首页 > 最新文献

Osteoarthritis imaging最新文献

英文 中文
AUTOMATING IMAGING BIOMARKER ANALYSIS FOR KNEE OSTEOARTHRITIS USING AN OPEN-SOURCE MRI-BASED DEEP LEARNING PIPELINE 使用开源的基于核磁共振的深度学习管道对膝关节骨关节炎进行自动成像生物标志物分析
Osteoarthritis imaging
Pub Date : 2025-01-01 DOI: 10.1016/j.ostima.2025.100288
A. Goyal , F. Belibi , V. Sahani , R. Pedersen , Y. Vainberg , A. Williams , C. Chu , B. Haddock , G. Gold , A.S. Chaudhari , F. Kogan , A.A. Gatti

INTRODUCTION

Quantitative MRI and [¹⁸F]NaF PET enable assessment of cartilage composition, bone shape, and subchondral bone metabolism in knee OA. Current workflows rely on manual segmentation that is time-consuming and subject to inter- and intra-reader variability. Furthermore, computing quantitative metrics requires considerable time and expertise. An open-source, automated, deep learning (DL) pipeline with standardized biomarker extraction has the potential to enhance reproducibility and make large-scale analysis accessible to clinical research communities, including non-technical users.

OBJECTIVE

Develop and validate an automated DL-based pipeline for comprehensive MRI-based segmentation and quantitative analysis of multiple knee tissues from multi-modal MR and PET images.

METHODS

We developed and open-sourced a comprehensive segmentation and analysis pipeline. A 2D U-Net was trained to segment 9 tissues using a dataset of 347 DESS and qDESS images: 3 bones (femur, tibia, patella), 4 cartilage regions (femoral, medial and lateral tibial, patellar), and 2 menisci (medial and lateral). Subchondral bone masks and femoral cartilage subregions were fitted automatically. Quantitative imaging biomarkers were computed as follows: cartilage T2 was computed analytically from qDESS scans; cartilage thickness was computed as the 3D Euclidean thickness of cartilage overlying the bone surface; meniscal volume was calculated as the product of voxel count and voxel volume; OA bone shape (BScore) was derived using a neural shape model; PET-derived subchondral bone metabolism was computed as regional SUVmean/max, and kinetic modeling via Hawkin’s method was used to extract KiNLR (bone mineralization rate) and K1 (perfusion to subchondral bone). To evaluate the pipeline, 20 unilateral qDESS and [¹⁸F]NaF PET knee scans (10 symptomatic OA, 10 controls) were analyzed by the automated pipeline, and two manual annotators. Manual and automated segmentations were compared using the Dice Similarity Coefficient (DSC) and average symmetric surface distance (ASSD). Biomarkers were compared using ICC and normalized mean RMSE (NRMSE).

RESULTS

All automated segmentations had good to excellent overlap measured using DSC (bone: 0.95-0.98; cartilage: 0.84-0.91; menisci: 0.85-0.89) and small surface errors (bone: 0.13-0.32 mm; cartilage: 0.11-0.21 mm; menisci: 0.17-0.30 mm). Notably, automated segmentations had better DSC and ASSD than the inter-rater comparison (Fig. 2). With the exception of cartilage thickness and patellar cartilage whole T2 values, all quantitative metrics showed excellent agreement with ICC >0.96 and NRMSE <0.1, comparable to inter-rater comparison. Bone metrics (BScore, SUV, PET kinetics) had ICC >0.96. Cartilage metrics had more variability, with the best reproducibility for whole cartilage T2 (ICC 0.89-0.98, NRMSE 0.01-0.04), then superficia
定量MRI和[¹⁸F]NaF PET可以评估膝关节OA患者的软骨组成、骨形状和软骨下骨代谢。当前的工作流程依赖于手动分割，这是耗时的，并受到阅读器之间和内部变化的影响。此外，计算定量度量需要大量的时间和专业知识。具有标准化生物标志物提取的开源、自动化、深度学习（DL）管道有可能提高可重复性，并为临床研究社区（包括非技术用户）提供大规模分析。目的：开发并验证一种自动化的基于dl的管道，用于从多模态MR和PET图像中对多个膝关节组织进行全面的基于mri的分割和定量分析。方法我们开发并开源了一个全面的细分和分析流程。使用347张DESS和qDESS图像数据集训练2D U-Net来分割9个组织：3块骨头（股骨，胫骨，髌骨），4个软骨区域（股骨，胫骨内侧和外侧，髌骨）和2个半月板（内侧和外侧）。软骨下骨面罩和股骨软骨亚区自动拟合。定量成像生物标志物计算如下：软骨T2通过qDESS扫描分析计算；软骨厚度计算为覆盖骨表面的软骨的三维欧几里得厚度；半月板体积计算为体素数与体素体积的乘积；采用神经形态模型推导OA骨形态（BScore）；pet衍生的软骨下骨代谢计算为区域SUVmean/max，通过Hawkin方法进行动力学建模，提取KiNLR（骨矿化率）和K1（软骨下骨灌注）。为了评估管道，我们使用自动管道和2个手动注释器分析了20个单侧qDESS和[¹⁸F]NaF PET膝关节扫描（10个症状性OA， 10个对照组）。使用Dice Similarity Coefficient （DSC）和平均对称表面距离（ASSD）对手动分割和自动分割进行比较。使用ICC和标准化平均RMSE （NRMSE）比较生物标志物。结果DSC测量所有自动分割的重叠度均为良好至优异(骨：0.95 ~ 0.98；软骨:0.84 - -0.91;半月板：0.85-0.89)，表面误差小(骨：0.13-0.32 mm；软骨：0.11-0.21 mm；半月板：0.17-0.30 mm)。值得注意的是，自动分割的DSC和ASSD优于内部比较（图2）。除了软骨厚度和髌骨软骨整体T2值外，所有定量指标均与ICC >；0.96和NRMSE <；0.1非常吻合，与间比较相当。骨骼指标（BScore， SUV， PET动力学）的ICC为0.96。软骨指标的可变性更大，T2全软骨的重现性最好（ICC 0.89-0.98, NRMSE 0.01-0.04），其次是T2浅层（ICC 0.93-0.99, NRMSE 0.01-0.05），最后是T2深层（ICC 0.7-0.97, NRMSE 0.01-0.06）。软骨厚度的再现性最差，但仍可与其他测量方法相媲美。半月板体积也显示出高度的一致性(ICC 0.93-0.97；NRMSE 0.05 - -0.10)。总的来说，我们发现大多数来自自动分割的指标与那些来自手动分割的指标是相当的。我们的开源、人工智能驱动的管道提供快速、准确的多模态膝关节MRI和PET数据分割和定量分析。接下来的步骤包括支持其他MR序列、多位点验证和3D切片器集成以促进翻译。该资源为OA研究和临床试验中可重复和可扩展的成像生物标志物分析提供了基础。
{"title":"AUTOMATING IMAGING BIOMARKER ANALYSIS FOR KNEE OSTEOARTHRITIS USING AN OPEN-SOURCE MRI-BASED DEEP LEARNING PIPELINE","authors":"A. Goyal ,&nbsp;F. Belibi ,&nbsp;V. Sahani ,&nbsp;R. Pedersen ,&nbsp;Y. Vainberg ,&nbsp;A. Williams ,&nbsp;C. Chu ,&nbsp;B. Haddock ,&nbsp;G. Gold ,&nbsp;A.S. Chaudhari ,&nbsp;F. Kogan ,&nbsp;A.A. Gatti","doi":"10.1016/j.ostima.2025.100288","DOIUrl":"10.1016/j.ostima.2025.100288","url":null,"abstract":"<div><h3>INTRODUCTION</h3><div>Quantitative MRI and [¹⁸F]NaF PET enable assessment of cartilage composition, bone shape, and subchondral bone metabolism in knee OA. Current workflows rely on manual segmentation that is time-consuming and subject to inter- and intra-reader variability. Furthermore, computing quantitative metrics requires considerable time and expertise. An open-source, automated, deep learning (DL) pipeline with standardized biomarker extraction has the potential to enhance reproducibility and make large-scale analysis accessible to clinical research communities, including non-technical users.</div></div><div><h3>OBJECTIVE</h3><div>Develop and validate an automated DL-based pipeline for comprehensive MRI-based segmentation and quantitative analysis of multiple knee tissues from multi-modal MR and PET images.</div></div><div><h3>METHODS</h3><div>We developed and open-sourced a comprehensive segmentation and analysis pipeline. A 2D U-Net was trained to segment 9 tissues using a dataset of 347 DESS and qDESS images: 3 bones (femur, tibia, patella), 4 cartilage regions (femoral, medial and lateral tibial, patellar), and 2 menisci (medial and lateral). Subchondral bone masks and femoral cartilage subregions were fitted automatically. Quantitative imaging biomarkers were computed as follows: cartilage T2 was computed analytically from qDESS scans; cartilage thickness was computed as the 3D Euclidean thickness of cartilage overlying the bone surface; meniscal volume was calculated as the product of voxel count and voxel volume; OA bone shape (BScore) was derived using a neural shape model; PET-derived subchondral bone metabolism was computed as regional SUVmean/max, and kinetic modeling via Hawkin’s method was used to extract KiNLR (bone mineralization rate) and K1 (perfusion to subchondral bone). To evaluate the pipeline, 20 unilateral qDESS and [¹⁸F]NaF PET knee scans (10 symptomatic OA, 10 controls) were analyzed by the automated pipeline, and two manual annotators. Manual and automated segmentations were compared using the Dice Similarity Coefficient (DSC) and average symmetric surface distance (ASSD). Biomarkers were compared using ICC and normalized mean RMSE (NRMSE).</div></div><div><h3>RESULTS</h3><div>All automated segmentations had good to excellent overlap measured using DSC (bone: 0.95-0.98; cartilage: 0.84-0.91; menisci: 0.85-0.89) and small surface errors (bone: 0.13-0.32 mm; cartilage: 0.11-0.21 mm; menisci: 0.17-0.30 mm). Notably, automated segmentations had better DSC and ASSD than the inter-rater comparison (Fig. 2). With the exception of cartilage thickness and patellar cartilage whole T2 values, all quantitative metrics showed excellent agreement with ICC &gt;0.96 and NRMSE &lt;0.1, comparable to inter-rater comparison. Bone metrics (BScore, SUV, PET kinetics) had ICC &gt;0.96. Cartilage metrics had more variability, with the best reproducibility for whole cartilage T2 (ICC 0.89-0.98, NRMSE 0.01-0.04), then superficia","PeriodicalId":74378,"journal":{"name":"Osteoarthritis imaging","volume":"5 ","pages":"Article 100288"},"PeriodicalIF":0.0,"publicationDate":"2025-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144524003","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
引用次数: 0
DO RATES OF FEMOROTIBIAL CARTILAGE LOSS IN KELLGREN-LAWRENCE 2 AND 3 KNEES DIFFER BETWEEN THOSE WITH MILD-MODERATE VS. SEVERE PATELLOFEMORAL STRUCTURAL DAMAGE? 轻度-中度髌骨-股骨结构损伤与重度髌骨-股骨结构损伤相比，kellgren-lawrence 2型和3型膝关节的股胫软骨丢失率不同吗？
Osteoarthritis imaging
Pub Date : 2025-01-01 DOI: 10.1016/j.ostima.2025.100311
F.W. Roemer , M.P. Jansen , S. Maschek , S. Mastbergen , A. Wisser , H.H. Weinans , F.J. Blanco , F. Berenbaum , M. Kloppenburg , I.K. Haugen , D.J. Hunter , A. Guermazi , W. Wirth

INTRODUCTION

Knees with radiographic disease severity of Kellgren-Lawrence (KL) 2 and 3 are commonly included in disease-modifying (DMOAD) clinical trials of knee osteoarthritis (OA). In an eligibility context, semi-quantitative (sq) MRI assessment has been used to define structural disease severity, rule out diagnoses of exclusion, and possibly define a structural phenotype. The KL system focuses on the femorotibial joint (FTJ) only, with MRI stratification being commonly limited to the FTJ. It is unclear whether sq MRI of the patellofemoral joint (PFJ) should be included for eligibility assessment.

