Musculoskeletal disorders are characterized by diverse pathological types and complex anatomical structures, placing high demands on the precision of imaging examinations. In recent years, artificial intelligence (AI) has shown great potential in musculoskeletal imaging, particularly in anatomical segmentation, lesion detection, quantitative measurement, and intelligent diagnosis. This review systematically summarizes advances in AI applications for degenerative joint diseases, sports injuries, fracture detection, osteoporosis screening, and musculoskeletal tumors, while also outlining its expanding roles in image reconstruction, quality control, and educational support. Furthermore, it highlights the developmental trends of large AI models in musculoskeletal imaging and discusses their potential and challenges in multimodal, multitask, and personalized clinical decision support. Although AI has reached or even surpassed expert-level performance in certain tasks, limitations remain in model generalization, data acquisition and annotation standards, cross-modality integration, and clinical adaptability. Future progress will require high-quality data construction, interdisciplinary collaboration, and the establishment of standardized frameworks to advance musculoskeletal AI toward more intelligent, efficient, and standardized clinical practice.

Objective To evaluate the prognostic value of CT—derived fractional flow reserve (CT-FFR) for predicting major adverse cardiovascular events (MACE) in diabetic patients with coronary artery disease (CAD). Methods A total of 876 CAD patients with coronary stenosis ranging from 25% to 80%, who underwent coronary CT angiography (CCTA) and had complete follow-up data, were prospectively enrolled. The median follow-up time was 2.9 years. Among them, 224 were diabetic and 652 were non-diabetic. Baseline clinical data were recorded, coronary stenosis severity was quantitatively assessed on CCTA, and CT-derived fractional flow reserve (CT-FFR) were measured for hemodynamic assessment. The primary clinical endpoint was MACE, including all-cause death, non-fatal myocardial infarction, and unplanned revascularization. The survival probabilities of diabetic patients and non-diabetic patients were analyzed respectively by the Kaplan-Meier method, and the survival differences between the two groups with CT-FFR≤0.80 and CT-FFR>0.80 were compared by the log-rank test. Univariate and multivariate Cox regression analyses were performed to identify independent predictors of MACE in diabetic and non-diabetic CAD patients. Results Compared with the non-diabetic group, the diabetic group had higher rates of hypertension, more patients with coronary stenosis severity≥50%, and higher levels of fasting blood glucose, systolic and diastolic blood pressure, and triglycerides；conversely, high-density lipoprotein, total bilirubin, and direct bilirubin levels were lower in the diabetic group (all P<0.05). Kaplan-Meier curve analysis showed that in both diabetic and non-diabetic patients, the survival probability of CT-FFR≤0.80 group was lower than CT-FFR>0.80 group (both P<0.05). Univariate and multivariate Cox regression analysis identified CT-FFR≤0.80 as an independent predictor of MACE in diabetic patients, whereas hyperlipidemia, coronary stenosis severity≥50%, and CT-FFR≤0.80 were independent predictors of MACE in non-diabetic patients. Conclusion CT-FFR can predict the 3-year occurrence of MACE in diabetic patients with CAD.

