ArchiveTo establish a prediction model for the conversion risk from mild cognitive impairment (MCI) to Alzheimer's disease (AD) based on structural MRI features, and employ interpretable algorithms to quantify the contribution of critical brain region features to model discrimination.
A total of 356 sets of T1-weighted structural MRI data were enrolled from the Alzheimer's Disease Neuroimaging Initiative (ADNI) database. All samples were stratified and randomly divided into a training set containing 237 subjects [119 mild cognitive impairment converters (MCIc) and 118 mild cognitive impairment non-converters (MCInc)] and an independent test set consisting of 119 subjects (60 MCIc and 59 MCInc). Five-fold cross-validation was conducted on the training set to evaluate model stability. FreeSurfer v7.4.1 software was used to extract 1,481 brain structural features for constructing the feature matrix. Boruta nonlinear wrapper feature selection was performed, and 17 core features were retained ultimately. Four models, including support vector machine (SVM), extreme gradient boosting (XGBoost), light gradient boosting machine (LightGBM) and tabular prior-data fitted network (TabPFN), were constructed using the original and filtered feature sets separately. Logistic regression combined with LASSO regression (LR+LASSO) was introduced as a classic linear baseline for comparative analysis. SHapley Additive exPlanations (SHAP) was further applied to interpret brain region features associated with high conversion risk.
After Boruta feature screening, the separability of samples in the principal component space was improved substantially. The 17 retained features mainly involve brain regions such as the hippocampus, amygdala, temporal lobe and parietal lobe. On the independent test set, the Boruta+TabPFN pipeline achieved an accuracy of 0.765, an area under curve (AUC) of 0.823 and a false negative rate (FNR) of 0.200. Compared with LR+LASSO, Boruta+TabPFN achieved better performance in accuracy, precision, specificity and AUC. However, LR+LASSO yielded slightly higher recall at the target recall threshold, indicating that the linear baseline model prioritizes high sensitivity at the expense of specificity.
Effective nonlinear feature screening is required to build robust discriminative models with high-dimensional structural MRI features. Although Boruta+TabPFN did not obtain the optimal value for every single evaluation metric, it delivered steady performance in AUC, recall and false negative control, which better satisfies the clinical requirement of identifying high-risk individuals in the early screening of MCI-to-AD conversion. SHAP interpretability analysis demonstrated that the risk of MCI progression to AD is driven by coordinated morphological changes across multiple brain regions including the hippocampus, amygdala, temporal lobe, parietal lobe and limbic system, providing interpretable imaging evidence for early disease risk evaluation.
Portable microwave stroke detection devices often suffer from sparse spatial sampling due to a limited number of antennas. This limitation leads to severe ill-posed and ill-conditioned problems in microwave inverse scattering imaging. To address these technical challenges, this paper proposes a cascaded imaging architecture based on a physics-constrained Swin Transformer, termed deep learning-enhanced truncated singular value decomposition (DL-TSVD) algorithm. By leveraging the global self-attention mechanism of the deep learning network, this architecture achieves high-fidelity mapping from a low-dimensional 8×8 sparse scattering matrix to a 16×16 dense matrix. Subsequently, the dense matrix is fed into the truncated singular value decomposition (TSVD) algorithm to accomplish the high-resolution inversion of the brain's permittivity distribution. In terms of S-parameter fitting accuracy, the complex-domain root mean square error of the predicted matrix is significantly reduced by 67.53% compared with traditional interpolation methods. Additionally, phantom experiments verify the noise-resistant generalization ability of the DL-TSVD architecture under real hardware noise and antenna mutual coupling environments. The structural similarity index measure (SSIM) of the reconstructed images reaches 0.541. The results demonstrate that the proposed method effectively mitigates the ill-conditioned nature of the electromagnetic inverse scattering problem under sparse arrays. It breaks through the resolution bottleneck of spatial sampling without increasing the number of antennas, providing a valuable technical reference for portable microwave stroke monitoring systems.
