Background: Bone health and brain function may be closely interconnected through a complex bone–brain axis. The relationship between bone mineral density (BMD), vertebral perfusion, marrow composition, cerebral perfusion, brain volume, and cognitive decline, however, remain incompletely understood. Methods: Ninety-nine female subjects (mean age 65.00 ± 5.00 years) with clinically suspected mild cognitive impairment underwent dual-energy X-ray absorptiometry, carotid ultrasound, and multimodal magnetic resonance imaging (MRI) of the brain and lumbar spine to measure BMD, bone perfusion, marrow fat content as well as cerebral perfusion, cerebral volume, cerebral white matter burden and large vessel atherosclerosis. Cognitive function was assessed using the Hong Kong Montreal Cognitive Assessment (HK-MoCA). Bone, cerebral, vascular, and cognitive measures were correlated using Spearman correlation coefficients and compared in group comparisons. Results: Lower BMD was correlated with reduced subcortical cerebral blood flow (CBF) (r = 0.27, p = 0.031) and lower total brain parenchymal volume (r = 0.25, p = 0.021). Reduced bone marrow perfusion and increased marrow fat content were also associated with lower total brain parenchymal volume (r = 0.24, p = 0.023 and r = −0.26, p = 0.025). Subjects with the lowest L3 vertebral body perfusion or highest marrow fat content had significantly reduced total brain and hippocampal volumes (p = 0.029–0.049) compared with those with the highest perfusion or lowest marrow fat content. Conclusions: This study shows an association between lower BMD, reduced vertebral perfusion, and increased marrow fat with reduced brain parenchymal volumes and reduced brain perfusion. Further studies are warranted to clarify these relationships and explore the underlying shared mechanisms affecting bone health and cerebral microvascular and structural brain changes.

Joint space narrowing (JSN) in rheumatoid arthritis (RA) can progress even during clinical remission. Conventional imaging lacks sensitivity for detecting early risks. This study evaluates AI-based analysis of bilateral joint-space asymmetry on hand radiographs to detect future JSN progression in radiographically normal RA. Forty-six radiographically normal RA patients underwent bilateral hand radiography and high-resolution peripheral quantitative computed tomography (HR-pQCT) at baseline, 12, and 24 months. Five joint space parameters were extracted from HR-pQCT. Gaussian mixture modeling (GMM) stratified patients into progressive and non-progressive phenotypes using 24 month data. An AI pipeline was developed to measure bilateral joint space asymmetry from baseline X-rays using deep learning–based segmentation and registration. Predictive models were constructed using logistic regression and evaluated via ROC curves. GMM identified 17 progressive and 29 non-progressive patients. HR-pQCT-derived intra-joint space width inconsistency (JSW.Inc) increased significantly over time in progressors (p = 0.0337). At baseline, HR-pQCT-based JSW.Inc showed fair predictive power (AUC = 0.696). AI-derived asymmetry at PIP3 and MCP3 joints showed stronger baseline prediction (AUCs = 0.739 and 0.714, respectively). By month 12, predictive performance improved across models (AUCs up to 0.836 for HR-pQCT and 0.748 for AI-based metrics). AI-assisted analysis of bilateral joint space asymmetry from hand radiography may provide an accessible method for detecting radiographic JSN progression, with prognostic findings that should be interpreted as exploratory and hypothesis-generating. This method complements conventional imaging tools and may support future investigations into early risk stratification.