Objective To investigate local cardiac functional changes in individuals infected with human immunodeficiency virus (HIV) using cardiac magnetic resonance feature tracking (CMR-FT) technology. Methods A total of 47 HIV-infected patients were prospectively enrolled, including 30 asymptomatic individuals (Asymptomatic HIV group), 17 patients with acquired immunodeficiency syndrome (AIDS group), and 21 age-matched healthy controls (HC group). Overall cardiac function was evaluated using CMR cine sequences, and myocardial strain parameters were measured with CMR-FT technology. The left ventricle was divided into 16 segments, and segmental strain reduction was defined using thresholds of global longitudinal strain (GLS)=-13.4%, global circumferential strain (GCS)=-16.8%, and global radial strain (GRS)=43.9%. The proportion of segments with reduced strain was compared across the three groups. Differences in clinical and imaging data among the three groups were compared. Differences in clinical information between the asymptomatic HIV group and the AIDS group were compared using the chi-square test or Fisher's exact test. Results For global cardiac function parameters, both the asymptomatic HIV group and the AIDS group had significantly lower left ventricular ejection fraction (LVEF), right ventricular ejection fraction (RVEF), right ventricular end-systolic volume (RVESV), and right ventricular cardiac output (RVCO) compared to the HC group (all P<0.05). However, there were no significant differences in overall cardiac function parameters between the asymptomatic HIV and AIDS groups(all P>0.05). Regarding myocardial strain parameters, GLS, GCS, and GRS were significantly lower in both the asymptomatic HIV group and the AIDS group compared to the HC group (all P<0.05), with no statistically significant differences between the asymptomatic HIV and AIDS groups for these parameters (all P>0.05). In the analysis of segmental GCS, the AIDS group showed a higher proportion of segments with reduced GCS compared to the asymptomatic HIV group, while no significant differences were found in the segmental strain parameters of other types (all P>0.05). Conclusion GCS assessed by CMR-FT technology can serve as an imaging biomarker for early cardiac function impairment following HIV infection, providing valuable information for clinical assessment.

Objective To explore the diagnostic value of contrast-enhanced cardiac CT for detecting left atrial appendage (LAA) thrombus in patients with atrial fibrillation (AF). Methods All of 116 patients with AF were prospectively selected. All patients received contrast-enhanced cardiac CT and transesophageal echocardiography (TEE). Based on the TEE findings, the patients were divided into the LAA thrombus group (45 cases) and the LAA non-thrombus group (71 cases). Using TEE as the gold standard, the diagnostic performance of contrast-enhanced cardiac CT for LAA thrombus was assessed. The Kappa test was used to analyze the agreement between CT and TEE diagnosis. The Chi-square test was applied to compare the accuracy of first- and second-phase CT scans, and the t test was used to compare the structural and volumetric indexes of left atrium and LAA between the two groups. Multivariate Logistic regression analysis was applied to analyze the independent risk factors for LAA thrombus. Results For diagnosing LAA thrombus in AF patients, the first-phase enhanced CT showed a sensitivity of 91.11% and a specificity of 88.73%, with excellent diagnostic consistency (κ=0.786); the second-phase scan showed a sensitivity of 95.56% and a specificity of 92.96%, with excellent diagnostic consistency (κ=0.874).Compared with the non-thrombus group, patients in the thrombus group had significantly larger maximum left atrial volume (LAVmax), minimum left atrial volume (LAV_min), maximum and minimum left atrial anteroposterior diameters (LAD_max, LAD_min), LAA volume (LAAV), and LAA orifice dimensions (long diameter, short diameter, and vertical diameter) (all P<0.05), whereas LAA ejection fraction (LAA-EF) was significantly lower (P<0.05). Multivariate Logistic regression analysis showed that increased LAAV and long orifice diameter were independent risk factors for LAA thrombus in AF patients (both P<0.05), while increased LAA-EF was a protective factor (P<0.05). Conclusion Contrast-enhanced cardiac CT examination have high diagnostic value for detecting LAA thrombus in AF patients. Structural and functional parameters of the atrium and LAA can provide additional insights into cardiac pump function assessment.