OBJECTIVE

The aim was to assess whether rates of quantitative femorotibial (FT) cartilage loss are increased for knees with semiquantitatively (sq)-defined severe patellofemoral (PF) cartilage damage and/or large bone marrow lesions (BMLs) vs. those without over a period of 24 months.

METHODS

626 knees with Kellgren-Lawrence 2 and 3 from the FNIH and IMI-APPROACH studies were included. MRI assessment was performed using the MRI Osteoarthritis Knee Score (MOAKS) instrument. Medial FT quantitative cartilage thickness loss was derived from baseline and 24-month manual segmentations and was compared between knees with severe vs. mild-moderate PF cartilage damage and between knees with vs. without large PF BMLs. Between-group comparisons were performed using analysis of variance (ANOVA) and were stratified by baseline medial FT cartilage damage severity (defined as mild, moderate, or severe).

RESULTS

410 (65%) knees were categorized as mild, 92 (15%) as moderate, and 124 (20%) as severe medial FT cartilage damage. For almost all categories of FT cartilage damage, the difference in quantitative medial FT cartilage loss was not statistically significant (Table 1). Only for the category of knees with moderate medial FT cartilage damage, statistically higher rates of quantitative medial FT cartilage loss were observed for those with large PF BMLs compared to those without (-0.245 ± 0.304 mm vs. -0.134 ± 0.218 mm) (Table 2).

CONCLUSION

For the large majority of sq-defined FT cartilage damage categories, no statistically significant differences in FT rates of quantitative cartilage loss were detected. Screening for PF cartilage damage and BMLs does not appear to be required in a disease-modifying OA drug trial.
患有Kellgren-Lawrence (KL) 2级和3级放射学疾病严重程度的膝关节通常包括在膝关节骨关节炎（OA）的疾病改善（DMOAD）临床试验中。在合格的背景下，半定量（sq） MRI评估已用于确定结构性疾病的严重程度，排除排除性诊断，并可能确定结构性表型。KL系统仅聚焦于股胫关节（FTJ）， MRI分层通常局限于FTJ。目前尚不清楚是否应将髌股关节（PFJ）的sq MRI纳入资格评估。目的：评估在24个月的时间里，有半定量（sq）定义的严重髌骨股骨（PF）软骨损伤和/或大骨髓病变（BMLs）的膝关节，与没有严重髌骨股骨（PF）软骨损伤的膝关节相比，定量股胫（FT）软骨损失的发生率是否增加。方法纳入626例来自FNIH和IMI-APPROACH研究的kelgren - lawrence 2和3膝关节。采用MRI骨关节炎膝关节评分（MOAKS）仪进行MRI评估。内侧FT定量软骨厚度损失来自基线和24个月的手工分割，并比较严重和轻度PF软骨损伤的膝关节以及有和没有大PF软骨损伤的膝关节之间的差异。采用方差分析（ANOVA）进行组间比较，并按基线内侧FT软骨损伤严重程度（定义为轻度、中度或重度）分层。结果410例膝关节（65%）为轻度，92例（15%）为中度，124例（20%）为重度内侧FT软骨损伤。对于几乎所有类型的FT软骨损伤，内侧FT软骨定量损失的差异无统计学意义（表1）。仅对于中度内侧FT软骨损伤的膝关节类别，统计学上观察到，与没有中度内侧FT软骨损伤的膝关节相比，具有较大PF bml的膝关节内侧FT软骨定量损失率更高（-0.245±0.304 mm vs -0.134±0.218 mm）（表2）。结论对于绝大多数sq定义的FT软骨损伤类别，在定量软骨损失的FT率上没有发现统计学上的显著差异。在改善疾病的OA药物试验中，似乎不需要筛选PF软骨损伤和bls。
{"title":"DO RATES OF FEMOROTIBIAL CARTILAGE LOSS IN KELLGREN-LAWRENCE 2 AND 3 KNEES DIFFER BETWEEN THOSE WITH MILD-MODERATE VS. SEVERE PATELLOFEMORAL STRUCTURAL DAMAGE?","authors":"F.W. Roemer ,&nbsp;M.P. Jansen ,&nbsp;S. Maschek ,&nbsp;S. Mastbergen ,&nbsp;A. Wisser ,&nbsp;H.H. Weinans ,&nbsp;F.J. Blanco ,&nbsp;F. Berenbaum ,&nbsp;M. Kloppenburg ,&nbsp;I.K. Haugen ,&nbsp;D.J. Hunter ,&nbsp;A. Guermazi ,&nbsp;W. Wirth","doi":"10.1016/j.ostima.2025.100311","DOIUrl":"10.1016/j.ostima.2025.100311","url":null,"abstract":"<div><h3>INTRODUCTION</h3><div>Knees with radiographic disease severity of Kellgren-Lawrence (KL) 2 and 3 are commonly included in disease-modifying (DMOAD) clinical trials of knee osteoarthritis (OA). In an eligibility context, semi-quantitative (sq) MRI assessment has been used to define structural disease severity, rule out diagnoses of exclusion, and possibly define a structural phenotype. The KL system focuses on the femorotibial joint (FTJ) only, with MRI stratification being commonly limited to the FTJ. It is unclear whether sq MRI of the patellofemoral joint (PFJ) should be included for eligibility assessment.</div></div><div><h3>OBJECTIVE</h3><div>The aim was to assess whether rates of quantitative femorotibial (FT) cartilage loss are increased for knees with semiquantitatively (sq)-defined severe patellofemoral (PF) cartilage damage and/or large bone marrow lesions (BMLs) vs. those without over a period of 24 months.</div></div><div><h3>METHODS</h3><div>626 knees with Kellgren-Lawrence 2 and 3 from the FNIH and IMI-APPROACH studies were included. MRI assessment was performed using the MRI Osteoarthritis Knee Score (MOAKS) instrument. Medial FT quantitative cartilage thickness loss was derived from baseline and 24-month manual segmentations and was compared between knees with severe vs. mild-moderate PF cartilage damage and between knees with vs. without large PF BMLs. Between-group comparisons were performed using analysis of variance (ANOVA) and were stratified by baseline medial FT cartilage damage severity (defined as mild, moderate, or severe).</div></div><div><h3>RESULTS</h3><div>410 (65%) knees were categorized as mild, 92 (15%) as moderate, and 124 (20%) as severe medial FT cartilage damage. For almost all categories of FT cartilage damage, the difference in quantitative medial FT cartilage loss was not statistically significant (<strong>Table 1</strong>). Only for the category of knees with moderate medial FT cartilage damage, statistically higher rates of quantitative medial FT cartilage loss were observed for those with large PF BMLs compared to those without (-0.245 ± 0.304 mm vs. -0.134 ± 0.218 mm) (<strong>Table 2</strong>).</div></div><div><h3>CONCLUSION</h3><div>For the large majority of sq-defined FT cartilage damage categories, no statistically significant differences in FT rates of quantitative cartilage loss were detected. Screening for PF cartilage damage and BMLs does not appear to be required in a disease-modifying OA drug trial.</div></div>","PeriodicalId":74378,"journal":{"name":"Osteoarthritis imaging","volume":"5 ","pages":"Article 100311"},"PeriodicalIF":0.0,"publicationDate":"2025-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144524025","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
引用次数: 0
AGREEMENT BETWEEN IN VIVO AND EX VIVO PHOTON-COUNTING CT MEASURES OF SUBCHONDRAL BONE FEATURES IN PATIENTS WITH KNEE OSTEOARTHRITIS 膝关节骨性关节炎患者软骨下骨特征的体内和体外光子计数ct测量的一致性
Osteoarthritis imaging
Pub Date : 2025-01-01 DOI: 10.1016/j.ostima.2025.100304
C.T. Nielsen , M. Boesen , M. Henriksen , J.U. Nybing , S.W. Bardenfleth , C.K. Rasmussen , M.W. Brejnebøl , A.S. Poulsen , S.M. Aljuboori , K.I. Bunyoz , S. Overgaard , A. Troelsen , H. Bliddal , H. Gudbergsen , F. Müller

INTRODUCTION

Bone changes are integral to the onset and progression of OA. Many aspects remain poorly understood due to the inability to assess bone architecture in vivo. Research has relied on ex vivo imaging, hindering evaluation of early-stage disease and longitudinal analysis. Conventional CT lacks the resolution to visualise subchondral bone microstructure. While ex vivo Photon Counting CT (PCCT) has demonstrated imaging comparable to μCT, its ability to capture bone microstructure in vivo in knee OA patients under clinical conditions remains unproven.

OBJECTIVE

The aim of this study was to compare in vivo and ex vivo PCCT of subchondral bone features in patients with knee OA.

METHODS

Pre-surgery in vivo and post-surgery ex vivo PCCT (Siemens Naeotom Alpha, Siemens Healthineers, Germany) of the tibial plateau from participants with severe knee OA referred to arthroplasty surgery from January 2022 through September 2023 were compared. Acquisition/reconstruction details: a tube current of 120 kV, a matrix size of 1024 × 1024, a slice thickness of 0.2 mm, and a FOV of 150 × 150 mm. 18 in vivo/ex vivo PCCT pairs were included. The ex vivo scans was registered to the in vivo scans. Linear regression and Bland-Altman plots were used to assess correlation and agreement between in vivo and ex vivo measures of bone volume fraction (BV/TV), trabecular thickness (Tb.Th.), and attenuation in healthy and sclerotic trabecular bone. Delineated areas of bone sclerosis were compared using the Dice coefficient and Hausdorff distance, Fig. 1.

RESULTS

Comparing in vivo and ex vivo scans strong correlations were found for BV/TV, R2=0.82 and attenuation in both healthy, R2=0.89, and sclerotic, R2=0.79, bone, while a moderate correlation was found for Tb.Th., R2=0.55. Bias for BV/TV and Tb.Th. was -4.1% and -0.598mm, respectively, and -41.4 HU and -81.1 HU for healthy and sclerotic bone, respectively. A proportional bias was observed for BV/TV and Tb.Th., Fig. 2. There was excellent agreement between the segmentations of sclerotic areas, Dice coefficient = 0.91 and Hausdorff distance = 0.11mm.