Background: Rheumatoid arthritis is a chronic inflammatory disease with progressive joint destruction resulting in functional loss. High-resolution peripheral quantitative CT (HR-pQCT) is a novel technique for detailed volumetric assessment of joint space morphology and bone erosions. We hypothesized that effective control of inflammation in early rheumatoid arthritis (ERA) patients who can achieve sustained simple disease activity index (SDAI) remission will have less progression of joint damage and functional loss than patients who cannot achieve sustained SDAI remission (SDI). Objectives: To elucidate the effects of achieving SDI on the progression of joint space outcomes in patients with ERA assessed using HR-pQCT. Design: This was a prospective study with 109 participants receiving 1-year tight-control treatment aiming at SDAI remission. Methods: A total of 109 patients received 1-year tight-control treatment aiming at SDAI remission. SDI was defined as achieving SDAI remission at months 6, 9, and 12. The primary outcome was the change in joint space size and volume at metacarpophalangeal joints 2–4, determined using HR-pQCT at 12 months. Results: At 12 months, HR-pQCT image analysis showed that 14 out of 109 (12.8%) patients achieved SDI (SDI group). No significant differences in changes in joint space and erosion parameters were observed between the SDI and non-SDI groups. Erosion volume was reduced significantly in both groups. While new erosions (0% vs 16%, p = 0.338) and erosion progression (0% vs 27.3%, p = 0.592) were only observed in the non-SDI group, partial erosion healing was numerically more frequently observed in the SDI group (27.3% vs 10.3%, p = 0.260). At baseline, disease activity parameters were positively correlated with joint space parameters, while after 12 months, a negative correlation emerged between disease activity parameters and both mean joint space width (JSW) and minimum JSW. Conclusion: Although HR‑pQCT did not show a significant difference in joint damage progression at 12 months between patients with and without sustained remission, the overall cohort—managed with a tight-control strategy targeting SDAI remission—exhibited minimal bone and joint damage progression over 1 year.

Objectives To distinguish pathological from nonpathological bone erosions in rheumatoid arthritis (RA) by comparing erosion prevalence and size with healthy controls (HCs) at baseline, as well as long-term evolution and treatment response using high-resolution peripheral quantitative computed tomography (HR-pQCT). Methods We enrolled 247 patients with RA and 78 age- and sex-matched HCs for baseline HR-pQCT of the second metacarpal head. Patients with RA from 2 cohorts—early RA (ERA; n = 98; symptom duration ≤2 years; treated-to-target [T2T] in year 1) and established RA (n = 149; received usual care)—were rescanned after a median of 8 years to quantify erosion volumetric changes. Results At baseline, small erosions (< 1 mm³) were similarly prevalent in RA and HCs, whereas intermediate (1-5 mm³) and large (> 5 mm³) erosions were more frequent in RA (P < .003). Over 8 years, small erosions in RA were highly stable with minimal volumetric change, whereas large erosions decreased substantially in volume. In patients with large erosions, logistic regression indicated that being in the ERA cohort independently predicted regression of large erosion (odds ratio: 6.93; 95% CI: 1.57-30.56; P = .011). Conclusions Small erosions are common in both RA and healthy individuals and remain stable long term in RA, supporting their interpretation as physiological/biomechanical (noninflammatory) variants rather than RA pathology. In contrast, large erosions are specific to RA and show significant capacity for repair under early, aggressive T2T management.

Background Generalized knee tissue segmentation, such as cartilage and meniscus in magnetic resonance imaging (MRI), plays a vital role in the clinical assessment of knee osteoarthritis (OA). However, domain variability between MRI datasets poses a significant challenge for the application of robust segmentation methods in real-world clinical settings. Existing unsupervised domain adaptation (UDA) approaches, which rely on one-to-one assumptions between the source and target domains, often fail to preserve knee tissues such as cartilage and meniscus, which are critical for OA diagnosis in diverse clinical settings. Methods We propose a source-independent segmentation approach tailored for multi-domain knee MRI datasets. Our method emphasizes knee tissue regions to reduce domain gaps and label inconsistencies. By introducing a stepwise adaptation strategy, segmentation performance was refined progressively from intermediate domains to the final target domain. Pseudo-label attention mechanisms were integrated into the adaptation pipeline, enabling iterative fine-tuning of domain-specific segmentations while leveraging unidirectional generative adversarial networks to enhance tissue-specific adaptation. This iterative training process ensures the generation of reliable pseudo-labels, thereby improving segmentation accuracy in diverse clinical MRI datasets. Results We demonstrated the effectiveness of our approach on the OA initiative dataset as the source domain and self-collected, T1-weighted fast field echo (T1FFE) as the intermediate domain and three-dimensional fast spin echo (3D FSE) as the final target domain. Our method achieved an average dice scores of 0.8701 and 0.7990 for source and target domains, respectively, surpassing the typical UDA methods explored in our experiments. Conclusion The experiments conducted on clinical MRI data, spanning OA severity from healthy knees to KL Grades 1-4, validated the effectiveness of the proposed domain adaptation method in precise segmentation of the cartilage and meniscus.