To explore the diagnostic and evaluation value of musculoskeletal ultrasound elastography in the standing position for nonspecific low back pain (NLBP).
A total of 16 patients with NLBP, who had not received treatment within two weeks prior to enrollment, were recruited as the observation group from Shanghai Ruijin Rehabilitation Hospital between June 2024 and June 2025, and 8 healthy people in the same period were selected as the control group. Before treatment or during physical examination, basic data such as gender, age, and body mass index (BMI) were collected, and musculoskeletal ultrasound related indicators (thickness of transversus abdominis, elastic modulus of transversus abdominis, thickness of lumboiliocostoid muscle, elastic modulus of lumboiliocostoid muscle) were measured in standing position. At the same time, the visual analogue scale (VAS) was used to evaluate the degree of pain in the observation group. R 4.5.2 software was used for statistical analysis and diagnostic modeling, and the receiver operating characteristic (ROC) curve was drawn. The area under curve (AUC), sensitivity, specificity and cut-off value of each diagnostic model were calculated, and the diagnostic efficacy of each model was compared.
There was no significant difference in gender, age, and BMI between the observation group and the control group (P>0.05). There was no significant difference in the thickness of transversus abdominis and lumboiliocostalis between the observation group and the control group (P>0.05). The elastic modulus of transverse abdominis in the observation group was significantly lower than that in the control group, and the elastic modulus of lumboiliocostalis in the observation group was significantly higher than that in the control group, and the differences were statistically significant (P<0.05). Pearson correlation analysis showed that among the diagnostic models, the quadratic power model of elastic modulus ratio (the ratio between the elastic modulus of transverse abdominis muscle and the elastic modulus of lumboiliocostoid muscle) had the best diagnostic performance, with an AUC of 1.000, followed by the elastic modulus of lumboiliocostoid muscle and elastic modulus ratio model, with AUC of 0.852 and 0.812, respectively. In addition, the VAS score of the observation group was positively correlated with the elastic modulus of the lumbar iliocostalis muscle in the standing position (P<0.001), and the VAS score of the observation group was negatively correlated with the ratio of elastic modulus in the standing position (P<0.001). In the control group, the ratio of elastic modulus in standing position was between 1.22 and 1.90, while in the observation group, the ratio was between 0.50 and 1.22 or >1.90.
The ratio of the elastic modulus of the transverse abdominis muscle to that of the lumboiliocostoid muscle measured by ultrasound in standing position can be used as an objective parameter, which is correlated with the subjective VAS score of NLBP patients, suggesting that the evaluation of muscle and bone elastography in standing position is expected to be one of the objective evaluation methods for NLBP patients, and provide quantitative basis for early diagnosis, efficacy evaluation and rehabilitation plan development.
With the development of single-cell sequencing technology, life science research has gradually moved from population average measurement to cell resolution, which provides a new research method for understanding the heterogeneity and cell state changes within tissues. The continuous progress of this technology system is not only due to the improvement of sequencing chemistry, but also closely related to the continuous optimization of instrument platforms such as microfluidic droplets, molecular barcoding, long-read sequencing and spatially resolved detection. Different technical routes have differences in sequencing throughput, read length characteristics and noise structure, which largely affect the subsequent data processing methods and promote the continuous adjustment and development of bioinformatics analysis methods. Combined with the main engineering paths of the single-cell sequencing instrument platform, this paper combs the representative solutions such as physical sorting, microfluidic system, long-read sequencing and space technology. At the same time, we discuss the characteristics of data structure shaped by barcodes and unique molecular identifiers (UMI), the statistical modeling problem caused by high sparsity, and the requirements of computational framework for large-scale data analysis. In addition, the development trend of single-cell sequencing and bioinformatics platforms in automation, multimodal integration and computing power improvement are summarized, in order to provide reference for the optimization of related technical systems.
