effect of internal derangements and degenerative bone changes on the minimum thickness of the roof of the glenoid fossa in temporomandibular joint. Oral Radiol. (2020) 36(1):25–31. doi: 10.1007/s11282-019-00373-6 [DOI] [PubMed] [Google Scholar]
21. 草丹 D, 動 S, 江 E, 玄 A, 雲 I, 冬 H. 内部紊乱和退行性骨变化对颞下颌关节的骨窝顶最小厚度的影响. 口腔放射学. (2020) 36(1):25–31. doi: 10.1007/s11282-019-00373-6 [ DOI] [ PubMed] [ Google 学者] 21. [DOI] [PubMed] [Google Scholar]21. [DOI] [PubMed] [谷歌学术搜索]

[R22] 22.Massilla Mani F, Sivasubramanian SS. A study of temporomandibular joint osteoarthritis using computed tomographic imaging. Biomed J. (2016) 39 (3):201–6. doi: 10.1016/j.bj.2016.06.003. [DOI] [PMC free article] [PubMed] [Google Scholar]22. [DOI] [PMC 免费文章] [PubMed] [Google 学术搜索]

[R23] 23.Rando C, Waldron T. TMJ Osteoarthritis: a new approach to diagnosis. Am J Phys Anthropol. (2012) 148(1):45–53. doi: 10.1002/ajpa.22039 [DOI] [PubMed] [Google Scholar] 23. [DOI] [PubMed] [Google Scholar]23. [DOI] [PubMed] [谷歌学术搜索]

[R24] 24.Cevidanes LH, Walker D, Schilling J, Sugai J, Giannobile W, Paniagua B, et al. 3D Osteoarthritic changes in TMJ condylar morphology correlates with specific systemic and local biomarkers of disease. Osteoarthritis Cartilage. (2014) 22(10):1657–67. doi: 10.1016/j.joca.2014.06.014 [DOI] [PMC free article] [PubMed] [Google Scholar]24. [DOI] [PMC 免费文章] [PubMed] [Google Scholar]

[R25] 25.Kofler B, Jenetten L, Runge A, Degenhart G, Fischer N, Hörmann R, et al. ALADA dose optimization in the computed tomography of the temporal bone: the diagnostic potential of different low-dose CT protocols. Diagnostics (Basel). (2021) 11(10):1894. doi: 10.3390/diagnostics11101894 [DOI] [PMC free article] [PubMed] [Google Scholar]25. [DOI] [PMC 免费文章] [PubMed] [Google 学术搜索]

[R26] 26.3D Slicer. https://www.slicer.org, accessed. July 23 2022.
26.3D Slicer. https://www.slicer.org, 访问日期. 2022 年 7 月 23 日. 26.

[R27] 27.ITK-SNAP. http://www.itksnap.org, accessed. July 23 2022.
27.ITK-SNAP. http://www.itksnap.org, 2022 年 7 月 23 日访问. 27.

[R28] 28.Shoukri B, Prieto J, Ruellas A, Yatabe M, Sugai J, Styner M, et al. Minimally invasive approach for diagnosing TMJ osteoarthritis. J Dent Res. (2019) 98 (10):1103–11. doi: 10.1177/0022034519865187 [DOI] [PMC free article] [PubMed] [Google Scholar]28. [DOI] [PMC 免费文章] [PubMed] [Google 学术搜索]

[R29] 29.Nieminen M, Casula V, Nevalainen M, Saarakkala S. Osteoarthritis year in review 2018: imaging. Osteoarthritis Cartilage. (2019) 27(3):401–11. doi: 10.1016/j.joca.2018.12.009 [DOI] [PubMed] [Google Scholar] 29. [DOI] [PubMed] [Google Scholar]29. [DOI] [PubMed] [谷歌学术搜索]

[R30] 30.Ebrahim F, Ruellas A, Paniagua B, Benavides E, Jepsen K, Wolford L, et al. Accuracy of biomarkers obtained from cone beam computed tomography in assessing the internal trabecular structure of the mandibular condyle. Oral Surg Oral Med Oral Pathol Oral Radiol, 2017. 124(6): p. 588–99. doi: 10.1016/j.oooo.2017.08.013 [DOI] [PMC free article] [PubMed] [Google Scholar]30. [DOI] [PMC 免费文章] [PubMed] [Google Scholar]