Most mistakes relate to either a lesion being missed or when a lesion is seen a wrong diagnosis being made. Clinicians and patients understandably want an accurate diagnosis. An accurate diagnosis, however, is not always possible given the complexities and inherent limitations of imaging. There is an inevitable trade off between committing to trying to provide a definitive diagnosis and occasionally getting it wrong. Better, and more useful, to being occasionally wrong though than habitually noncommittal.

Background: To differentiate pathological bone erosions from non-pathological in rheumatoid arthritis (RA) by comparing erosion prevalence and size against healthy controls (HCs) at baseline and evaluating their long-term evolution and response to therapy using high-resolution peripheral quantitative computed tomography (HR-pQCT). Methods: A total of 247 RA patients and 78 age- and sex-matched HCs underwent baseline HR-pQCT scans of the second metacarpal head. RA patients, consisting of two cohorts (early RA [ERA]; n=98, with symptom duration [Formula: see text]2 years and treat-to-target [T2T] within the first year, and established RA [EstRA]; n=149, receiving usual care), were reassessed after a median follow-up of 8 years to track volumetric changes in erosions. Results: At baseline, the prevalence of small erosions (<1 mm3) was comparable between RA patients and HCs. In contrast, intermediate (1–5 mm3) and large erosions (>5 mm3) were significantly more common in the RA cohort (P<0.003). Over the 8-year follow-up in RA patients, small erosions demonstrated remarkable stability with minimal change in size, while large erosions demonstrated a significant reduction in volume. In a subgroup of patients with large erosions, multivariable logistic regression analysis revealed that ERA patients had a significantly higher likelihood of regression of large erosions compared to EstRA patients (OR 6.93, 95% CI 1.57-30.56, p = 0.011). Conclusion: Small erosions are common in both RA patients and healthy individuals and remain stable over the long term, suggesting they are likely physiological variants rather than pathological features of RA. Conversely, large erosions are pathological but demonstrate a significant healing potential, especially when early and aggressive T2T therapy is initiated.

Injuries of the posteromedial meniscocapsular junction encompass a spectrum of injuries at or near the posteromedial meniscocapsular junction. These so-called ramp lesions have become increasingly well recognized, defined, and understood. Such injuries range in severity from perimeniscal edema-like signal to tears at or near the meniscocapsular junction, with or without meniscocapsular separation. Since the term's introduction in 1998, the definition of a ramp lesion has evolved to include not only tears of the perimeniscal attachments (i.e., the meniscocapsular and meniscotibial attachments) but also tears of the peripheral third of the posterior horn of the medial meniscus. Radiologists should seek to accurately recognize the presence of a ramp lesion on MRI, to help guide orthopedic surgeons in inspecting the posteromedial meniscocapsular region at the time of arthroscopy. This Special Series Review provides an imaging-focused description of ramp lesions of the knee. The article presents pertinent anatomy of the posteromedial meniscocapsular junction, pathogenesis underlying the occurrence of ramp lesions, criteria for diagnosing these lesions on MRI, potential pitfalls, and observed injury patterns. The goals of treatment are additionally discussed, highlighting the rationale behind the use of different treatment approaches for stable and unstable ramp lesions.