This paper systematically evaluates the efficacy and safety of the domestic electromagnetic muscle toning device called NEOSCULPT, which is based on high-intensity focused electromagnetic field (HIFEM) technology for abdominal and gluteal contouring.
A prospective self-controlled clinical trial was conducted in two phases (gluteal and abdominal), each involving 32 healthy subjects. All subjects received targeted electromagnetic interventions twice a week for 30 minutes each session over a period of 4 weeks. Body circumference measurements, MRI imaging analysis, body composition analysis, and blood and urine biochemical tests were performed at baseline, at the end of the intervention, and one month post-intervention. Adverse events were recorded.
Following the gluteal intervention, the average overall hip circumference increased by 2.2% (P<0.001), and the average waist-to-hip ratio in females improved by 6.7% (P<0.01). Following the abdominal intervention, male subjects showed a 6.3% increase in abdominal muscle area (P<0.01) and a 10.4% decrease in fat area (P<0.001); female subjects showed an 11.8% increase in abdominal muscle area (P<0.001) and a 7.5% decrease in fat area (P<0.01). The overall efficacy rate for muscle enhancement and fat reduction was 80%, with an average reduction of 1% in total body fat percentage (P<0.001) and an average reduction of 8.2% in abdominal subcutaneous fat thickness (P<0.001). These effects were generally maintained during the follow-up period. Apart from a physiological increase in creatine kinase levels post-exercise, no abnormal fluctuations were observed in blood and urine biochemical parameters. Only one case experienced changes in renal function, which recovered after treatment, with no serious adverse events occurring.
The NEOSCULPT electromagnetic muscle toning device significantly increased participants' gluteal circumference and improved gluteal firmness, while simultaneously achieving dual effects of abdominal muscle growth and fat reduction, with excellent clinical application safety. These findings provide high-quality evidence-based support for the clinical application of domestically produced non-invasive muscle enhancement devices.
Cardiovascular and respiratory diseases are leading global public health burdens, and accurate acquisition and analysis of cardiopulmonary sounds are critical for early disease diagnosis. Traditional auscultation is subjective and non-quantifiable, while commercial single-channel electronic stethoscopes face issues such as signal aliasing and limited intelligence. Multi-channel electronic auscultation systems have emerged as a research hotspot due to their advantages of rich spatial information and strong anti-interference capabilities. This article reviews the research progress of multi-channel electronic stethoscopes, focusing on the optimization of hardware acquisition systems such as multi-position compatibility and flexible sensor array applications, as well as innovations in algorithms for separating heart and lung sounds such as blind source separation algorithms and deep learning methods, and clinical applications including vital sign monitoring, abnormal sound identification, and auxiliary diagnosis. We highlight current technical breakthroughs and bottlenecks in integration, robustness, and intelligence, and propose future directions including standardized dataset construction, physics-informed algorithm design, and clinical translation pathways to promote the field's development.
Evaluate the diagnostic performance of a deep learning model in assisting primary care ultrasound doctors in differentiating benign and malignant breast nodules.
This study retrospectively collected female patients who were diagnosed with BI-RADS category 3 and 4 breast nodules during routine ultrasound examinations at Shanghai Sixth People's Hospital from September 2022 to September 2024. The clinical data and ultrasound images of these patients were collected.. Five deep learning models (GoogLeNet, VGG11, AlexNet, ResNet-34, ViT) were constructed, and the optimal model was selected based on metrics such as area under curve (AUC). Visual heatmaps of the ultrasound images were generated using the optimal model. Community ultrasound physicians independently reviewed the images without knowledge of the pathological results. Two weeks later, the physicians re-reviewed the images with the assistance of the generated heatmaps. Diagnostic performance was compared among three approaches: deep learning model alone, physician independent review, and deep learning-assisted physician review, using accuracy, sensitivity, specificity, and precision as metrics.