CONCLUSION

In patients with severe knee OA, BV/TV and attenuation can be obtained with high correlation and small bias between in vivo and ex vivo scans. Tb.Th. showed moderate correlation and larger bias. Subchondral bone sclerosis, a key OA feature, is well translated from ex vivo to in vivo PCCT. Longitudinal studies using in vivo PCCT are feasible, but caution may be advised when measuring Tb.Th.
骨改变是骨性关节炎发病和进展不可或缺的一部分。由于无法评估体内的骨结构，许多方面仍然知之甚少。研究依赖于离体成像，阻碍了早期疾病的评估和纵向分析。常规CT缺乏显示软骨下骨微观结构的分辨率。虽然离体光子计数CT （PCCT）已经证明了与μCT相当的成像能力，但其在临床条件下捕获膝关节OA患者体内骨骼微观结构的能力仍未得到证实。目的比较膝关节OA患者软骨下骨特征的体内和体外PCCT。方法比较2022年1月至2023年9月期间接受关节置换术的严重膝OA患者的术前体内和术后离体PCCT （Siemens Naeotom Alpha, Siemens Healthineers, Germany）。采集/重建细节：管电流为120 kV，矩阵尺寸为1024 × 1024，切片厚度为0.2 mm，视场为150 × 150 mm。包括18对体内/离体PCCT。离体扫描与体内扫描相匹配。采用线性回归和Bland-Altman图来评估体内和体外骨量分数（BV/TV）、骨小梁厚度（Tb.Th.）和健康和硬化骨小梁衰减之间的相关性和一致性。使用Dice系数和Hausdorff距离对所描绘的骨硬化区域进行比较，见图1。结果在体内和离体扫描比较，BV/TV与健康骨（R2=0.82）和硬化骨（R2=0.79）有较强的相关性，BV/TV与tb有中等相关性。R2 = 0.55。对BV/TV和th的偏见。健康骨和硬化骨分别为-41.4 HU和-81.1 HU，为-4.1%和-0.598mm。BV/TV和th呈比例偏倚。，图2。硬化区分割结果吻合良好，Dice系数 = 0.91，Hausdorff距离 = 0.11mm。结论在严重膝关节OA患者中，BV/TV和衰减在体内和离体扫描之间具有高相关性和小偏差。Tb.Th。相关性中等，偏倚较大。软骨下骨硬化是骨性关节炎的一个重要特征，它可以很好地从离体转化为体内PCCT。使用体内PCCT进行纵向研究是可行的，但在测量tth时应谨慎。
{"title":"AGREEMENT BETWEEN IN VIVO AND EX VIVO PHOTON-COUNTING CT MEASURES OF SUBCHONDRAL BONE FEATURES IN PATIENTS WITH KNEE OSTEOARTHRITIS","authors":"C.T. Nielsen ,&nbsp;M. Boesen ,&nbsp;M. Henriksen ,&nbsp;J.U. Nybing ,&nbsp;S.W. Bardenfleth ,&nbsp;C.K. Rasmussen ,&nbsp;M.W. Brejnebøl ,&nbsp;A.S. Poulsen ,&nbsp;S.M. Aljuboori ,&nbsp;K.I. Bunyoz ,&nbsp;S. Overgaard ,&nbsp;A. Troelsen ,&nbsp;H. Bliddal ,&nbsp;H. Gudbergsen ,&nbsp;F. Müller","doi":"10.1016/j.ostima.2025.100304","DOIUrl":"10.1016/j.ostima.2025.100304","url":null,"abstract":"<div><h3>INTRODUCTION</h3><div>Bone changes are integral to the onset and progression of OA. Many aspects remain poorly understood due to the inability to assess bone architecture in vivo. Research has relied on ex vivo imaging, hindering evaluation of early-stage disease and longitudinal analysis. Conventional CT lacks the resolution to visualise subchondral bone microstructure. While ex vivo Photon Counting CT (PCCT) has demonstrated imaging comparable to μCT, its ability to capture bone microstructure in vivo in knee OA patients under clinical conditions remains unproven.</div></div><div><h3>OBJECTIVE</h3><div>The aim of this study was to compare in vivo and ex vivo PCCT of subchondral bone features in patients with knee OA.</div></div><div><h3>METHODS</h3><div>Pre-surgery in vivo and post-surgery ex vivo PCCT (Siemens Naeotom Alpha, Siemens Healthineers, Germany) of the tibial plateau from participants with severe knee OA referred to arthroplasty surgery from January 2022 through September 2023 were compared. Acquisition/reconstruction details: a tube current of 120 kV, a matrix size of 1024 × 1024, a slice thickness of 0.2 mm, and a FOV of 150 × 150 mm. 18 in vivo/ex vivo PCCT pairs were included. The ex vivo scans was registered to the in vivo scans. Linear regression and Bland-Altman plots were used to assess correlation and agreement between in vivo and ex vivo measures of bone volume fraction (BV/TV), trabecular thickness (Tb.Th.), and attenuation in healthy and sclerotic trabecular bone. Delineated areas of bone sclerosis were compared using the Dice coefficient and Hausdorff distance, Fig. 1.</div></div><div><h3>RESULTS</h3><div>Comparing in vivo and ex vivo scans strong correlations were found for BV/TV, R<sup>2</sup>=0.82 and attenuation in both healthy, R<sup>2</sup>=0.89, and sclerotic, R<sup>2</sup>=0.79, bone, while a moderate correlation was found for Tb.Th., R<sup>2</sup>=0.55. Bias for BV/TV and Tb.Th. was -4.1% and -0.598mm, respectively, and -41.4 HU and -81.1 HU for healthy and sclerotic bone, respectively. A proportional bias was observed for BV/TV and Tb.Th., Fig. 2. There was excellent agreement between the segmentations of sclerotic areas, Dice coefficient = 0.91 and Hausdorff distance = 0.11mm.</div></div><div><h3>CONCLUSION</h3><div>In patients with severe knee OA, BV/TV and attenuation can be obtained with high correlation and small bias between in vivo and ex vivo scans. Tb.Th. showed moderate correlation and larger bias. Subchondral bone sclerosis, a key OA feature, is well translated from ex vivo to in vivo PCCT. Longitudinal studies using in vivo PCCT are feasible, but caution may be advised when measuring Tb.Th.</div></div>","PeriodicalId":74378,"journal":{"name":"Osteoarthritis imaging","volume":"5 ","pages":"Article 100304"},"PeriodicalIF":0.0,"publicationDate":"2025-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144524124","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
引用次数: 0
EFFECT OF LATERAL MENISCUS POSTERIOR ROOT TEARS ON CARTILAGE AND MENISCAL MECHANICS 外侧半月板后根撕裂对软骨和半月板力学的影响
Osteoarthritis imaging
Pub Date : 2025-01-01 DOI: 10.1016/j.ostima.2025.100353
J.S. Broberg, E. Hoptioncann, A. Kimbowa, A. Yung, K. Bale, I. Hacihaliloglu, P. Lodhia, D.R. Wilson

INTRODUCTION

Measuring cartilage and meniscal mechanics in loaded knees is essential to understanding the effects of lateral meniscus posterior root tears (LMPRTs) and the effectiveness of meniscal repair procedures that seek to protect the joint from degeneration. Studies have assessed mechanics with thin-film pressure sensors or finite element models, but their conclusions are limited by the invasiveness or inherent assumptions of the techniques employed. Ultra-high field MRI provides sufficient resolution to measure cartilage and meniscal mechanics during loading in a compatible loading device, without requiring disruption or simulation of the articulating joint surfaces. However, no studies have evaluated the impact of LMPRTs on the cartilage and meniscal mechanics in a human cadaveric knee using such a method.

OBJECTIVE

Test the hypothesis that LMPRTs increase femoral and tibial cartilage strain and meniscal extrusion.

METHODS

Six human knee lateral compartments (mean age 70 yrs) were tested. Anatomical alignment in full extension was maintained during preparation. The lateral meniscus and its roots, meniscotibial ligament, and attachment to the popliteus, as well as the ACL, were preserved. Specimens were placed in a novel pneumatic compression apparatus customized for use a 9.4T MRI scanner. Morphologic scans with a resolution of 0.06 × 0.12 × 0.4 mm were acquired before loading and after 2 hours of loading (Figure 1). The load applied was constant and equivalent to 48% body weight to simulate two-legged standing. An artificial LMPRT was then created, and specimens were left unloaded until testing the next day with the same protocol. Joint tissues were manually segmented for both intact and LMPRT conditions, in both loaded and unloaded states. Flattened cartilage profiles were generated to calculate cartilage strain in the axial direction, with negative strain indicating compression. The mean and maximum strains in the tibiofemoral contact area were determined in both the femoral and tibial cartilage. Meniscal extrusion was measured as the perpendicular distance between the external edge of the meniscus and the line bisecting the external edge of the tibial plateau and femoral condyle in the most anterior slice of the popliteus’ insertion. All measures were compared between conditions with paired Student’s t-tests with significance set to 0.05.

RESULTS

Maximum compressive strain in the tibiofemoral contact region of the femoral (p = 0.013) and tibial (p = 0.010) cartilage increased significantly after the LMPRT (Figure 2). The increase in mean compressive strain in the tibiofemoral contact region after the LMPRT was not significantly different for the femoral (p = 0.103) or tibial (p = 0.065) cartilage. Likewise, the increase in meniscal extrusion after the LMPRT was not significantly different (p = 0.143). Specimens with a greater incr
测量负重膝的软骨和半月板力学对于理解外侧半月板后根撕裂（lprts）的影响和半月板修复程序的有效性至关重要，以保护关节免受退变。研究已经用薄膜压力传感器或有限元模型评估了力学，但他们的结论受到所采用技术的侵入性或固有假设的限制。超高场MRI提供了足够的分辨率来测量软骨和半月板力学在一个兼容的加载装置中加载，而不需要破坏或模拟关节表面。然而，没有研究使用这种方法评估lprts对人尸体膝关节软骨和半月板力学的影响。目的验证lprts增加股骨、胫骨软骨劳损和半月板挤压的假说。方法对6例平均年龄70岁的人膝关节外侧腔室进行检测。在准备过程中保持完全伸展的解剖对齐。外侧半月板及其根、半月板胫韧带、与腘肌的连接以及前交叉韧带均被保留。将标本放置在为9.4T MRI扫描仪定制的新型气动压缩装置中。加载前和加载2小时后分别获得分辨率为0.06 × 0.12 × 0.4 mm的形态学扫描（图1）。施加的负荷是恒定的，相当于48%的体重来模拟两条腿站立。然后创建一个人工lprt，并将标本卸载，直到第二天按照相同的方案进行测试。在加载和卸载状态下，对完整和lprt条件下的关节组织进行手动分割。生成扁平软骨剖面，计算软骨轴向应变，负应变表示压缩。测定股骨软骨和胫骨软骨在胫股接触区的平均应变和最大应变。半月板挤压测量为半月板外缘与胫骨平台外缘和股骨髁在腘肌止点最前方切面的平分线之间的垂直距离。所有测量值在不同条件下的比较采用配对学生t检验，显著性设为0.05。结果lprt后股骨（p = 0.013）和胫骨（p = 0.010）软骨胫股接触区最大压缩应变显著升高（图2）。股骨（p = 0.103）和胫骨（p = 0.065）软骨在lprt后胫股接触区平均压缩应变的增加无显著差异。同样，lprt后半月板挤压的增加也没有显著差异（p = 0.143）。在lprt后半月板挤压增加较大的标本，在lprt后最大软骨应变往往增加较大。结论lprt后最大软骨应变升高反映了软骨应力升高，与软骨退变有关。我们发现在软骨应变增加的标本中，更多的半月板挤压突出了软骨和半月板力学之间的潜在关系，以及通过lprt修复恢复正常半月板力学的重要性。这种方法研究膝关节力学的一个关键优势是能够同时评估半月板和软骨力学，对对齐和关键软组织的破坏最小。该方法具有评估半月板修复技术有效性的潜力。
{"title":"EFFECT OF LATERAL MENISCUS POSTERIOR ROOT TEARS ON CARTILAGE AND MENISCAL MECHANICS","authors":"J.S. Broberg,&nbsp;E. Hoptioncann,&nbsp;A. Kimbowa,&nbsp;A. Yung,&nbsp;K. Bale,&nbsp;I. Hacihaliloglu,&nbsp;P. Lodhia,&nbsp;D.R. Wilson","doi":"10.1016/j.ostima.2025.100353","DOIUrl":"10.1016/j.ostima.2025.100353","url":null,"abstract":"<div><h3>INTRODUCTION</h3><div>Measuring cartilage and meniscal mechanics in loaded knees is essential to understanding the effects of lateral meniscus posterior root tears (LMPRTs) and the effectiveness of meniscal repair procedures that seek to protect the joint from degeneration. Studies have assessed mechanics with thin-film pressure sensors or finite element models, but their conclusions are limited by the invasiveness or inherent assumptions of the techniques employed. Ultra-high field MRI provides sufficient resolution to measure cartilage and meniscal mechanics during loading in a compatible loading device, without requiring disruption or simulation of the articulating joint surfaces. However, no studies have evaluated the impact of LMPRTs on the cartilage and meniscal mechanics in a human cadaveric knee using such a method.</div></div><div><h3>OBJECTIVE</h3><div>Test the hypothesis that LMPRTs increase femoral and tibial cartilage strain and meniscal extrusion.</div></div><div><h3>METHODS</h3><div>Six human knee lateral compartments (mean age 70 yrs) were tested. Anatomical alignment in full extension was maintained during preparation. The lateral meniscus and its roots, meniscotibial ligament, and attachment to the popliteus, as well as the ACL, were preserved. Specimens were placed in a novel pneumatic compression apparatus customized for use a 9.4T MRI scanner. Morphologic scans with a resolution of 0.06 × 0.12 × 0.4 mm were acquired before loading and after 2 hours of loading (Figure 1). The load applied was constant and equivalent to 48% body weight to simulate two-legged standing. An artificial LMPRT was then created, and specimens were left unloaded until testing the next day with the same protocol. Joint tissues were manually segmented for both intact and LMPRT conditions, in both loaded and unloaded states. Flattened cartilage profiles were generated to calculate cartilage strain in the axial direction, with negative strain indicating compression. The mean and maximum strains in the tibiofemoral contact area were determined in both the femoral and tibial cartilage. Meniscal extrusion was measured as the perpendicular distance between the external edge of the meniscus and the line bisecting the external edge of the tibial plateau and femoral condyle in the most anterior slice of the popliteus’ insertion. All measures were compared between conditions with paired Student’s t-tests with significance set to 0.05.</div></div><div><h3>RESULTS</h3><div>Maximum compressive strain in the tibiofemoral contact region of the femoral (p = 0.013) and tibial (p = 0.010) cartilage increased significantly after the LMPRT (Figure 2). The increase in mean compressive strain in the tibiofemoral contact region after the LMPRT was not significantly different for the femoral (p = 0.103) or tibial (p = 0.065) cartilage. Likewise, the increase in meniscal extrusion after the LMPRT was not significantly different (p = 0.143). Specimens with a greater incr","PeriodicalId":74378,"journal":{"name":"Osteoarthritis imaging","volume":"5 ","pages":"Article 100353"},"PeriodicalIF":0.0,"publicationDate":"2025-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144524180","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
引用次数: 0
REVEALING EARLY SUBCHONDRAL BONE STRUCTURAL CHANGES IN OSTEOARTHRITIS PROGRESSION IN A COLLAGENASE-INDUCED MOUSE MODEL USING MICRO COMPUTED TOMOGRAPHY 在胶原酶诱导的小鼠模型中使用显微计算机断层扫描揭示骨关节炎进展的早期软骨下骨结构变化
Osteoarthritis imaging
Pub Date : 2025-01-01 DOI: 10.1016/j.ostima.2025.100301
H. Liu, Z. Li, C.E. Davey, K.S. Stok