Objective To investigate the predictive value of a model based on iodine concentration parameters derived from dual-energy computed tomography (DECT) for recurrence of laryngeal squamous cell carcinoma (LSCC) within 2 years after surgery in patients who underwent neoadjuvant chemotherapy(NAC). Methods A total of 133 LSCC patients (126 males, 7 females; mean age, 62±7 years) who underwent NAC and had pathologically confirmed diagnoses after surgery were retrospectively enrolled. Normallized iodine concentration (NIC) values of the tumors were measured in the arterial and venous phases on DECT before and after NAC, and the NIC change rate were calculated. According to the two-year follow-up results, patients were divided into a recurrence group (n=32) and a non-recurrence group (n=101). Differences in clinical and imaging parameters between groups were compared using t-test, Mann-Whitney U test, chi-square test, and Fisher's exact test. Variables with significant differences were entered into binary logistic regression to identify independent predictors of recurrence and to construct a nomogram model. Model performance was assessed using receiver operating characteristic (ROC) curves, area under the curve (AUC), sensitivity, and specificity, while calibration curves were used to evaluate accuracy. Results In the training cohort, significant differences between the recurrence and non-recurrence groups were observed in pre-treatment lymph node status, treatment response based on RECIST criteria, post-treatment NIC in the venous phase, and the NIC venous-phase change rate (all P<0.05). Logistic regression analysis identified pre-treatment lymph node status, RECIST response, and NIC venous-phase change rate as independent predictors of postoperative recurrence (all P<0.05). The nomogram constructed from these three factors achieved AUC values >0.80 in both the training and validation cohorts, with calibration curves demonstrating excellent agreement between predicted and observed outcomes. Conclusion Pre-treatment lymph node status, RECIST treatment response, and NIC venous-phase change rate can effectively predict the risk of two-year recurrence after surgery in LSCC patients receiving NAC.

Objective To explore the value of a predictive model based on MRI radiomics features for early postoperative recurrence in patients with esophageal cancer. Methods A total of 224 patients with esophageal cancer confirmed by postoperative pathological examination were retrospectively collected. All patients underwent preoperative multimodal MRI examination. Based on the 1-year postoperative follow-up, patients were divided into a recurrence group and a non-recurrence group. They were randomly assigned to a training set (157 cases; 58 recurrence, 99 non-recurrence) and a validation set (67 cases; 14 recurrence, 53 non-recurrence) in a 7∶3 ratio. Radiomics features were extracted from T₂WI BLADE (motion-corrected) sequence and contrast-enhanced T₁-weighted StarVIBE (free-breathing, radial K-space acquisition). Radiomics feature selection was performed, and Logistic regression was used to construct prediction models. The predictive performance of each model was evaluated using receiver operating characteristic (ROC) curve and the area under the curve (AUC) was calculate. Differences in AUC values were compared using the Delong test. Clinical net benefits of the model was assessed via decision curve analysis. Results Fourteen and 15 radiomics features were selected from the T₂WI BLADE and T₁WI StarVIBE sequences, respectively. Based on these features, T₂WI BLADE radiomics models, T₁WI StarVIBE enhanced radiomics models, and combined radiomics models were constructed. The AUC values of these three models in the training and validation sets were 0.778, 0.889, 0.920, and 0.699, 0.884, 0.911, respectively. Delong's test showed that the combined model had significantly better performance in predicting early recurrence than either single-sequence model (all P<0.05). Decision curve analysis shows that the combined model provided greater clinical net benefit and closely aligned with clinical outcomes. Conclusion MRI-based radiomics features can effectively predict the risk of early postoperative recurrence in patients with surgically esophageal cancer. The combined radiomics model demonstrates superior predictive performance compared to single-sequence models.