[R31] 31.Le C, Deleat-Besson R, Turkestani NA, Cevidanes L, Bianchi J, Zhang W, et al. TMJOAI: An artificial web-based intelligence tool for early diagnosis of the temporomandibular joint osteoarthritis. Clin Image Based Proced Distrib Collab Learn Artif Intell Combat COVID 19 Secur Priv Preserv Mach Learn, 2021 Sep-Oct;12969: (2021). pp. 78–87. doi: 10.1007/978-3-030-90874-4_8. [DOI] [PMC free article] [PubMed] [Google Scholar]31. [DOI] [PMC 免费文章] [PubMed] [Google Scholar]

[R32] 32.Zhang W, Bianchi J, Turkestani NA, Le C, Deleat-Besson R, Ruellas A, et al. Temporomandibular joint osteoarthritis diagnosis using privileged learning of protein markers. Annu Int Conf IEEE Eng Med Biol Soc. (2021) 2021:1810–3. doi: 10.1109/EMBC46164.2021.9629990 [DOI] [PMC free article] [PubMed] [Google Scholar]32. [DOI] [PMC 免费文章] [PubMed] [Google 学术搜索]

[R33] 33.Brosset S, Dumont M, Cevidanes L, Soroushmehr R, Bianchi J, Gurgel M, et al. Web infrastructure for data management, storage and computation. Proc SPIE Int Soc Opt Eng. (2021) 11600:116001N. doi: 10.1117/12.2582283 [DOI] [PMC free article] [PubMed] [Google Scholar]33. [DOI] [PMC 免费文章] [PubMed] [Google Scholar]

[R34] 34.Chen T, Guestrin C. Xgboost: A scalable tree boosting system. In Proceedings of the 22nd Acm Sigkdd International Conference on Knowledge Discovery and Data Mining (ACM, 2016). pp. 785–794. [Google Scholar] 34. [Google Scholar] 34. [谷歌学术搜索]

[R35] 35.Ke G, Meng Q, Finley T, Wang T, Chen W, Ma W, et al. Lightgbm: a highly efficient gradient boosting decision tree. Adv Neural Inf Process Syst. (2017):3146–54. doi: 10.5555/3294996.3295074 [DOI] [Google Scholar] 35. [DOI] [Google Scholar] 35. [DOI] [谷歌学术搜索]

[R36] 36.Garrett E, Laky F, Meier K. User protocol for evaluation of qualitative test performance; approved guideline 2nd edn. 28 No. 3. CLSI/NCCLS document EP12-A2. Wayne, PA: Clinical and Laboratory Standards Institute; (2008) [Google Scholar]
36.加里特 E, Laky F, Meier K. 用户协议用于评估定性测试性能；批准的指南第 2 版。28 期第 3 号。CLSI/NCCLS 文件 EP12-A2。宾夕法尼亚州韦恩：临床与实验室标准研究所；（2008 年） [Google Scholar] 36. [Google Scholar] 36. [谷歌学术搜索]

[R37] 37.Lundberg SM, Lee SI. A unified approach to interpreting model predictions. 31st Conference on Neural Information Processing Systems (NIPS; 2017), Long Beach, CA, USA. p. 30. [Google Scholar]
37.伦德伯格 SM，李 SI。对模型预测的统一解释方法。第 31 届神经信息处理系统会议（NIPS；2017 年），美国长滩，美国。第 30 页。[谷歌学者] 37. [Google Scholar] 37. [谷歌学术搜索]

[R38] 38.Lipton ZC, Elkan C, Narayanaswamy B. Thresholding classifiers to maximize F1 score. arXiv preprint arXiv:1402.1892 (2014). [DOI] [PMC free article] [PubMed] [Google Scholar]38. [DOI] [PMC 免费文章] [PubMed] [Google 学术搜索]

[R39] 39.Micro, Macro & Weighted Averages of F1 Score, Clearly Explained, Available at: https://towardsdatascience.com/micro-macro-weighted-averages-of-f1-score-clearly-explained-b603420b292f#989c, (accessed July 23 2022).
39. 微调、宏调及 F1 分数的加权平均，清晰解释，可访问：https://towardsdatascience.com/micro-macro-weighted-averages-of-f1-score-clearly-explained-b603420b292f#989c，访问日期：2022 年 7 月 23 日。 39.

PERMALINK 永久链接

Tamara Mackie

Najla Al Turkestani

Jonas Bianchi

Tengfei Li

Antonio Ruellas

Marcela Gurgel

Erika Benavides

Fabiana Soki

Lucia Cevidanes

Abstract 摘要Abstract Abstract

Introduction 简介Introduction 介绍

Materials and methodsMaterials and methods 材料和方法

Study design and participantsStudy design and participants 研究设计和受试者

Clinical signs and symptomsClinical signs and symptoms 临床体征和症状