Background: How different gender-specific bone mineral density cutpoint T-scores are associated with different hip fragility fracture (FFx) prediction sensitivity has not been well studied. This article presents an updated analysis of hip FFx prediction among older people by a dual-energy X-ray absorptiometry (DXA) measure, using literature results and our own Chinese data. Methods: We systematically searched literature reports on DXA T-score results measured at the timepoint of a hip FFx. With osteoporotic fractures in women (MsOS) and in men (MrOS) Hong Kong studies, at baseline 2,000 Chinese women (mean: 72.5 years) and 2,000 Chinese men (mean: 72.3 years) were recruited. Female participants were followed up for 8.8 +/- 1.5 years, and 69 FFx were recorded. Male participants were followed up for 9.9 +/- 2.8 years, and 63 hip FFx were recorded. Results: Ten articles published femoral neck (FN) and/or total hip (TH) T-score at the timepoint of a hip FFx with separated females' or males' T-score data. We estimated that, if a DXA exam were taken shortly before the FFx accident, females' FN, females' TH, males' FN, or males' TH T-scores on average predicted 66.9%, 70.4%, 66.5%, and 67.8% of the hip FFx. For the MsOS and MrOS Hong Kong studies, a combination of baseline FN and TH T-score predicted >50% of the cases with a follow-up hip FFx. A combination of baseline FN T-score, TH T-score, lumbar spine T-score, and spine fracture-like deformity assessment predicted 68.1% of the female cases with a follow-up hip FFx, and 63.4% of the male cases with a follow-up hip FFx. Conclusions: If a DXA scan is regularly performed, approximately 70% of the hip FFx incidents can be predicted for older women and men.

Background: Rheumatoid arthritis (RA) features joint inflammation and bone erosions that impair function. High-resolution peripheral quantitative computed tomography (HR-pQCT) detects very early erosions not visible using radiography, enabling detailed monitoring of structural change in early RA (ERA). The baseline total erosion score assessed by HR-pQCT at two MCP joints, is known to correlate with the Health Assessment Questionnaire (HAQ) score1. Whether erosive progression/regression detected on HR-pQCT impacts HAQ score at medium-term needs to be assessed. We aimed to determine whether HR-pQCT-detected changes in erosion predict medium-term changes in functional status in ERA patients. Methods: We recruited 101 patients with ERA (symptom duration < 2 years). HR-pQCT scans were obtained at baseline and after 2 years. Erosions were assessed at the second to fourth metacarpophalangeal joints (MCPJ 2-4) of the most affected hand or dominant hand when both hands are equally affected. Repair was defined as either a decrease in erosion counts by [Formula: see text] 1 or a reduction in total erosion volume exceeding least significant change (LSC). Progression was defined as either an increase in erosion counts by [Formula: see text] 1 or an increase in total erosion volume exceeding the LSC. Results: Of the 101 patients at baseline, 78 (77%) were women, with a mean age of 57.2 ± 12.3 years. Disease activity and functional status (assessed by HAQ) improved significantly after 2 years of protocolized treatment. The proportion receiving b/tsDMARDs increased from 0.9% at baseline to 23.8% at month 24 (p<0.001). Paired HR-pQCT analyses at baseline and month 24 were available for 98 patients. Thirty-seven (37.8%) patients had no erosion at both time points. Among the 61 patients who had erosions at either baseline or month 24, 16 (26.2%) achieved partial repair (repair group), while 38 (62.3%) showed progression. In the repair group, change in total erosion volume correlated positively with improvement in HAQ score ([Formula: see text]= 0.719, p = 0.003; Figure 1). This association remained significant after adjusting for change in disease activity (p = 0.009; 95% CI 0.178–1.047). Conclusion: Bone erosion repair detected by HR-pQCT may independently contribute to functional improvement in patients with early RA.

Objective: To determine whether wrist bone erosive burden assessed by high-resolution peripheral quantitative computed tomography (HR-pQCT) is associated with functional disability in patients with rheumatoid arthritis (RA). Methods: A total of 232 RA patients who underwent wrist HR-pQCT imaging were enrolled. Erosive burden was quantified by the number and volume of erosions in the distal radius, lunate, and scaphoid. Functional outcomes were assessed using the Health Assessment Questionnaire Disability Index (HAQ-DI), with disability stratified according to HAQ-DI scores. The relationship between wrist bone erosion burden and both the presence and severity of disability was evaluated. Results: Analysis was performed on 232 patients (mean age 62 ± 10 years, 80% female) with evaluable HR-pQCT scans. Carpal erosions were detected in 88% of patients, most frequently in the lunate (80%), followed by the scaphoid (53%) and radius (42%). Both the number and volume of erosions in the radius and lunate significantly correlated with functional disability ([Formula: see text]up to 0.21, P<0.01). Multivariate ordinal analysis adjusting for potential confounders identified the presence of radial erosions (OR[Formula: see text]2.16, 95% CI 1.26–3.71, P [Formula: see text] 0.005), erosion counts (OR[Formula: see text]1.20-1.41, p<0.001), and volume (OR[Formula: see text]1.01-1.02, P [Formula: see text] 0.001) in the wrist bones were independently associated with higher disability grades in RA patients. Additionally, wrist joint destruction independently conferred threefold higher odds of higher disability grades in RA patients (p[Formula: see text]0.021). Conclusion: HR-pQCT-detected radial erosions and higher erosion numbers and volume in the wrist bones were strongly associated with functional disability in RA patients, independent of disease activity, disease duration, and treatment regimen.