A total of 430 patients with breast nodules were included and divided into a training set (n=301) and a validation set (n=129). Among the different learning models constructed based on breast ultrasound images, the ResNet-34 model performed best in predicting benign and malignant breast nodules, with AUC, accuracy, and sensitivity superior to other models in both the training and validation sets. Therefore, the ResNet-34 model was selected as the optimal model. Among the three approaches, the diagnostic efficacy of independent review by primary care physicians was the lowest. The deep learning model slightly outperformed primary care physicians in terms of accuracy, sensitivity, and precision, while the diagnostic efficacy of the deep learning model assisting ultrasound physicians in reading was the strongest, with all evaluation indicators significantly improved. In terms of the improvement rate, the sensitivity and F1 value showed the most significant increase, rising by 34.3% and 29.9% respectively.
Deep learning models based on ultrasound images can effectively distinguish the benign and malignant nature of breast nodules. Combined with the generated visual heatmaps, they can significantly enhance the diagnostic efficacy of assisting primary ultrasound physicians in distinguishing between benign and malignant breast nodules, providing valuable guidance for selecting treatment plans for patients with breast nodules.
To construct a full life cycle quality management model for traditional Chinese medicine (TCM) diagnostic and treatment equipment and evaluate its application effects.
A total of 357 TCM diagnostic and treatment equipment units under management and 360 patients receiving physiotherapy at a tertiary Class-A TCM hospital were selected as the research subjects. From January to December 2024, the traditional equipment management model was implemented for these 357 units, and 180 patients were included as the traditional equipment management group; from January to June 2025, the full life cycle quality management model was implemented for the same 357 units, with the information management platform as the core hub and the unique electronic identity file of the equipment as the data mainline, covering the entire process of procurement and acceptance, usage monitoring, preventive maintenance, fault response and scrap replacement, and 180 patients were included as the full life cycle quality management group. Monthly average failure rate, intact rate, idle rate, preventive maintenance execution rate, adverse event incidence, maintenance response time, and patient clinical efficacy were compared between the two groups.
Compared with the traditional equipment management group (2.35%, 3.08%, 2.80%), the full life cycle quality management group showed significantly lower monthly average failure rate (0.67%), idle rate (0.56%), and adverse event incidence (0.56%); maintenance response time was shorter [(2.00±0.45)h vs. (6.48±1.23)h]; intact rate (96.64% vs. 85.15%) and preventive maintenance execution rate (95.52% vs. 68.35%) were markedly higher; and the total clinical effective rate (93.33% vs. 78.33%) was significantly improved. All differences were statistically significant (P<0.05).
Compared with the traditional management model, the full life cycle quality management model for TCM diagnostic and treatment equipment significantly reduces equipment failure rates and adverse event incidence, enhances operational efficiency and patient clinical outcomes, and demonstrates considerable value for broader implementation.
To investigate the prognostic value of pre-treatment skeletal muscle index at the third lumbar vertebra (L3-SMI) for survival in patients with cervical cancer, and to analyze its predictive performance across different International Federation of Gynecology and Obstetrics (FIGO) stages.
Clinical data and imaging studies of patients with cervical cancer diagnosed and treated at Hunan Cancer Hospital from January 2014 to December 2019 were retrospectively collected. Skeletal muscle was segmented at the mid-vertebral level of the third lumbar vertebra, and skeletal muscle area (SMA) was measured. L3-SMI was calculated as SMA divided by body surface area (BSA). The optimal cut-off value of L3-SMI for predicting overall survival (OS) in cervical cancer patients was determined using receiver operating characteristic (ROC) curve analysis. Survival differences between patients with different L3-SMI levels were compared using the Kaplan-Meier method and Log-Rank test. The independent prognostic value of L3-SMI was analyzed using the Cox proportional hazards regression model.