INTRODUCTION

The deployment of micro-computed tomography (microCT) enables quantitative morphometric analysis (QMA) to quantify morphological and structural changes caused by OA in mouse knee joint with excellent spatial resolution. Previous studies quantifying microstructural changes to subchondral tibiae in fortnightly intervals, report bone loss and trabecular thinning as early as two weeks post disease induction in mouse models. However, evidence suggests that the subchondral bone turnover may occur earlier than two weeks post disease induction in a mouse OA model.

OBJECTIVE

To reveal early bone microstructural changes associated with OA progression in a mouse model with a high temporal resolution using microCT and QMA.

METHODS

Seventy-five male C57BL/10 mice aged nine weeks were recruited and randomly assigned to three cross-sectional cohorts, i.e., baseline (n = 4), control (n = 24) and OA (n = 47) cohorts. Forty-seven ten-week-old mice assigned to OA cohort received intra-articular injection of 10 unit of filtered collagenase dissolved in 6 µl physiological saline to the right joints (OA group) through the patellar ligament. A similar volume of saline was intraarticularly injected to the left contralateral joints (CTLR group). Prior to scanning, mice were euthanized at 0-, 1-, 2-, 3-, 4-, 5-, 6-, 7-, and 8-weeks post ten-week-old. Scans were performed using microCT (vivaCT80, SCANCO Medical AG, Brüttisellen, Switzerland) with a source voltage of 70 kVp, an integration time of 350 ms, a current of 114 µA, a nominal resolution of 10.4 µm, and 500 projections with each scan taking around 20 minutes. QMA was performed to quantify changes to subchondral bone microstructure associated with OA progression. To detect differences between treatments at each time point, a linear mixed-effect model was used. Individual mice were considered as random effects, time points (1- to 8- weeks post collagenase injection) and treatment (CT, CTLR, and OA) were considered as fixed effects.

RESULTS

Representative segmented microCT images from CT and OA group can be found in Figure 1 A. Typical osteoarthritic characteristics were observed in OA group at multiple time points, with changes detectable as early as one week post disease induction, shown in Figure 1 B. Specifically, comparing joints from CT and CTLR groups, smaller trabecular thickness, Tb.Th, were observed at both lateral and medial sides in OA femora, in accordance with the increasing trabecular spacing, Tb.Sp, and decreasing trabecular number, Tb.N.

CONCLUSION

This study, for the first time, demonstrated that prominent bone changes could be detected as early as one week after disease induction. These findings underscore the necessity of early quantification to capture rapidly changing bone microstructure alterations in early s
微计算机断层扫描（microCT）的部署使定量形态分析（QMA）能够以极好的空间分辨率量化小鼠膝关节OA引起的形态和结构变化。先前的研究每隔两周量化软骨下胫骨的微结构变化，报告早在小鼠模型疾病诱导后两周骨丢失和小梁变薄。然而，有证据表明，在小鼠OA模型中，软骨下骨转换可能早于疾病诱导后两周发生。目的利用微ct和QMA技术揭示高时间分辨率小鼠骨关节炎进展的早期骨微结构变化。方法招募9周龄雄性C57BL/10小鼠75只，随机分为基线组（n = 4）、对照组（n = 24）和OA组（n = 47）3个横断面队列。选取10周龄OA组小鼠47只，经髌骨韧带向右侧关节（OA组）关节内注射溶解于6µl生理盐水中过滤后的胶原酶10单位。左对侧关节内注射等量生理盐水（CTLR组）。在扫描前，小鼠在0、1、2、3、4、5、6、7和8周时被安乐死。扫描使用microCT （vivaCT80, SCANCO Medical AG, br ttisellen, Switzerland）进行，源电压为70 kVp，积分时间为350 ms，电流为114µa，标称分辨率为10.4µm， 500个投影，每次扫描大约需要20分钟。QMA用于量化与骨关节炎进展相关的软骨下骨微结构的变化。为了检测各时间点处理间的差异，采用线性混合效应模型。个体小鼠被认为是随机效应，时间点（胶原酶注射后1- 8周）和治疗（CT、CTLR和OA）被认为是固定效应。Figure 1a为CT组和OA组具有代表性的微CT分割图像。OA组在多个时间点观察到典型的骨关节炎特征，早在疾病诱导后一周就可以检测到变化，如图1 b所示。其中，对比CT组和CTLR组的关节，骨小梁厚度较小，Tb。随着骨小梁间距的增加，在OA股骨的外侧和内侧均观察到Th。结论本研究首次证实，早在疾病诱导后1周就可以检测到明显的骨变化。这些发现强调了早期量化的必要性，以捕获早期OA中快速变化的骨微观结构改变，从而有可能实现早期诊断、干预和治疗。
{"title":"REVEALING EARLY SUBCHONDRAL BONE STRUCTURAL CHANGES IN OSTEOARTHRITIS PROGRESSION IN A COLLAGENASE-INDUCED MOUSE MODEL USING MICRO COMPUTED TOMOGRAPHY","authors":"H. Liu,&nbsp;Z. Li,&nbsp;C.E. Davey,&nbsp;K.S. Stok","doi":"10.1016/j.ostima.2025.100301","DOIUrl":"10.1016/j.ostima.2025.100301","url":null,"abstract":"<div><h3>INTRODUCTION</h3><div>The deployment of micro-computed tomography (microCT) enables quantitative morphometric analysis (QMA) to quantify morphological and structural changes caused by OA in mouse knee joint with excellent spatial resolution. Previous studies quantifying microstructural changes to subchondral tibiae in fortnightly intervals, report bone loss and trabecular thinning as early as two weeks post disease induction in mouse models. However, evidence suggests that the subchondral bone turnover may occur earlier than two weeks post disease induction in a mouse OA model.</div></div><div><h3>OBJECTIVE</h3><div>To reveal early bone microstructural changes associated with OA progression in a mouse model with a high temporal resolution using microCT and QMA.</div></div><div><h3>METHODS</h3><div>Seventy-five male C57BL/10 mice aged nine weeks were recruited and randomly assigned to three cross-sectional cohorts, i.e., baseline (n = 4), control (n = 24) and OA (n = 47) cohorts. Forty-seven ten-week-old mice assigned to OA cohort received intra-articular injection of 10 unit of filtered collagenase dissolved in 6 µl physiological saline to the right joints (OA group) through the patellar ligament. A similar volume of saline was intraarticularly injected to the left contralateral joints (CTLR group). Prior to scanning, mice were euthanized at 0-, 1-, 2-, 3-, 4-, 5-, 6-, 7-, and 8-weeks post ten-week-old. Scans were performed using microCT (vivaCT80, SCANCO Medical AG, Brüttisellen, Switzerland) with a source voltage of 70 kVp, an integration time of 350 <em>ms</em>, a current of 114 µA, a nominal resolution of 10.4 µm, and 500 projections with each scan taking around 20 minutes. QMA was performed to quantify changes to subchondral bone microstructure associated with OA progression. To detect differences between treatments at each time point, a linear mixed-effect model was used. Individual mice were considered as random effects, time points (1- to 8- weeks post collagenase injection) and treatment (CT, CTLR, and OA) were considered as fixed effects.</div></div><div><h3>RESULTS</h3><div>Representative segmented microCT images from CT and OA group can be found in <strong>Figure 1 A</strong>. Typical osteoarthritic characteristics were observed in OA group at multiple time points, with changes detectable as early as one week post disease induction, shown in <strong>Figure 1 B</strong>. Specifically, comparing joints from CT and CTLR groups, smaller trabecular thickness, Tb.Th, were observed at both lateral and medial sides in OA femora, in accordance with the increasing trabecular spacing, Tb.Sp, and decreasing trabecular number, Tb.N.</div></div><div><h3>CONCLUSION</h3><div>This study, for the first time, demonstrated that prominent bone changes could be detected as early as one week after disease induction. These findings underscore the necessity of early quantification to capture rapidly changing bone microstructure alterations in early s","PeriodicalId":74378,"journal":{"name":"Osteoarthritis imaging","volume":"5 ","pages":"Article 100301"},"PeriodicalIF":0.0,"publicationDate":"2025-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144524186","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
引用次数: 0
CHARACTERIZING MENISCAL CALCIFICATIONS WITH PHOTON COUNTING-BASED DUAL-ENERGY COMPUTED TOMOGRAPHY 基于光子计数的双能计算机断层扫描表征半月板钙化
Osteoarthritis imaging
Pub Date : 2025-01-01 DOI: 10.1016/j.ostima.2025.100303
E. Nevanranta , V.-P. Karjalainen , M. Brix , I. Hellberg , A. Turkiewicz , B. Shakya , P. Önnerfjord , S. Ylisiurua , A. Sjögren , K. Elkhouly , V. Hughes , J. Tjörnstrand , S. Saarakkala , M. Englund , M.A.J. Finnilä

INTRODUCTION

Meniscal calcifications, including basic calcium phosphate (BCP) and calcium pyrophosphate (CPP), are commonly associated with OA and may disrupt meniscal function, contributing to joint degeneration. However, the role of specific calcification types in OA is not fully understood due to the lack of non-invasive imaging techniques that can differentiate them in vivo. While Raman spectroscopy accurately distinguishes BCP from CPP, it is limited to 2D and requires destructive histological processing. In contrast, dual-energy computed tomography (DECT) has shown potential for differentiating calcifications in both in vivo and ex vivo, but its performance varies across previous studies. The integration of photon-counting detectors (PCD) in CT imaging improves spatial resolution and enables multi-energy acquisition, enhancing in vivo calcification characterization.

OBJECTIVE

We evaluated the capability of dual-energy computed tomography with a photon counting detector (PCD-DECT) to differentiate BCP and CPP calcification deposits in the posterior horns of human menisci ex vivo, using Raman spectroscopy as the reference.