Ultrasound is as accurate as MRI in the detection of most brachial pathologies but tends to be underutilized in clinical practice compared to MRI. The main reason for this under-usage is a relative lack of knowledge regarding how to perform brachial plexus ultrasound and a lack of awareness of the ultrasound appearances of brachial pathologies. This review serves to re-address this imbalance by providing a practical overview on how to perform brachial plexus ultrasound as well as highlighting the ultrasound appearances of common pathologies likely to be encountered in everyday clinical practice.

Clinicians should consider baseline BMD, age, and sex when assessing osteoporosis risk. Women over 65 with severe osteopenia need close monitoring. Transition rates reached 10.7

PURPOSE:To propose and evaluate an accelerated T 1 ρ $$ {T}_{1\rho } $$ quantification method that combines T 1 ρ $$ {T}_{1\rho } $$ -weighted fast spin echo (FSE) images and proton density (PD)-weighted anatomical FSE images, leveraging deep learning models for T 1 ρ $$ {T}_{1\rho } $$ mapping. The goal is to reduce scan time and facilitate integration into routine clinical workflows for osteoarthritis (OA) assessment. METHODS:This retrospective study utilized MRI data from 40 participants (30 OA patients and 10 healthy volunteers). A volume of PD-weighted anatomical FSE images and a volume of T 1 ρ $$ {T}_{1\rho } $$ -weighted images acquired at a non-zero spin-lock time were used as input to train deep learning models, including a 2D U-Net and a multi-layer perceptron (MLP). T 1 ρ $$ {T}_{1\rho } $$ maps generated by these models were compared with ground truth maps derived from a traditional non-linear least squares (NLLS) fitting method using four T 1 ρ $$ {T}_{1\rho } $$ -weighted images. Evaluation metrics included mean absolute error (MAE), mean absolute percentage error (MAPE), regional error (RE), and regional percentage error (RPE). RESULTS:The best-performed deep learning models achieved RPEs below 5% across all evaluated scenarios. This performance was consistent even in reduced acquisition settings that included only one PD-weighted image and one T 1 ρ $$ {T}_{1\rho } $$ -weighted image, where NLLS methods cannot be applied. Furthermore, the results were comparable to those obtained with NLLS when longer acquisitions with four T 1 ρ $$ {T}_{1\rho } $$ -weighted images were used. CONCLUSION:The proposed approach enables efficient T 1 ρ $$ {T}_{1\rho } $$ mapping using PD-weighted anatomical images, reducing scan time while maintaining clinical standards. This method has the potential to facilitate the integration of quantitative MRI techniques into routine clinical practice, benefiting OA diagnosis and monitoring.

Purpose To propose and evaluate an accelerated T-1 rho quantification method that combines T-1 rho -weighted fast spin echo (FSE) images and proton density (PD)-weighted anatomical FSE images, leveraging deep learning models for T-1 rho mapping. The goal is to reduce scan time and facilitate integration into routine clinical workflows for osteoarthritis (OA) assessment. Methods This retrospective study utilized MRI data from 40 participants (30 OA patients and 10 healthy volunteers). A volume of PD-weighted anatomical FSE images and a volume of T-1 rho -weighted images acquired at a non-zero spin-lock time were used as input to train deep learning models, including a 2D U-Net and a multi-layer perceptron (MLP). T-1 rho maps generated by these models were compared with ground truth maps derived from a traditional non-linear least squares (NLLS) fitting method using four T-1 rho -weighted images. Evaluation metrics included mean absolute error (MAE), mean absolute percentage error (MAPE), regional error (RE), and regional percentage error (RPE). Results The best-performed deep learning models achieved RPEs below 5% across all evaluated scenarios. This performance was consistent even in reduced acquisition settings that included only one PD-weighted image and one T-1 rho -weighted image, where NLLS methods cannot be applied. Furthermore, the results were comparable to those obtained with NLLS when longer acquisitions with four T-1 rho -weighted images were used. Conclusion The proposed approach enables efficient T-1 rho mapping using PD-weighted anatomical images, reducing scan time while maintaining clinical standards. This method has the potential to facilitate the integration of quantitative MRI techniques into routine clinical practice, benefiting OA diagnosis and monitoring.