A total of 207 patients were enrolled in this study. Survival analysis showed that L3-SMI was not significantly correlated with OS in the early-stage group (FIGO stages I and IIA, P>0.05), but was significantly correlated with OS in the advanced-stage group (FIGO stages IIB, III and IV, P<0.05). The optimal cut-off value of L3-SMI for predicting OS in advanced-stage patients was 39.8 cm2/m2, with an area under curve (AUC) of 0.668. Accordingly, 101 advanced-stage patients were divided into the high L3-SMI group (≥39.8 cm2/m2, n=51) and the low L3-SMI group (<39.8 cm2/m2, n=50). Kaplan-Meier analysis demonstrated that patients in the high L3-SMI group had significantly better OS than those in the low L3-SMI group, with a statistically significant difference (Log-Rank χ2=8.20, P=0.004 2). Multivariate Cox regression analysis revealed that low L3-SMI remained an independent risk factor for OS in advanced-stage patients after adjusting for age, FIGO stage, and pathological type (HR=2.561, 95%CI: 1.040~6.336, P=0.039).
Pre-treatment L3-SMI is an independent prognostic factor for survival in patients with advanced-stage cervical cancer (FIGO stage IIB and above), and this biomarker has clinical value in predicting the prognosis of advanced cervical cancer patients.
To investigate the clinical efficacy of a 10-day intervention combining progressive core training with individually adjusted intensity, neuromuscular electrical stimulation (NMES), and manual therapy.
A total of 50 postpartum patients with diastasis recti abdominis (DRA) who visited the Pelvic Floor Clinic of our hospital from December 2024 to January 2026 were enrolled and randomly divided into a control group and an experimental group, with 25 cases in each group. The control group received NMES and manual therapy, while the experimental group received progressive core training in addition to the same NMES and manual therapy. Both groups were treated once daily for 10 consecutive days. The inter-rectus distance (IRD), abdominal circumference, and Oswestry disability index (ODI) scores were compared between the two groups before and after the intervention.
After 10 consecutive days of intervention, both groups showed significant improvements in IRD measured at three sites (4.5 cm above the umbilicus, at the umbilicus, and 4.5 cm below the umbilicus), abdominal circumference, and ODI scores compared with baseline (all P<0.001). Compared with the control group, the experimental group demonstrated significantly greater improvements in IRD at all three sites, abdominal circumference, and ODI scores after the intervention (all P<0.05).
Under the regimen of once-daily intervention for 10 consecutive days, the combination of progressive core training, NMES, and manual therapy effectively reduces IRD and abdominal circumference, and improves low back pain-related functional disability in postpartum patients with diastasis recti abdominis.
To investigate the application value of dual-energy X-ray absorptiometry (DXA) combined with magnetic resonance imaging (MRI) in evaluating the fracture severity and predicting the functional prognosis in patients with osteoporotic vertebral compression fracture (OVCF).
70 patients with OVCF admitted to our hospital from December 2023 to July 2025 were included. All patients underwent DXA and MRI examinations, recording imaging parameters such as bone mineral density (BMD) values, T values, vertebral compression rate, bone marrow edema volume, and signal intensity ratio. According to the Genant semi quantitative method, the patients were divided into mild group, moderate group, and severe group. Compare the differences in imaging parameters, visual analogue scale (VAS) scores, and Oswestry disability index (ODI) among patients with different fracture severity. Pearson analysis was used to evaluate the correlation between imaging parameters, VAS scores, and ODI. According to the ODI improvement rate at 6 months after treatment, the patients were divided into good prognosis group and poor prognosis group. Multivariate logistic regression analysis was used to identify independent risk factors for poor prognosis. Receiver operating characteristic (ROC) curves were used to evaluate the predictive value of each indicator and combined model for poor prognosis.