METHODS

This study included 82 medial and lateral meniscus samples from 21 deceased donors without known knee OA and 20 TKR patients with medial compartment OA. Samples were imaged using an experimental cone-beam CT setup with PCD, operating at 120 kVp and 0.2 mA. Low energy (LE) data were collected in the 20-50 keV range, and high energy (HE) data in the 50-120 keV range, with a final voxel size of 37 µm. Only calcified samples identified using Raman spectroscopy (n = 36), 8 CPP and 28 BCP samples, were included to the analysis. Calcifications were segmented and divided between BCP and CPP groups. Subsequently, LE, HE, and Dual Energy Index (DEI) values were measured for each calcification. We used linear mixed models to estimate associations between LE and HE variables and the calcification type, and to compare the DEI values between the calcification types. Estimates are presented with 95% confidence intervals.

RESULTS

Figure 1A-C shows a 3D visualization of menisci with and without different calcifications. The results showed that CPP calcifications had consistently lower LE values than BCP for corresponding HE values. The difference increased with higher HE values, peaking at 500 HU with a difference of 166.1 HU (95% CI: 73.4, 258.8), while the smallest difference occurs at -100 HU, where the difference is 33.81 HU (95% CI: -40.38, 107.99) HU. The differences between LE and HE values are shown in Figure 1D-E. Additionally, estimated mean DEI values were higher in BCP calcifications compared to CPP, with an estimated difference of 0.035 (95%CI: 0.011, 0.059). Detailed results are shown in Table 1.

CONCLUSION

Our findings show that BCP and CPP m
半月板钙化，包括碱性磷酸钙（BCP）和焦磷酸钙（CPP），通常与OA相关，并可能破坏半月板功能，导致关节退变。然而，由于缺乏能够在体内区分特定钙化类型的非侵入性成像技术，因此尚未完全了解特定钙化类型在OA中的作用。虽然拉曼光谱可以准确地区分BCP和CPP，但它仅限于二维，并且需要破坏性的组织学处理。相比之下，双能计算机断层扫描（DECT）已经显示出在体内和体外鉴别钙化的潜力，但其性能在以前的研究中有所不同。光子计数探测器（PCD）在CT成像中的集成提高了空间分辨率，实现了多能采集，增强了体内钙化表征。目的以拉曼光谱为参考，评价双能光子计数检测器（PCD-DECT）对人半月板离体后角BCP和CPP钙化沉积的鉴别能力。方法本研究包括21例无已知膝关节OA的已故供体和20例有内侧室OA的TKR患者的82个内外侧半月板样本。使用带PCD的实验锥束CT装置对样品进行成像，工作电压为120 kVp和0.2 mA。低能量（LE）数据在20-50 keV范围内采集，高能量（HE）数据在50-120 keV范围内采集，最终体素尺寸为37µm。只有通过拉曼光谱鉴定的钙化样品（n = 36），8个CPP和28个BCP样品被纳入分析。在BCP组和CPP组对钙化进行分段和划分。随后，测量每个钙化的LE、HE和双能指数（DEI）值。我们使用线性混合模型来估计LE和HE变量与钙化类型之间的关联，并比较钙化类型之间的DEI值。估计值以95%置信区间表示。结果图1A-C显示了有无不同钙化的半月板的三维可视化。结果表明，CPP钙化的LE值始终低于BCP对应的HE值。HE值越高，差异越大，在500 HU时达到峰值，差异为166.1 HU (95% CI: 73.4, 258.8)，而在-100 HU时差异最小，差异为33.81 HU (95% CI: -40.38, 107.99) HU。LE和HE值的差异如图1D-E所示。此外，与CPP相比，BCP钙化的估计平均DEI值更高，估计差异为0.035 （95%CI: 0.011, 0.059）。详细结果见表1。结论BCP和CPP半月板钙化在LE和HE以及PCD-DECT测量的DEI值上存在差异。该方法揭示了钙化类型之间的平均差异，而在未来，更先进的PCD探测器可以提高对单个钙化的精确识别。总之，PCD-DECT成功地实现了半月板钙化类型的体外评估，突出了其在未来体内应用的潜力，以更好地了解钙化机制和评估对钙化靶向治疗的反应。
{"title":"CHARACTERIZING MENISCAL CALCIFICATIONS WITH PHOTON COUNTING-BASED DUAL-ENERGY COMPUTED TOMOGRAPHY","authors":"E. Nevanranta ,&nbsp;V.-P. Karjalainen ,&nbsp;M. Brix ,&nbsp;I. Hellberg ,&nbsp;A. Turkiewicz ,&nbsp;B. Shakya ,&nbsp;P. Önnerfjord ,&nbsp;S. Ylisiurua ,&nbsp;A. Sjögren ,&nbsp;K. Elkhouly ,&nbsp;V. Hughes ,&nbsp;J. Tjörnstrand ,&nbsp;S. Saarakkala ,&nbsp;M. Englund ,&nbsp;M.A.J. Finnilä","doi":"10.1016/j.ostima.2025.100303","DOIUrl":"10.1016/j.ostima.2025.100303","url":null,"abstract":"<div><h3>INTRODUCTION</h3><div>Meniscal calcifications, including basic calcium phosphate (BCP) and calcium pyrophosphate (CPP), are commonly associated with OA and may disrupt meniscal function, contributing to joint degeneration. However, the role of specific calcification types in OA is not fully understood due to the lack of non-invasive imaging techniques that can differentiate them <em>in vivo</em>. While Raman spectroscopy accurately distinguishes BCP from CPP, it is limited to 2D and requires destructive histological processing. In contrast, dual-energy computed tomography (DECT) has shown potential for differentiating calcifications in both <em>in vivo</em> and <em>ex vivo</em>, but its performance varies across previous studies. The integration of photon-counting detectors (PCD) in CT imaging improves spatial resolution and enables multi-energy acquisition, enhancing <em>in vivo</em> calcification characterization.</div></div><div><h3>OBJECTIVE</h3><div>We evaluated the capability of dual-energy computed tomography with a photon counting detector (PCD-DECT) to differentiate BCP and CPP calcification deposits in the posterior horns of human menisci <em>ex vivo</em>, using Raman spectroscopy as the reference.</div></div><div><h3>METHODS</h3><div>This study included 82 medial and lateral meniscus samples from 21 deceased donors without known knee OA and 20 TKR patients with medial compartment OA. Samples were imaged using an experimental cone-beam CT setup with PCD, operating at 120 kVp and 0.2 mA. Low energy (LE) data were collected in the 20-50 keV range, and high energy (HE) data in the 50-120 keV range, with a final voxel size of 37 µm. Only calcified samples identified using Raman spectroscopy (n = 36), 8 CPP and 28 BCP samples, were included to the analysis. Calcifications were segmented and divided between BCP and CPP groups. Subsequently, LE, HE, and Dual Energy Index (DEI) values were measured for each calcification. We used linear mixed models to estimate associations between LE and HE variables and the calcification type, and to compare the DEI values between the calcification types. Estimates are presented with 95% confidence intervals.</div></div><div><h3>RESULTS</h3><div>Figure 1A-C shows a 3D visualization of menisci with and without different calcifications. The results showed that CPP calcifications had consistently lower LE values than BCP for corresponding HE values. The difference increased with higher HE values, peaking at 500 HU with a difference of 166.1 HU (95% CI: 73.4, 258.8), while the smallest difference occurs at -100 HU, where the difference is 33.81 HU (95% CI: -40.38, 107.99) HU. The differences between LE and HE values are shown in Figure 1D-E. Additionally, estimated mean DEI values were higher in BCP calcifications compared to CPP, with an estimated difference of 0.035 (95%CI: 0.011, 0.059). Detailed results are shown in Table 1.</div></div><div><h3>CONCLUSION</h3><div>Our findings show that BCP and CPP m","PeriodicalId":74378,"journal":{"name":"Osteoarthritis imaging","volume":"5 ","pages":"Article 100303"},"PeriodicalIF":0.0,"publicationDate":"2025-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144524188","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
引用次数: 0
AUTOMATED QUANTIFICATION OF MENISCUS EXTRUSION IN MRI VIA AI FOUNDATION MODEL: PROOF OF CONCEPT USING A TRAINING-FREE FEW-SHOT SEGMENTATION APPROACH 基于ai基础模型的mri半月板挤压的自动量化：使用无训练的少镜头分割方法的概念验证
Osteoarthritis imaging
Pub Date : 2025-01-01 DOI: 10.1016/j.ostima.2025.100333
Z. Zhou , X. He , Y. Hu , H.A. Khan , F. Liu , M. Jarraya

INTRODUCTION

Manual assessment of meniscus extrusion (ME) in magnetic resonance (MR) images is time-consuming and prone to variability, limiting efficiency in clinical and research settings. While deep learning methods have shown promise in MR image segmentation, their reliance on task-specific training and large annotated datasets limits scalability and adaptability.

OBJECTIVE

Building upon our previously developed AI foundation model, we aim to establish a fully automated pipeline for quantifying ME in knee MRI with our model training and eliminate the need for large annotated datasets.

METHODS

By providing a support set including a minimal number of segmentation examples, the AI Foundation Model enables accurate segmentation of knee anatomy and reliable ME measurement in a training-free, few-shot manner. In the study, we analyzed 3T MR images acquired using either T2-weighted or proton density MR sequences from 10 patients with mild osteoarthritis. Manual segmentations of femur, tibia, medial, and lateral menisci were performed by experts. Two patients, one with T2-weighted and one with proton density images, were randomly selected to build the support set. The remaining 8 patients comprised the testing set, which was used for both automated segmentation and model evaluation. Segmentation performance was assessed using the Dice Coefficient. For ME evaluation, an experienced radiologist manually identified the slice containing the tibial spine and measured extrusion as the reference. Automated ME measurement was computed from the segmentation by detecting the femoral condyle and tibial plateau edge, then measuring the distance from the most medial point of the medial meniscus to a reference line connecting the femoral condyle and tibial plateau edge.

RESULTS

The average Dice Coefficient was 94.07 ± 3.97% for the femur, 97.09 ± 0.93% for the tibia, 82.91 ± 6.72% for the medial meniscus, and 85.49 ± 5.24% for the lateral meniscus. ME measurements predicted by the model were also compared with ground truth values. The human measured ME was 4.26 ± 1.46 mm, while the model-predicted ME was 4.18 ± 1.16 mm.