Objective To investigate the predictive value of gadoxetic acid-enhanced MRI features and clinicopathological characteristics for postoperative recurrence in patients with hepatocellular carcinoma (HCC). Methods A retrospective cohort of 209 patients with pathologically confirmed HCC who underwent surgical resection was analyzed. Based on postoperative follow-up results, patients were divided into a recurrence group (n=82) and a non-recurrence group (n=127). Clinicopathological parameters and MRI features were compared between groups. A nomogram model for predicting postoperative recurrence-free survival (RFS) in HCC patients was constructed based on the results of multivariate Cox regression analysis. Model performance was assessed using receiver operating characteristic (ROC) curve analysis, calibration curve analysis, and decision curve analysis (DCA). Kaplan-Meier curves were generated to compare RFS between subgroups, and the Log-rank test was used for statistical comparison. Results Multivariate Cox regression identified CK19 positivity, microvascular invasion (MVI) positivity, and the presence of the mosaic sign as independent risk factors for postoperative recurrence of HCC (all P<0.05). The nomogram model demonstrated a higher area under the ROC curve (AUC) compared with any single factor (CK19, MVI, or mosaic sign). The model showed good fit and calibration, and DCA demonstated good clinical utility. Median RFS was 30 months for CK19-positive patients, 17 months for MVI-positive patients, and 26 months for patients with mosaic sign positivity. Differences in 1-, 2-, and 3-year RFS rates according to CK19, MVI, and mosaic sign status were statistically significant (Log-rank test, all P<0.05). Conclusion CK19 positivity, MVI positivity, and mosaic sign positivity are independent risk factors for postoperative recurrence of HCC, and may serve as potential biomarkers for predicting poor prognosis after curative resection.

Objective To explore a noninvasive method for evaluating the efficacy of neoadjuvant chemoradiotherapy (nCRT) in patients with mid-low rectal cancer and to provide new insights for selecting optimal treatment strategies. Methods A total of 212 patients who underwent preoperative nCRT and radical resection with pathologically confirmed were retrospectively enrolled. Based on postoperative pathology, patients were divided into a pathological complete response(pCR) group(30 cases) and a non-pCR (npCR) group (182 cases). All patients were randomly assigned to a training set (170 cases; pCR 25 cases, npCR 145 cases) and a validation set(42 cases; pCR 5 cases, npCR 37 cases) at an 8∶2 ratio. Clinical features in the training set were screened using the Boruta algorithm, and selected features were combined with imaging indicators [tumor stage, tumor length, circumferential resection margin (CRM), extramural vascular invasion (EMVI), and MRI tumor regression grade (mrTRG)] to build three supervised machine learning classification models—logistic regression (LR), naïve Bayes (NB), and neural network (NN)—for predicting nCRT efficacy. Model performance was evaluated on the validation set. Interobserver agreement for imaging indicators was assessed using intraclass correlation coefficients (ICC) and Kappa statistics. Predictive performance was assessed with receiver operating characteristic (ROC) curves, area under the curve (AUC), accuracy, sensitivity,and specificity. Results Interobserver agreement between two radiologists for imaging indicators was excellent (ICC>0.9, all κ>0.8). Among the three models, the NN model showed the best predictive performance (training set: AUC 0.931, accuracy 0.900, sensitivity 0.890, specificity 0.875; validation set: AUC 0.687, accuracy 0.855, sensitivity 0.866, specificity 0.723). Conclusion Machine learning models constructed through feature selection can effectively predict the efficacy of nCRT in mid-low rectal cancer. Combining multiple indicators improves the preoperative predictive accuracy of mrTRG.