With the fracture severity increased, the BMD and T values gradually decreased, the vertebral compression rate, bone marrow edema volume, signal intensity ratio, VAS scores, and ODI gradually increased in the three groups, the differences were statistically significant (P<0.05). Pearson correlation analysis showed that BMD values were negatively correlated with VAS scores and ODI (P<0.05), the bone marrow edema volume was positively correlated with VAS scores and ODI (P<0.05). Multivariate logistic regression analysis indicated that T values (OR=2.416, 95%CI: 1.213~4.798), vertebral compression ratio (OR=3.121, 95%CI: 1.472~6.618), and bone marrow edema volume (OR=2.183, 95%CI: 1.134~4.209) were independent risk factors for poor prognosis (P<0.05). ROC curve analysis showed that the area under curve (AUC) of T values, vertebral compression ratio, and bone marrow edema volume for predicting the poor prognosis were 0.763, 0.781, and 0.752, while the AUC of the combined model was 0.892.
T values, vertebral compression rate, and bone marrow edema volume are closely related to the fracture severity and functional prognosis in patients with OVCF. DXA combined with MRI can comprehensively evaluate OVCF from multiple dimensions such as bone mass, vertebral structure, and bone marrow response, which is helpful in predicting the prognosis, and has important value in guiding individualized clinical treatment.
To analyze the differences in fat deposition in the pancreas, kidneys and vertebral bodies between patients with metabolic syndrome (MetS) and healthy subjects using the mDixon-Quant quantitative technique, and to investigate metabolic parameters among MetS patients with different visceral fat deposition patterns.
This prospective study enrolled 58 MetS patients admitted to the Department of Endocrinology, the Second Affiliated Hospital of Mudanjiang Medical University between June 2023 and September 2024 into the MetS group. Meanwhile, 52 healthy volunteers were recruited as the control group. The fat fraction (FF) and T2* values of the pancreas, renal cortex, renal medulla and vertebral bodies were measured via mDixon-Quant sequence. Clinical data including waist circumference, body mass index (BMI), triglycerides, total cholesterol and high-density lipoprotein cholesterol were collected. The K-means clustering algorithm was applied to stratify MetS patients into subgroups according to fat deposition patterns, and metabolic indicators were compared across subgroups.
Waist circumference, BMI, triglyceride levels, as well as FF values of the pancreas, renal cortex and vertebral bodies were significantly higher in the MetS group than in the control group, whereas high-density lipoprotein cholesterol levels and T2* values of the pancreas, renal cortex and renal medulla were markedly lower (all P<0.05). Clustering analysis classified MetS patients into three phenotypes: multi-organ fat deposition, pancreas-skeleton predominant fat deposition, and kidney-predominant fat deposition. Patients with multi-organ fat deposition exhibited significantly higher waist circumference, triglycerides and total cholesterol compared with the other two subgroups (P<0.05).
The FF values of the pancreas, kidneys and vertebral bodies are significantly elevated in MetS patients. Among all phenotypes, individuals with multi-organ fat deposition present higher triglyceride and total cholesterol concentrations. The concept of multi-organ synergistic lesions can provide imaging evidence for precise subtyping and risk stratification of metabolic syndrome.
To analyze the diagnostic value of low-dose CT in gouty arthritis.
A total of 60 patients with acute gout attacks who visited Yunan County Hospital of Traditional Chinese Medicine in Yunfu City from January to December 2025 were enrolled in this study. Using a random number table, they were assigned to Group A and Group B, with 30 patients in each group. Group A underwent low-dose CT, while Group B underwent conventional CT. Joint aspiration fluid examined under polarized light microscopy served as the gold standard. The diagnostic performance of both groups in identifying gouty arthritis was compared, and radiation doses were calculated for each group.
There were no significant differences in diagnostic performance parameters between the two groups (P>0.05). The effective dose of Group A was (0.27±0.08) mSv, the dose-length product was (46.23±8.54) mGy·cm, and the volumetric CT dose index was (4.73±0.87) mGy, all of which were significantly lower than those of Group B [(0.96±0.32) mSv, (300.54±51.37) mGy·cm, (23.19±3.85) mGy] (all P<0.05).