CONCLUSION

This study demonstrates that the foundation model enables reliable and fully automated quantification of meniscus extrusion from knee MR images without requiring training or large annotated datasets. With only two support examples, the model achieved accurate segmentation and ME measurement across eight testing subjects, underscoring its efficiency and strong generalization. Its consistent performance across key anatomical structures highlights its potential for expert-level evaluation in both clinical and research settings with minimal manual effort. Further work will explore semi-automated expansion of the support set and extension to diverse MRI protocols and osteoarthritis severities, and validation on
磁共振（MR）图像中半月板挤压（ME）的人工评估耗时且容易变化，限制了临床和研究环境的效率。虽然深度学习方法在MR图像分割中显示出前景，但它们对特定任务训练和大型注释数据集的依赖限制了可扩展性和适应性。在我们之前开发的AI基础模型的基础上，我们的目标是建立一个全自动的管道，通过我们的模型训练来量化膝关节MRI中的ME，并消除对大型注释数据集的需求。方法通过提供一个支持集，包括最少量的分割示例，人工智能基础模型能够以无训练、少镜头的方式准确分割膝关节解剖和可靠的ME测量。在研究中，我们分析了10例轻度骨关节炎患者使用t2加权或质子密度MR序列获得的3T MR图像。由专家进行股骨、胫骨、内侧和外侧半月板的手工分割。随机选择2例患者，其中1例为t2加权图像，1例为质子密度图像，建立支持集。其余8例患者组成测试集，用于自动分割和模型评估。使用Dice系数评估分割性能。对于ME评估，经验丰富的放射科医生手动识别包含胫骨脊柱的切片并测量挤压作为参考。通过检测股骨髁和胫骨平台边缘的分割，然后测量内侧半月板最中间点到连接股骨髁和胫骨平台边缘的参考线的距离，计算自动ME测量。结果股骨的平均Dice系数为94.07 ±3.97%，胫骨为97.09 ± 0.93%，内侧半月板为82.91 ± 6.72%，外侧半月板为85.49 ± 5.24%。模型预测的ME测量值也与地面真值进行了比较。人体测量ME为4.26 ± 1.46 mm，而模型预测ME为4.18 ± 1.16 mm。结论：本研究表明，该基础模型能够可靠且全自动地从膝关节MR图像中量化半月板挤压，而无需训练或大型注释数据集。仅用两个支持例，该模型就实现了8个测试对象的准确分割和ME测量，突出了其效率和较强的泛化能力。它在关键解剖结构上的一致表现突出了它在临床和研究环境中以最小的人工工作量进行专家级评估的潜力。进一步的工作将探索支持集的半自动扩展，扩展到不同的MRI协议和骨关节炎严重程度，并在更大规模的数据集上进行验证。
{"title":"AUTOMATED QUANTIFICATION OF MENISCUS EXTRUSION IN MRI VIA AI FOUNDATION MODEL: PROOF OF CONCEPT USING A TRAINING-FREE FEW-SHOT SEGMENTATION APPROACH","authors":"Z. Zhou ,&nbsp;X. He ,&nbsp;Y. Hu ,&nbsp;H.A. Khan ,&nbsp;F. Liu ,&nbsp;M. Jarraya","doi":"10.1016/j.ostima.2025.100333","DOIUrl":"10.1016/j.ostima.2025.100333","url":null,"abstract":"<div><h3>INTRODUCTION</h3><div>Manual assessment of meniscus extrusion (ME) in magnetic resonance (MR) images is time-consuming and prone to variability, limiting efficiency in clinical and research settings. While deep learning methods have shown promise in MR image segmentation, their reliance on task-specific training and large annotated datasets limits scalability and adaptability.</div></div><div><h3>OBJECTIVE</h3><div>Building upon our previously developed AI foundation model, we aim to establish a fully automated pipeline for quantifying ME in knee MRI with our model training and eliminate the need for large annotated datasets.</div></div><div><h3>METHODS</h3><div>By providing a support set including a minimal number of segmentation examples, the AI Foundation Model enables accurate segmentation of knee anatomy and reliable ME measurement in a training-free, few-shot manner. In the study, we analyzed 3T MR images acquired using either T2-weighted or proton density MR sequences from 10 patients with mild osteoarthritis. Manual segmentations of femur, tibia, medial, and lateral menisci were performed by experts. Two patients, one with T2-weighted and one with proton density images, were randomly selected to build the support set. The remaining 8 patients comprised the testing set, which was used for both automated segmentation and model evaluation. Segmentation performance was assessed using the Dice Coefficient. For ME evaluation, an experienced radiologist manually identified the slice containing the tibial spine and measured extrusion as the reference. Automated ME measurement was computed from the segmentation by detecting the femoral condyle and tibial plateau edge, then measuring the distance from the most medial point of the medial meniscus to a reference line connecting the femoral condyle and tibial plateau edge.</div></div><div><h3>RESULTS</h3><div>The average Dice Coefficient was 94.07 ± 3.97% for the femur, 97.09 ± 0.93% for the tibia, 82.91 ± 6.72% for the medial meniscus, and 85.49 ± 5.24% for the lateral meniscus. ME measurements predicted by the model were also compared with ground truth values. The human measured ME was 4.26 ± 1.46 mm, while the model-predicted ME was 4.18 ± 1.16 mm.</div></div><div><h3>CONCLUSION</h3><div>This study demonstrates that the foundation model enables reliable and fully automated quantification of meniscus extrusion from knee MR images without requiring training or large annotated datasets. With only two support examples, the model achieved accurate segmentation and ME measurement across eight testing subjects, underscoring its efficiency and strong generalization. Its consistent performance across key anatomical structures highlights its potential for expert-level evaluation in both clinical and research settings with minimal manual effort. Further work will explore semi-automated expansion of the support set and extension to diverse MRI protocols and osteoarthritis severities, and validation on","PeriodicalId":74378,"journal":{"name":"Osteoarthritis imaging","volume":"5 ","pages":"Article 100333"},"PeriodicalIF":0.0,"publicationDate":"2025-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144524261","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
引用次数: 0
REGIONAL VARIATION IN TRAPEZIOMETACARPAL BONE MICROARCHITECTURE IN FEMALES WITH OSTEOARTHRITIS USING HR-PQCT 利用hr-pqct观察女性骨关节炎患者的骨梯跖骨微结构的区域差异
Osteoarthritis imaging
Pub Date : 2025-01-01 DOI: 10.1016/j.ostima.2025.100299
M.T. Kuczynski , C. Hasselaar , G. Dhaliwal , C. Hiscox , N.J. White , S.L. Manske

INTRODUCTION

The trapeziometacarpal (TMC) joint, comprised of the trapezium (TRP) and first metacarpal (MC1) bones, is a mechanically complex, saddle-shaped joint. Studies have estimated that the peak forces acting on the TMC joint are up to five times higher than the corresponding external forces [1]. Moreover, cadaveric studies have shown non-uniform cartilage loss in TMC joint with OA [2]. While several cadaveric studies have investigated TMC joint cartilage and bone changes, evaluation of subchondral bone changes in the TMC joint in vivo is lacking.

OBJECTIVE

The objective of this study was to investigate differences in bone microarchitecture in anatomical quadrants of the TMC joint in women with TMC OA compared to age- and sex-matched controls. We hypothesized that women with TMC OA will exhibit quadrant-specific differences in bone microarchitecture compared to controls. Specifically, we hypothesized that the volar region of the TMC joint will demonstrate an increase in trabecular thickness, bone volume, and volumetric bone mineral density due to localized bone adaptations as a response to increased loading in the volar region.

METHODS

14 females diagnosed with symptomatic TMC OA (mean age: 60 ± 6.5 years) and 12 similarly aged female controls (mean age: 59 ± 5.7 years) were scanned using HR-pQCT (XtremeCT2, Scanco Medical). A standard HR-pQCT scanning protocol was used (61 µm3 voxels). Images were preprocessed using a Laplace-Hamming filter and segmented with a fixed threshold (15% of the maximum intensity). A bone coordinate system was automatically defined for the MC1 and TRP [3], and used to separate each bone into four anatomical quadrants: 1) radial-dorsal (RD), 2) radial-volar (RV), 3) ulnar-dorsal (UD), and 4) ulnar-volar (UV). For each whole bone and quadrant, we computed volumetric bone mineral density (vBMD, mg HA/cm3), bone volume fraction (BV/TV, %), and bone thickness (B.Th, mm). A mixed ANOVA was used to compare bone measures in each bone and quadrant between groups.

RESULTS

We did not observe a significant difference in total bone parameters between groups for the MC1 or TRP. However, we found a statistically significant interaction effect between the volar and dorsal quadrants of the TRP and group for B.Th (p = 0.02, Figure 1, Table 1). Compared to controls, the mean B.Th in the TRP of the OA group was 1.9% lower in the RD quadrant, 7.5% lower in the UD quadrant, 4.8% greater in the RV quadrant, and 6.2% greater in the UV quadrant.

CONCLUSION

Our results suggest that whole bone TMC microarchitecture may not differ between OA and controls; however, we found significant differences in quadrant bone microarchitecture. This suggests that the MC1 and TRP undergo localized bone microarchitectural changes to adapt to the loading of the TMC joint. Further, our results s
由斜方骨（TRP）和第一掌骨（MC1）组成的斜方骨（TMC）关节是一个机械复杂的鞍形关节。研究估计，作用在TMC关节上的峰值力比相应的外力[1]高出5倍。此外，尸体研究显示，患有OA的TMC关节存在不均匀的软骨丢失。虽然一些尸体研究已经研究了TMC关节软骨和骨的变化，但缺乏对TMC关节软骨下骨在体内变化的评估。目的：本研究的目的是研究与年龄和性别匹配的对照组相比，女性TMC骨性关节炎关节解剖象限骨微结构的差异。我们假设，与对照组相比，患有TMC骨性关节炎的女性将在骨微结构方面表现出象限特异性差异。具体来说，我们假设TMC关节的掌侧区域将表现出小梁厚度、骨体积和体积骨矿物质密度的增加，这是由于局部骨适应作为掌侧区域负荷增加的反应。方法采用HR-pQCT （XtremeCT2, Scanco Medical）对14例诊断为症状性TMC OA的女性（平均年龄：60±6.5岁）和12例年龄相似的女性（平均年龄：59±5.7岁）进行扫描。采用标准的HR-pQCT扫描方案（61µm3体素）。使用拉普拉斯-汉明滤波器对图像进行预处理，并使用固定阈值（最大强度的15%）对图像进行分割。为MC1和TRP[3]自动定义骨坐标系，并用于将每个骨划分为四个解剖象限：1)桡侧-背侧（RD）， 2)桡侧-掌侧（RV）， 3)尺侧-背侧（UD）和4)尺侧-掌侧（UV）。对于每个全骨和象限，我们计算了体积骨矿物质密度（vBMD, mg HA/cm3），骨体积分数（BV/TV, %）和骨厚度（b.t, mm）。采用混合方差分析比较各组间各骨和象限的骨测量值。结果MC1和TRP组间骨总参数无显著差异。然而，我们发现TRP的掌侧和背侧象限与B.Th组之间存在统计学上显著的相互作用效应（p = 0.02，图1，表1）。与对照组相比，OA组TRP的平均B.Th在RD象限低1.9%，在UD象限低7.5%，在RV象限高4.8%，在UV象限高6.2%。结论OA与对照组全骨TMC微结构无明显差异；然而，我们发现象限骨微结构有显著差异。这表明MC1和TRP发生了局部骨微结构变化以适应TMC关节的负荷。此外，我们的研究结果表明，掌侧区域的骨厚度可能随着TMC OA而增加。TMC关节韧带有助于在关节中分配力，这可能会影响TMC OA。Koff等人发现OA患者TMC关节掌侧区软骨变薄，这可能是由于负荷增加所致。在本研究中，骨没有进一步细分为小梁区和皮质区，因为斜方骨在这些区域之间没有明确的分离。结合小样本量，这可以解释各组之间vBMD和BV/TV缺乏意义的原因。软骨下硬化骨厚50%的尸体斜方肌与OA bb0。因此，开发一种算法来可靠地分离这些梯形区域，可能会进一步了解TMC OA对皮质骨和小梁骨的区域影响。
{"title":"REGIONAL VARIATION IN TRAPEZIOMETACARPAL BONE MICROARCHITECTURE IN FEMALES WITH OSTEOARTHRITIS USING HR-PQCT","authors":"M.T. Kuczynski ,&nbsp;C. Hasselaar ,&nbsp;G. Dhaliwal ,&nbsp;C. Hiscox ,&nbsp;N.J. White ,&nbsp;S.L. Manske","doi":"10.1016/j.ostima.2025.100299","DOIUrl":"10.1016/j.ostima.2025.100299","url":null,"abstract":"<div><h3>INTRODUCTION</h3><div>The trapeziometacarpal (TMC) joint, comprised of the trapezium (TRP) and first metacarpal (MC1) bones, is a mechanically complex, saddle-shaped joint. Studies have estimated that the peak forces acting on the TMC joint are up to five times higher than the corresponding external forces [1]. Moreover, cadaveric studies have shown non-uniform cartilage loss in TMC joint with OA [2]. While several cadaveric studies have investigated TMC joint cartilage and bone changes, evaluation of subchondral bone changes in the TMC joint <em>in vivo</em> is lacking.</div></div><div><h3>OBJECTIVE</h3><div>The objective of this study was to investigate differences in bone microarchitecture in anatomical quadrants of the TMC joint in women with TMC OA compared to age- and sex-matched controls. We hypothesized that women with TMC OA will exhibit quadrant-specific differences in bone microarchitecture compared to controls. Specifically, we hypothesized that the volar region of the TMC joint will demonstrate an increase in trabecular thickness, bone volume, and volumetric bone mineral density due to localized bone adaptations as a response to increased loading in the volar region.</div></div><div><h3>METHODS</h3><div>14 females diagnosed with symptomatic TMC OA (mean age: 60 ± 6.5 years) and 12 similarly aged female controls (mean age: 59 ± 5.7 years) were scanned using HR-pQCT (XtremeCT2, Scanco Medical). A standard HR-pQCT scanning protocol was used (61 µm<sup>3</sup> voxels). Images were preprocessed using a Laplace-Hamming filter and segmented with a fixed threshold (15% of the maximum intensity). A bone coordinate system was automatically defined for the MC1 and TRP [3], and used to separate each bone into four anatomical quadrants: 1) radial-dorsal (RD), 2) radial-volar (RV), 3) ulnar-dorsal (UD), and 4) ulnar-volar (UV). For each whole bone and quadrant, we computed volumetric bone mineral density (vBMD, mg HA/cm<sup>3</sup>), bone volume fraction (BV/TV, %), and bone thickness (B.Th, mm). A mixed ANOVA was used to compare bone measures in each bone and quadrant between groups.</div></div><div><h3>RESULTS</h3><div>We did not observe a significant difference in total bone parameters between groups for the MC1 or TRP. However, we found a statistically significant interaction effect between the volar and dorsal quadrants of the TRP and group for B.Th (p = 0.02, Figure 1, Table 1). Compared to controls, the mean B.Th in the TRP of the OA group was 1.9% lower in the RD quadrant, 7.5% lower in the UD quadrant, 4.8% greater in the RV quadrant, and 6.2% greater in the UV quadrant.</div></div><div><h3>CONCLUSION</h3><div>Our results suggest that whole bone TMC microarchitecture may not differ between OA and controls; however, we found significant differences in quadrant bone microarchitecture. This suggests that the MC1 and TRP undergo localized bone microarchitectural changes to adapt to the loading of the TMC joint. Further, our results s","PeriodicalId":74378,"journal":{"name":"Osteoarthritis imaging","volume":"5 ","pages":"Article 100299"},"PeriodicalIF":0.0,"publicationDate":"2025-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144523382","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
引用次数: 0
OPTIMIZED DEEP LEARNING METHOD FOR AUTOMATED SEGMENTATION OF BONE MARROW LESIONS 骨髓病变自动分割的优化深度学习方法
Osteoarthritis imaging
Pub Date : 2025-01-01 DOI: 10.1016/j.ostima.2025.100319
Q. Shihua , W. Qiong , S. Juan , B.D. Jeffrey , M. Timothy , Z. Ming