Objective To analyze the application value of dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) and diffusion-weighted imaging (DWI) in the prognostic evaluation of patients with glioma. Methods A total of 184 glioma patients confirmed by surgical resection or biopsy pathology were retrospectively selected. Based on postoperative follow-up survival time, patients were divided into a poor prognosis group (survival time < 24 months, n=64) and a good prognosis group (n=120). The general data, DCE-MRI parameters [volume transfer constant (K^trans), extracellular space volume fraction (v_e)], and DWI parameters [apparent diffusion coefficient (ADC), mean diffusion coefficient (MD)] were compared between the two groups using independent samples t test and Chi-square test. Univariate and multivariate Logistic regression were used to identify the prognostic factors predicting the prognosis of glioma patients, and a nomogram model was constructed accordingly. Receiver operating characteristic (ROC) curve and area under the curve (AUC) were used to analyze the predictive value of each single factor and nomogram model. Calibration curve was used to evaluate the calibration of the model, and decision curve was used to evaluate the net benefit of the model. Results The proportion of patients undergoing partial tumor resection and the levels of K^trans and v_e were significantly higher in the poor prognosis group than in the good prognosis group(all P<0.05). The proportion of patients receiving postoperative chemoradiotherapy and the levels of ADC and MD were significantly lower in the poor prognosis group than in the good prognosis group (all P<0.05). Partial tumor resection, K^trans, and v_e were independent risk factors for the prognosis of glioma patients, while postoperative chemoradiotherapy, ADC, and MD were protective factors for the prognosis of glioma patients (all P<0.05). A nomogram model was constructed by combining these six indicators. ROC curve analysis showed that the method of tumor resection, postoperative chemoradiotherapy, K^trans, v_e, ADC, MD, and the nomogram model could predict the prognosis of glioma patients. The nomogram model had the best prediction performance, with an AUC=0.829 (95%CI: 0.764-0.893). The calibration curve showed good consistency between the predicted probability and the actual probability of glioma patient prognosis. The decision curve analysis showed that the model achieved good clinical net benefit when the risk threshold interval was 37% to 88%. Conclusion DCE-MRI and DWI can be used to evaluate the prognosis of glioma patients. The nomogram model constructed based on clinical, DCE-MRI, and DWI factors has high clinical application value.

Vulnerable carotid plaques and cerebral small vessel disease (CSVD) are both significant etiologies of ischemic stroke, and they are closely interrelated. Imaging modalities, including ultrasound, CT angiography (CTA), and MR high-resolution vessel wall imaging (HR-VWI), enable the assessment of various vulnerable plaque characteristics. This review focuses on the correlation between vulnerable plaques and CSVD, aiming to provide a theoretical basis for further research and clinical application.

Neurovascular coupling (NVC) refers to the brain's regulatory capacity to adjust cerebral blood flow in response to neuronal activity, representing a coordinated function between neural activity and hemodynamics. Impairment in NVC diminishes connectivity between distinct brain regions, thereby impairing cognitive function. Currently, various neuroimaging techniques including arterial spin labeling, resting-state functional MRI, and functional near-infrared spectroscopy are employed to evaluate NVC function. These methods have confirmed that declines in NVC are closely linked to cognitive impairments in conditions such as vascular cognitive impairment, Alzheimer's disease, diabetic cognitive decline, and cognitive deficits in end-stage renal disease patients. This review aims to provide an overview of recent advances in neuroimaging research on the relationship between NVC and cognitive disorders.

Radiation-induced brain injury (RIBI) is a common complication following radiotherapy for nasopharyngeal carcinoma, often leading to cognitive impairment and other neurological sequelae. Multimodal MRI, including perfusion-weighted imaging (PWI), susceptibility-weighted imaging (SWI), blood oxygen level dependent-functional MRI (BOLD-fMRI), diffusion MRI, voxel-based morphometry (VBM), and magnetic resonance spectroscopy (MRS), provides multidimensional insights into the pathophysiological mechanisms of RIBI by revealing vascular damage, neural dysfunction, structural injury, and metabolic abnormalities. Artificial intelligence techniques can further integrate multimodal MRI data to enable early prediction and individualized assessment of RIBI. This review summarizes recent advances in the application of multimodal MRI in RIBI associated with nasopharyngeal carcinoma.

Extrahepatic cholangiocarcinoma (EHCC) is a highly aggressive malignancy with an extremely poor prognosis. Early diagnosis and precise treatment are crucial for improving patient outcomes. Conventional imaging examinations are limited by subjective assessment, and diagnostic accuracy remains to be further enhanced. Radiomics, which applies advanced computational techniques to deeply mine conventional imaging data, enables comprehensive characterization of tumor heterogeneity and is playing an increasingly important role in the precision medicine of EHCC. This review summarizes recent advances in the application of radiomics for predicting EHCC differentiation and staging, tumor microenvironment, lymph node metastasis, and postoperative recurrence. Furthermore, current challenges in radiomics research on EHCC are discussed, and future perspectives for its development are outlined.