There was no significant difference in diagnostic performance between low-dose CT and conventional CT for gouty arthritis; however, low-dose CT reduces radiation dose and lowers the risk of radiation exposure for patients.
To investigate the clinical application value of 64 slice spiral CT head and neck CT angiography (CTA) in the diagnosis of head and neck vascular lesions.
80 patients with suspected head and neck vascular disease in our hospital from June 2024 to December 2025 were selected as the research object. All patients underwent 64 slice spiral CT head and neck CTA examination and digital subtraction angiography (DSA) examination. The detection performance, diagnostic efficacy indicators, examination-related indicators, and incidence of adverse reactions of 64-slice spiral CT head and neck CTA technology were evaluated for various types of head and neck vascular lesions.
There was no statistically significant difference in the detection performance of various types of head and neck vascular lesions between the two examination methods (P>0.05). Using surgical pathology results or clinical comprehensive diagnosis as the gold standard, the sensitivity, specificity, and accuracy of 64-slice spiral CT head and neck CTA technology were 93.94%, 100.00%, and 95.00%, respectively, compared with those of DSA (96.97%, 100.00%, and 97.50%), there was no statistically significant difference (P>0.05). The examination time [(12.35±3.18) min], radiation dose [(320.45±50.63) mGy·cm] and contrast agent dosage [(35.76±4.32) mL] of 64 slice spiral CT head and neck CTA were lower than those of DSA [(35.67±5.42) min, (650.32±80.45) mGy·cm, (58.63±6.27) mL], and the difference was statistically significant (P<0.05). The incidence of adverse reactions of 64 slice spiral CT head and neck CTA was 2.50%, which was lower than 11.25% of DSA, and the difference was statistically significant (P<0.05).
The application of head and neck CTA technology in the diagnosis of head and neck vascular disease can improve the diagnostic accuracy and diagnostic efficiency through the advantages of 64 slice spiral CT, and has the advantages of rapid examination, low radiation dose, low dosage of contrast agent, high safety and so on, which can be used as one of the preferred schemes for the clinical diagnosis of head and neck vascular disease.
To quantitatively compare the left-right asymmetry of the sacroiliac joints between postpartum women with low back pain and healthy postpartum women using multi-slice spiral CT (MSCT), and to investigate its imaging features and diagnostic value for postpartum sacroiliac joint subluxation.
35 postpartum women with low back pain treated in our hospital between January 2023 and December 2024 were selected as the study group (postpartum low back pain group), while 35 healthy postpartum women served as the control group (healthy postpartum group). 16-slice spiral CT was used to scan the sacroiliac joints, and multiplanar reconstruction was employed to for image processing. A total of 18 measurement points were selected in the anterior, middle, and posterior regions from S1 to S3 to calculate the degree of left-right asymmetry. Differences in asymmetry between the two groups at each measurement point were compared, and receiver operating characteristic (ROC) curve analysis was used to identify diagnostic cutoff values.
There was no significant difference in age, postpartum duration, or other baseline characteristics between the two groups. Significant left-right asymmetry was observed at 14 measurement points (P<0.05). The degree of asymmetry at all significant locations was greater in the postpartum low back pain group than in the control group, with the most pronounced difference observed in the posterior region of S2-2 measurement point [(0.860±0.608) mm vs. (0.491±0.205) mm, P=0.001 7)]. ROC curve analysis showed that the asymmetry in the anterior region of S1-1 level had the best diagnostic performance (AUC=0.738), with a cutoff value of 0.3 mm.
Postpartum women with low back pain exhibit significant left-right asymmetry in the sacroiliac joints, with increased asymmetry at multiple anatomical locations. A left-right asymmetry ≥0.3 mm in the anterior region of S1-1 level can serve as a key reference standard for evaluating postpartum sacroiliac joint subluxation.
To investigate the clinical diagnostic value and safety of endoscopic ultrasound-guided fine-needle aspiration/biopsy (EUS-FNA/B) in upper gastrointestinal and extramural lesions.