INTRODUCTION

Bone Marrow Lesions (BMLs), characterized by high-signal intensity on fat-suppressed MRIs, are associated with the progression of knee osteoarthritis (OA). In early OA or when joint damage is not visible on radiographs, BMLs are predictive markers for progression. However, their irregular distribution, potentially large size, and low-contrast boundaries challenge BML segmentation.

OBJECTIVE

This study introduces a novel training strategy for enhancing automated BML segmentation accuracy

METHODS

We aimed to optimize a deep learning method for automatic BML detection and segmentation in MRI, using the Osteoarthritis Initiative (OAI) dataset split into 70% training (210 participants), 15% validation (45 participants), and 15% testing (45 participants), totaling 1025, 190, and 201 MRIs, respectively. Images were employed using data augmentation like brightness, contrast, and geometric transformations. We applied a closing operation, a morphological technique combining dilation and erosion, to smooth edges, addressing the coarse manual labels that impair training. Several models (U-net, SwinUnetR, AttentionUnet, and U-net++) were trained with single-label (BML) and dual-label (BML + femur bone) outputs. Model performance was measured with the Dice Similarity Coefficient (DSC) for overlap and HD95 for boundary error. Cross-entropy and Dice loss functions improved sensitivity during training, particularly in dual-label channels where the femur bone location helped constrain BML positions. We also applied Pixel-Wise Voting (PWV) to improve segmentation stability and accuracy by averaging results from image variations, reducing false positives, and enhancing final segmentation outcomes.

RESULTS

UNet++ model with dual-label (BML + femur bone) yielded the best accuracy, outperforming U-net, SwinUnetR, and AttentionUnet. Figure 1 shows its predicted region (yellow) overlapping well with the manually labeled BML and aligning with boundaries. Specifically, the dual-label Unet model with PWV improved DSC from 62.21% to 64.88% for BML and to 96.52% for bone, while HD95 dropped to 26.82% for BML and 15.52% for bone. SwinUnetR with dual-label and PWV also showed improved DSC (65.06% to 66.70% for BML; 96.34% for bone) and reduced HD95 to 28.31% for BML and 11.54% for bone. AttentionUnet exhibited notable PWV improvements in bone segmentation. Overall, Unet++ achieved the highest performance with dual-label and PWV, increasing DSC from 66.16% to 68.48% for BML and 96.66% for bone, with the lowest HD95 values.

CONCLUSION

This study employed augmentation strategies, a closing operation, and both single- and dual-label analyses to train four models—Unet, SwinUnetR, AttentionUnet, and Unet++. Cross-entropy loss and Pixel-Wise Voting (PWV) enhanced model performance, with dual-label consistently outperforming single-label, es
骨髓病变（BMLs）在脂肪抑制的mri上表现为高信号强度，与膝骨关节炎（OA）的进展有关。在早期骨性关节炎或关节损伤在x线片上不可见时，骨性损伤是病情进展的预测标志。然而，它们的不规则分布、潜在的大尺寸和低对比度边界对BML分割提出了挑战。方法利用骨关节炎倡议（OAI）数据集，将其分为70%的训练（210人）、15%的验证（45人）和15%的测试（45人），共1025、190和201张MRI，旨在优化一种用于MRI中BML自动检测和分割的深度学习方法。图像使用数据增强，如亮度，对比度和几何变换。我们应用闭合操作，一种结合扩张和侵蚀的形态学技术，来平滑边缘，解决粗糙的手工标签，影响训练。使用单标签（BML）和双标签（BML + 股骨骨）输出训练多个模型（U-net、SwinUnetR、AttentionUnet和U-net++ +）。模型性能用Dice Similarity Coefficient （DSC）来衡量重叠，用HD95来衡量边界误差。交叉熵和骰子损失函数在训练过程中提高了灵敏度，特别是在双标签通道中，股骨的位置有助于限制BML的位置。我们还应用了像素明智投票（PWV），通过平均图像变化的结果来提高分割的稳定性和准确性，减少误报，并增强最终的分割结果。结果双标签（BML + 股骨骨）的tsunet++模型准确率最高，优于U-net、SwinUnetR和AttentionUnet。图1显示了它的预测区域（黄色）与手动标记的BML重叠良好，并与边界对齐。其中，带PWV的双标签Unet模型将BML的DSC从62.21%提高到64.88%，骨的DSC提高到96.52%，而HD95模型的BML和骨的DSC分别下降到26.82%和15.52%。双标签和PWV的SwinUnetR也显示出改善的DSC (BML的65.06%至66.70%)；BML和bone的HD95分别降至28.31%和11.54%。AttentionUnet在骨分割方面表现出显著的PWV改善。总的来说，Unet++在双标签和PWV下实现了最高的性能，BML的DSC从66.16%增加到68.48%，骨的DSC从96.66%增加到最低的HD95值。本研究采用增强策略、闭合操作、单标签和双标签分析对Unet、SwinUnetR、AttentionUnet和unet++四个模型进行了训练。交叉熵损失（Cross-entropy loss）和像素明智投票（Pixel-Wise Voting， PWV）增强了模型性能，双标签的表现始终优于单标签，尤其是PWV。我们的发现突出了自动分割作为研究人员强大工具的潜力。
{"title":"OPTIMIZED DEEP LEARNING METHOD FOR AUTOMATED SEGMENTATION OF BONE MARROW LESIONS","authors":"Q. Shihua ,&nbsp;W. Qiong ,&nbsp;S. Juan ,&nbsp;B.D. Jeffrey ,&nbsp;M. Timothy ,&nbsp;Z. Ming","doi":"10.1016/j.ostima.2025.100319","DOIUrl":"10.1016/j.ostima.2025.100319","url":null,"abstract":"<div><h3>INTRODUCTION</h3><div>Bone Marrow Lesions (BMLs), characterized by high-signal intensity on fat-suppressed MRIs, are associated with the progression of knee osteoarthritis (OA). In early OA or when joint damage is not visible on radiographs, BMLs are predictive markers for progression. However, their irregular distribution, potentially large size, and low-contrast boundaries challenge BML segmentation.</div></div><div><h3>OBJECTIVE</h3><div>This study introduces a novel training strategy for enhancing automated BML segmentation accuracy</div></div><div><h3>METHODS</h3><div>We aimed to optimize a deep learning method for automatic BML detection and segmentation in MRI, using the Osteoarthritis Initiative (OAI) dataset split into 70% training (210 participants), 15% validation (45 participants), and 15% testing (45 participants), totaling 1025, 190, and 201 MRIs, respectively. Images were employed using data augmentation like brightness, contrast, and geometric transformations. We applied a closing operation, a morphological technique combining dilation and erosion, to smooth edges, addressing the coarse manual labels that impair training. Several models (U-net, SwinUnetR, AttentionUnet, and U-net++) were trained with single-label (BML) and dual-label (BML + femur bone) outputs. Model performance was measured with the Dice Similarity Coefficient (DSC) for overlap and HD95 for boundary error. Cross-entropy and Dice loss functions improved sensitivity during training, particularly in dual-label channels where the femur bone location helped constrain BML positions. We also applied Pixel-Wise Voting (PWV) to improve segmentation stability and accuracy by averaging results from image variations, reducing false positives, and enhancing final segmentation outcomes.</div></div><div><h3>RESULTS</h3><div>UNet++ model with dual-label (BML + femur bone) yielded the best accuracy, outperforming U-net, SwinUnetR, and AttentionUnet. Figure 1 shows its predicted region (yellow) overlapping well with the manually labeled BML and aligning with boundaries. Specifically, the dual-label Unet model with PWV improved DSC from 62.21% to 64.88% for BML and to 96.52% for bone, while HD95 dropped to 26.82% for BML and 15.52% for bone. SwinUnetR with dual-label and PWV also showed improved DSC (65.06% to 66.70% for BML; 96.34% for bone) and reduced HD95 to 28.31% for BML and 11.54% for bone. AttentionUnet exhibited notable PWV improvements in bone segmentation. Overall, Unet++ achieved the highest performance with dual-label and PWV, increasing DSC from 66.16% to 68.48% for BML and 96.66% for bone, with the lowest HD95 values.</div></div><div><h3>CONCLUSION</h3><div>This study employed augmentation strategies, a closing operation, and both single- and dual-label analyses to train four models—Unet, SwinUnetR, AttentionUnet, and Unet++. Cross-entropy loss and Pixel-Wise Voting (PWV) enhanced model performance, with dual-label consistently outperforming single-label, es","PeriodicalId":74378,"journal":{"name":"Osteoarthritis imaging","volume":"5 ","pages":"Article 100319"},"PeriodicalIF":0.0,"publicationDate":"2025-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144523454","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
引用次数: 0
TOPOGRAPHY OF SEX-RELATED FEMOROTIBIAL CARTILAGE THICKNESS DIFFERENCES: A MATCHED MALE-FEMALE PAIR ANALYSIS CONTROLLING FOR AGE, BMI, AND HEIGHT 与性别相关的股胫软骨厚度差异的地形：一个匹配的男性-女性配对分析，控制年龄，bmi和身高
Osteoarthritis imaging
Pub Date : 2025-01-01 DOI: 10.1016/j.ostima.2025.100350
N. Spoelder , W. Wirth , T.D. Turmezei , F. Eckstein , D.A. Kessler , J.W. Mackay , M. Karperien , S.C. Mastbergen , M.P. Jansen

INTRODUCTION

Knee OA is both more common and progresses faster in women than in men. While it is well known that men exhibit thicker cartilage, it remains unclear whether this difference is inherently sex-based or attributable to confounding factors such as age, BMI, and/or height.

OBJECTIVE

The aim of this study was to evaluate regional differences in knee cartilage thickness between men and women without radiographic OA, who were matched for age, BMI, and height.