T₂* mapping, a quantitative MRI technique based on multi gradient-echo sequences, has been widely used in the structurally complex musculoskeletal system due to its advantages of rapid imaging, non-invasive acquisition, and high spatial resolution. It is primarily employed to evaluate metabolism and blood perfusion changes in skeletal muscle during exercise, degeneration and load response of articular cartilage, degenerative changes of spinal intervertebral discs and osteoporosis, as well as injury and postoperative repair processes of ligaments and tendons. This technique demonstrates high sensitivity in detecting changes in tissue composition, water content, and collagen structure. This review summarizes the principles of T₂* mapping sequences and recent advancements in their applications for musculoskeletal system assessment, and discusses prospects for clinical use.

Temporomandibular disorders (TMD) are common in dentistry, with complex etiologies and diverse clinical manifestations. Imaging techniques allow visualization and precise quantification of temporomandibular joint (TMJ) structures. MRI can reveal soft tissue abnormalities by analyzing disc morphology, degree of displacement, joint effusion, and lateral pterygoid muscle attachment patterns; while CT provides quantitative parameters of bony structures such as condylar morphology, size, and position. This review summarizes the associations between CT and MRI structural features of the TMJ with TMD severity and prognosis, aiming to provide a theoretical basis for precision diagnosis and treatment of TMD.

Time-dependent diffusion MRI (TDDMRI) is an MRI technique that combines oscillating gradient spin-echo (OGSE) sequences with pulsed gradient spin-echo (PGSE) sequences. It can comprehensively capture the time-dependent characteristics of water molecule diffusion and generate quantitative microstructural parameters, such as intracellular volume fraction (f_in), cell diameter (d), and cell density, through modeling. This allows for non-invasive quantification of tumor microstructure. This paper briefly introduces the principles of TDDMRI and commonly used models, and systematically reviews its recent applications in human tumors, including differential diagnosis, treatment response prediction, prognosis evaluation, precision diagnosis, and risk stratification.

Malignant perivascular epithelioid cell tumor(PEComa) of soft tissue is an extremely rare mesenchymal neoplasms exhibits certain characteristic imaging features on ultrasound, CT, and MRI. However, it is important to differentiate it from other mesenchymal tumors within soft tissue that containing fat or melanin components,such as clear cell sarcoma, liposarcoma and malignant melanoma. This paper summarizes the imaging characteristics of malignant soft tissue PEComa to assist in clinical diagnosis and differential diagnosis.

Objective To explore the diagnostic value of whole-spinal CT angiography (CTA) combined with MRI in spinal vascular malformations. Methods A total of 86 patients with suspected spinal vascular malformations were retrospectively collected, including 47 males and 39 females, with age of 25-62 years (mean age, 42.74±8.51 years). All patients underwent CTA, MRI, and digital subtraction angiography (DSA). DSA findings were used as the golden standard. The consistency between CTA, MRI, CTA combined with MRI, and DSA was analyzed using the Kappa test. The diagnostic performance of the three methods was compared using the Chi-square test. The concordance rates of CTA and MRI with DSA were analyzed. Results DSA confirmed 65 cases of spinal vascular malformations. The diagnostic consistency of CTA and MRI with DSA was moderate (Kappa=0.554, 0.442), while CTA combined with MRI showed higher consistency (Kappa=0.712). The sensitivity of the three methods differed significantly (P<0.05), with CTA combined with MRI showing higher sensitivity than either CTA or MRI alone (both P<0.05). The concordance rates of CTA and MRI with DSA were 88.68% (47/53) and 86.27% (44/51), respectively. Conclusion CTA combined with MRI demonstrates high consistency with DSA and high sensitivity in diagnosing spinal vascular malformations, significantly improving the diagnostic performance compared with either modality alone.