Background: To identify contributors to non-traumatic incident fractures on rheumatic disease patients who are on long term glucocorticoids (LTGC) Methods: Two hundred and twenty patients on LTGC (110 with vertebral fracture and 110 without vertebral fracture) who participated in a cross-sectional study in 2014-2015 were invited to have repeated assessments on 1) aBMD using dual-energy X-ray absorptiometry (DXA), 2) volumetric BMD (vBMD), microstructure and bone strength assessment of the wrist and tibia using high-resolution peripheral quantitative computed tomography (HR-pQCT) and 3) spine radiographs in the 5th year. Clinical covariates were recorded on questionnaires, and Fracture Risk Assessment Tool (FRAX) score was calculated accordingly. Non-traumatic incident fracture over the 5-year were documented. Receiver operating characteristic curve (ROC) analysis was performed to compare the strength of fracture prediction tools. Results: Out of the 140 patients who completed the 5th year assessments, 47 (33.6%) developed incident fractures. History of previous fracture, aBMD at hip and lumbar spine, T-score, trabecular vBMD, trabecular bone volume fraction and estimated bone strength at the tibia at baseline remained significantly different after adjusting for age between the group with and without incident fractures. The area under curve (AUC) of a prediction model comprised of age, history of previous fracture and average trabecular vBMD at tibia was comparable with that of the FRAX score (0.710 vs 0.679-0.702), and slightly outperformed than the AUC of DXA aBMD at hip and lumbar spine (0.628-0.668) under ROC analysis. Conclusion: Age, history of previous fracture and average trabecular vBMD at tibia could be the main contributors in building a prediction model for non-traumatic incident fracture in rheumatic disease patients on LTGC.

Abstract Background Frailty is common in patients undergoing cardiac surgery and is associated with poorer postoperative outcomes. Ultrasound examination of skeletal muscle morphology may serve as an objective assessment tool as lean muscle mass reduction is a key feature of frailty. Methods This study investigated the association of ultrasound-derived muscle thickness, cross-sectional area, and echogenicity of the rectus femoris muscle (RFM) with preoperative frailty and predicted subsequent poor recovery after surgery. Eighty-five patients received preoperative RFM ultrasound examination and frailty-related assessments: Clinical Frailty Scale (CFS) and 5-m gait speed test (GST5m). Association of each ultrasound measurement with frailty assessments was examined. Area under receiver-operating characteristic curve (AUROC) was used to assess the discriminative ability of each ultrasound measurement to predict days at home within 30 days of surgery (DAH30). Results By CFS and GST5m criteria, 13% and 34% respectively of participants were frail. RFM cross-sectional area alone demonstrated moderate predictive association for frailty by CFS criterion (AUROC: 0.76, 95% CI: 0.66–0.85). Specificity improved to 98.7% (95% CI: 93.6%-100.0%) by utilising RFM cross-sectional area as an ‘add-on’ test to a positive gait speed test, and thus a combined muscle size and function test demonstrated higher predictive performance (positive likelihood ratio: 40.4, 95% CI: 5.3–304.3) for frailty by CFS criterion than either test alone (p < 0.001). The combined ‘add-on’ test predictive performance for DAH30 (AUROC: 0.90, 95% CI: 0.81–0.95) may also be superior to either CFS or gait speed test alone. Conclusions Preoperative RFM ultrasound examination, especially when integrated with the gait speed test, may be useful to identify patients at high risk of frailty and those with poor outcomes after cardiac surgery. Trial registration The study was registered on the Chinese Clinical Trials Registry (ChiCTR2000031098) on 22 March 2020.