METHODS

Participants without radiographic signs of knee OA were selected from the Osteoarthritis Initiative (OAI). Men and women were matched based on height (±1 cm), age (±5 years), and BMI (±2 kg/m²), yielding 63 male-female pairs (n = 126; mean age 57 ± 8 years, BMI 26 ± 4 kg/m², height 170 ± 5 cm). Right knee 3T MRI scans were processed using a deep learning model to generate preliminary automatic segmentations of the outer femoral and tibial contours and the inner cartilage boundaries. These segmentations were manually refined in Stradview and converted into 3D surface models. Cartilage thickness was computed at each vertex as the distance from the cartilage surface to the underlying bone, measured along the normal vector using model-based deconvolution. The femoral, medial tibial, and lateral tibial surfaces and their associated thickness maps were spatially aligned to canonical templates using wxRegSurf. Statistical analyses were performed in MATLAB using the SurfStat package, applying statistical parametric mapping (SPM) with linear mixed models to evaluate paired male-female differences. Significance was set at p < 0.05.

RESULTS

Figure 1 shows the average cartilage thickness in men and women, as well as the differences between sexes. The difference map is predominantly blue, indicating thicker cartilage in men. In both sexes, cartilage was thicker on the lateral side than on the medial side. The trochlea had the greatest thickness overall, with a maximum of 3.98 mm in men and 3.30 mm in women. Statistically significant differences in cartilage thickness between men and women were observed in specific regions of the femur, medial tibia, and lateral tibia (Figure 2). In those regions in the femur, cartilage was thicker in men, with a mean thickness of 2.77 mm compared to 2.42 mm in women, a difference of 0.36 mm (15%). In both the statistically significant different regions of the medial and lateral tibia, cartilage thickness was 0.09 mm (4%) greater in men than in women, with means of 2.26 mm versus 2.17 mm and 2.19 mm versus 2.10 mm, respectively.

CONCLUSION

Despite similar height, age, and BMI, men exhibited thicker femorotibial cartilage than women. Statistically significant differences were found across all three joint surfaces, with the largest difference observed in the trochlea. These findings underscore the need for further research in
膝关节炎在女性中比男性更常见且进展更快。虽然我们都知道男性的软骨更厚，但目前还不清楚这种差异是天生的性别差异，还是由年龄、体重指数和/或身高等混杂因素造成的。目的：本研究的目的是评估年龄、BMI和身高匹配的无骨关节炎的男性和女性膝关节软骨厚度的区域差异。方法从骨关节炎倡议（OAI）中选择无膝关节OA影像学征象的参与者。根据身高（±1 cm）、年龄（±5岁）和BMI（±2 kg/m²）对男女进行配对，共得到63对男女配对(n = 126；平均年龄57±8岁，体重指数26±4 kg/m²，身高170±5 cm)。右膝3T MRI扫描使用深度学习模型进行处理，以生成股骨外侧和胫骨轮廓以及内部软骨边界的初步自动分割。这些分割是在Stradview中手工细化并转换为3D表面模型。在每个顶点处计算软骨厚度，作为软骨表面到下面骨骼的距离，沿着法向量使用基于模型的反褶积进行测量。使用wxRegSurf将股骨、胫骨内侧和胫骨外侧表面及其相关的厚度图在空间上与标准模板对齐。在MATLAB中使用SurfStat软件包进行统计分析，采用统计参数映射（SPM）和线性混合模型来评估成对的男女差异。p <为显著性；0.05.结果图1显示了男性和女性的平均软骨厚度，以及性别之间的差异。差异图以蓝色为主，表明男性的软骨较厚。在两性中，外侧软骨比内侧软骨厚。滑车整体厚度最大，男性最大3.98 mm，女性最大3.30 mm。在股骨、胫骨内侧和胫骨外侧的特定区域，男性和女性的软骨厚度在统计学上有显著差异（图2）。在股骨的这些区域，男性的软骨较厚，平均厚度为2.77 mm，而女性为2.42 mm，差异为0.36 mm（15%）。在胫骨内侧和外侧这两个具有统计学意义的不同区域，男性的软骨厚度比女性大0.09 mm(4%)，分别为2.26 mm比2.17 mm和2.19 mm比2.10 mm。结论：尽管身高、年龄和BMI相似，但男性的股胫软骨比女性厚。在所有三个关节表面上都发现了统计学上显著的差异，其中滑车的差异最大。这些发现强调需要进一步研究股骨胫骨软骨厚度的性别差异，作为女性膝关节OA患病率和严重程度更高的潜在因素。
{"title":"TOPOGRAPHY OF SEX-RELATED FEMOROTIBIAL CARTILAGE THICKNESS DIFFERENCES: A MATCHED MALE-FEMALE PAIR ANALYSIS CONTROLLING FOR AGE, BMI, AND HEIGHT","authors":"N. Spoelder ,&nbsp;W. Wirth ,&nbsp;T.D. Turmezei ,&nbsp;F. Eckstein ,&nbsp;D.A. Kessler ,&nbsp;J.W. Mackay ,&nbsp;M. Karperien ,&nbsp;S.C. Mastbergen ,&nbsp;M.P. Jansen","doi":"10.1016/j.ostima.2025.100350","DOIUrl":"10.1016/j.ostima.2025.100350","url":null,"abstract":"<div><h3>INTRODUCTION</h3><div>Knee OA is both more common and progresses faster in women than in men. While it is well known that men exhibit thicker cartilage, it remains unclear whether this difference is inherently sex-based or attributable to confounding factors such as age, BMI, and/or height.</div></div><div><h3>OBJECTIVE</h3><div>The aim of this study was to evaluate regional differences in knee cartilage thickness between men and women without radiographic OA, who were matched for age, BMI, and height.</div></div><div><h3>METHODS</h3><div>Participants without radiographic signs of knee OA were selected from the Osteoarthritis Initiative (OAI). Men and women were matched based on height (±1 cm), age (±5 years), and BMI (±2 kg/m²), yielding 63 male-female pairs (n = 126; mean age 57 ± 8 years, BMI 26 ± 4 kg/m², height 170 ± 5 cm). Right knee 3T MRI scans were processed using a deep learning model to generate preliminary automatic segmentations of the outer femoral and tibial contours and the inner cartilage boundaries. These segmentations were manually refined in Stradview and converted into 3D surface models. Cartilage thickness was computed at each vertex as the distance from the cartilage surface to the underlying bone, measured along the normal vector using model-based deconvolution. The femoral, medial tibial, and lateral tibial surfaces and their associated thickness maps were spatially aligned to canonical templates using wxRegSurf. Statistical analyses were performed in MATLAB using the SurfStat package, applying statistical parametric mapping (SPM) with linear mixed models to evaluate paired male-female differences. Significance was set at p &lt; 0.05.</div></div><div><h3>RESULTS</h3><div>Figure 1 shows the average cartilage thickness in men and women, as well as the differences between sexes. The difference map is predominantly blue, indicating thicker cartilage in men. In both sexes, cartilage was thicker on the lateral side than on the medial side. The trochlea had the greatest thickness overall, with a maximum of 3.98 mm in men and 3.30 mm in women. Statistically significant differences in cartilage thickness between men and women were observed in specific regions of the femur, medial tibia, and lateral tibia (Figure 2). In those regions in the femur, cartilage was thicker in men, with a mean thickness of 2.77 mm compared to 2.42 mm in women, a difference of 0.36 mm (15%). In both the statistically significant different regions of the medial and lateral tibia, cartilage thickness was 0.09 mm (4%) greater in men than in women, with means of 2.26 mm versus 2.17 mm and 2.19 mm versus 2.10 mm, respectively.</div></div><div><h3>CONCLUSION</h3><div>Despite similar height, age, and BMI, men exhibited thicker femorotibial cartilage than women. Statistically significant differences were found across all three joint surfaces, with the largest difference observed in the trochlea. These findings underscore the need for further research in","PeriodicalId":74378,"journal":{"name":"Osteoarthritis imaging","volume":"5 ","pages":"Article 100350"},"PeriodicalIF":0.0,"publicationDate":"2025-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144523628","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
引用次数: 0
首页 上一页
类型
全部 化学•材料 生命科学 医学 物理 工程技术 环境•农林 材料科学 地球科学 法学 管理学 化学 环境科学与生态学 计算机科学 教育学 经济学 农林科学 人文科学 生物学 数学 物理与天体物理 心理学 综合性期刊 其他 工业工程 理学 历史学 农学 文学 信息工程
数据库
全部 ACS Publications Elsevier ieeexplore Springer The Royal Society of Chemistry Wiley
期刊
Osteoarthritis imaging
全部 Acc. Chem. Res. ACS Applied Bio Materials ACS Appl. Electron. Mater. ACS Appl. Energy Mater. ACS Appl. Mater. Interfaces ACS Appl. Nano Mater. ACS Appl. Polym. Mater. ACS BIOMATER-SCI ENG ACS Catal. ACS Cent. Sci. ACS Chem. Biol. ACS Chemical Health & Safety ACS Chem. Neurosci. ACS Comb. Sci. ACS Earth Space Chem. ACS Energy Lett. ACS Infect. Dis. ACS Macro Lett. ACS Mater. Lett. ACS Med. Chem. Lett. ACS Nano ACS Omega ACS Photonics ACS Sens. ACS Sustainable Chem. Eng. ACS Synth. Biol. Anal. Chem. BIOCHEMISTRY-US Bioconjugate Chem. BIOMACROMOLECULES Chem. Res. Toxicol. Chem. Rev. Chem. Mater. CRYST GROWTH DES ENERG FUEL Environ. Sci. Technol. Environ. Sci. Technol. Lett. Eur. J. Inorg. Chem. IND ENG CHEM RES Inorg. Chem. J. Agric. Food. Chem. J. Chem. Eng. Data J. Chem. Educ. J. Chem. Inf. Model. J. Chem. Theory Comput. J. Med. Chem. J. Nat. Prod. J PROTEOME RES J. Am. Chem. Soc. LANGMUIR MACROMOLECULES Mol. Pharmaceutics Nano Lett. Org. Lett. ORG PROCESS RES DEV ORGANOMETALLICS J. Org. Chem. J. Phys. Chem. J. Phys. Chem. A J. Phys. Chem. B J. Phys. Chem. C J. Phys. Chem. Lett. Analyst Anal. Methods Biomater. Sci. Catal. Sci. Technol. Chem. Commun. Chem. Soc. Rev. CHEM EDUC RES PRACT CRYSTENGCOMM Dalton Trans. Energy Environ. Sci. ENVIRON SCI-NANO ENVIRON SCI-PROC IMP ENVIRON SCI-WAT RES Faraday Discuss. Food Funct. Green Chem. Inorg. Chem. Front. Integr. Biol. J. Anal. At. Spectrom. J. Mater. Chem. A J. Mater. Chem. B J. Mater. Chem. C Lab Chip Mater. Chem. Front. Mater. Horiz. MEDCHEMCOMM Metallomics Mol. Biosyst. Mol. Syst. Des. Eng. Nanoscale Nanoscale Horiz. Nat. Prod. Rep. New J. Chem. Org. Biomol. Chem. Org. Chem. Front. PHOTOCH PHOTOBIO SCI PCCP Polym. Chem.
﹀
×
引用
GB/T 7714-2015
复制
MLA
复制
APA
复制
导出至
BibTeX EndNote RefMan NoteFirst NoteExpress
×
0
微信
客服QQ
Book学术公众号 扫码关注我们
反馈
×
意见反馈
请填写您的意见或建议
请填写您的手机或邮箱
×
提示
您的信息不完整，为了账户安全，请先补充。
现在去补充
×
提示
您因"违规操作"
具体请查看互助需知
我知道了
×
提示
确定
Book学术官方微信
Book学术文献互助
Book学术文献互助群
群 号：604180095
Book学术
文献互助 智能选刊 最新文献 互助须知 联系我们：info@booksci.cn
Book学术提供免费学术资源搜索服务，方便国内外学者检索中英文文献。致力于提供最便捷和优质的服务体验。
Copyright © 2023 Book学术 All rights reserved.
ghs 京公网安备 11010802042870号 京ICP备2023020795号-1