A retrospective analysis was performed on 50 patients who underwent EUS-FNA/B from January 2024 to December 2025. The final comprehensive diagnosis was based on surgical pathology when available, or on follow-up imaging and treatment response in nonsurgical cases. Diagnostic agreement, sensitivity, specificity, predictive values, accuracy, AUC and complications were analyzed.
The final diagnosis included 18 benign and 32 malignant lesions. EUS-FNA/B identified 31 malignant and 19 benign lesions, with 30 true positives, 17 true negatives, 2 false negatives and 1 false positive. The sensitivity, specificity, positive predictive value, negative predictive value and accuracy were 93.75%, 94.44%, 96.77%, 89.47% and 94.00%, respectively. The AUC was 0.94. No serious complications occurred.
EUS-FNA/B has high diagnostic accuracy and favorable safety for upper gastrointestinal and extramural lesions, and can provide reliable evidence for clinical decision-making.
Super-resolution microscopy (SRM) technology exhibits enormous application potential in the early diagnosis and treatment of cancer, owing to its capability of achieving ultra-high-resolution imaging of cellular structures or small-molecule substances at the nanoscale and even single-molecule level. This article systematically introduces the imaging principles of several mainstream SRM technologies, summarizes the commonly used fluorescent probe types of different technologies and the characteristics and specific application scenarios of each probe. Then, it outlines the application cases of SRM in early cancer diagnosis by observing diagnostic targets related to carbohydrates, membrane proteins, and exosomes. This demonstrates its core application value in identifying molecular distribution, monitoring structural changes, and revealing early cancer characteristics. It further introduces the advantages and specific applications of SRM in treating cancer through monitoring the endocytosis of nanomedicines, enhancing the efficacy of photodynamic therapy, and clarifying the mechanism of drug resistance. Finally, it discusses and analyzes the current shortcomings of SRM in imaging speed, probe stability, and clinical applicability, and looks forward to future research directions, aiming to promote the further application of SRM in cancer precision diagnosis and treatment.
Environment-responsive smart membranes consist of two components: a porous membrane substrate and a polymeric functional switch that senses external environmental stimuli. Through external environmental stimuli, the physical and chemical properties of the smart membrane are altered, thereby enabling active regulation of separation performance or other functionalities. Due to their unique properties, environment-responsive smart membranes have been widely applied in various fields and are currently a research hotspot. This paper adopts a combination of quantitative and qualitative research methods to analyze the changing trend in patent applications, the distribution of major source countries, the ranking of major applicants, and the comparison of application volumes across different technical branches in the field of smart membrane technology. Furthermore, it provides a detailed analysis of key technologies, including the LCST phase transition characteristics of temperature-responsive membranes, imprinting and coating techniques for molecularly responsive membranes, three directions of light-responsive membranes—photoactuators, photochromism, photocatalysis, as well as graft blending and response sensitivity enhancement for pH-responsive membranes. The research shows that China is the major source country for environment-responsive smart membrane patent applications, with universities being the main applicants and enterprises accounting for a relatively low share. Temperature-responsive, molecularly responsive, light-responsive, and pH-responsive smart membranes are the key research and development directions. To this end, future efforts should focus on developing bio-stimuli-responsive membranes and biomimetic membranes, continuously optimizing the performance of temperature and pH-responsive membranes, expanding the applications of light-responsive membranes, promoting the functional transformation of molecularly responsive membranes, and accelerating industrial transformation through industry-university-research collaboration.
No. 86 Xueyuan South Road, Haidian District, Beijing
010-62662699(Operation and Technical Support)
qkjq@cast.org.cn
Copyright © 2025 China Association for Science and Technology. All rights reserved. For all open access content, the relevant licensing terms apply.
Sponsored by the Office of the Leading Group for Cybersecurity and Informatization of CAST, and supported by Science and Technology Review Publishing House
