Objective:To investigate the clinical value of a fully autonomous ultrasound robot in splenic ultrasound imaging.Methods:A retrospective study was conducted by enrolling 56 adult volunteers from the Third Medical Center of the Chinese PLA General Hospital between February 1-8，2024 as research subjects.A senior physician sequentially performed splenic ultrasound examinations using both the fully autonomous ultrasound robot and a matched portable ultrasound device. The acquired images were randomly coded and scored via a double-blind method by 3 physicians. The differences of the image quality scores and high-quality image proportions between the two groups were compared. Examination durations were recorded and compared between the two groups.Results:Both modalities successfully acquired splenic images in all 56 volunteers. No statistically significant differences were observed in image quality scores among the 3 physicians：（3.52 ± 1.31）points vs.（3.83 ± 1.23）points，（2.77 ± 1.23）points vs.（3.17 ± 1.17）points，and（3.48 ± 0.97）points vs.（3.79 ± 0.94）points（all P>0.05）. The numbers of images scoring ≥ 3 points showed no significant differences：45（80.36%） vs. 50（89.29%），30（53.57%） vs. 38（67.86%），and 48（85.71%） vs. 52（92.86%）（all P>0.05）. The fully autonomous ultrasound robot required significantly longer examination time［（60.86 ± 50.55）s vs.（7.95 ± 4.35）s， t=6.88， P＜0.01］. Conclusions:The fully autonomous ultrasound robot demonstrates comparable image quality and clinically acceptable image proportions to conventional portable ultrasound in splenic examinations. These findings suggest its potential equivalence to operator-dependent ultrasound for splenic imaging，supporting its feasibility as an alternative ultrasound modality despite longer procedural duration.

Objective:To construct a thyroid nodule segmentation model using ultrasound dynamic images and explore its potential for assisting in the screening of thyroid nodules.Methods:A total of 126 patients with thyroid nodules（comprising 150 nodules）who were diagnosed and treated at Xuzhou Cancer Hospital from April 2024 to December 2024 were prospectively enrolled. Two-dimensional ultrasound was performed to capture short-axis and long-axis video images of thyroid nodules，forming a dynamic ultrasound image dataset. The dataset was divided into training，validation，and test sets in a ratio of 6∶1∶3. After the training loss curve converged，the model that performed well on the validation set was selected for testing. Three-fold cross-validation was employed for training and testing. All 300 ultrasound videos were divided into three subsets. In each experiment，two subsets were used as the training set，and one subset was used as the test set to evaluate the model's generalization ability. A collaborative spatiotemporal diffusion model was established based on the dynamic trends and tissue texture details of thyroid nodules. Six widely used segmentation metrics were employed to evaluate the model's application capabilities.Results:The study included 126 patients with 150 thyroid nodules，300 dynamic ultrasound images，and video lengths of 3-4 seconds per nodule，resulting in 12 312 segmented images. The size of the thyroid nodules was（10.7 ± 10.6）mm（transverse diameter）×（8.4 ± 6.3）mm（anteroposterior diameter）. Among the nodules，62（41.3%）had clear boundaries，while 88（58.7%）had indistinct boundaries；61（40.7%）exhibited regular shapes，while 89（59.3%）were irregular；66（44.0%）had a taller-than-wide aspect ratio；and 70（46.7%）showed microcalcifications. The collaborative diffusion model based on dynamic ultrasound image segmentation achieved the following scores：a Jaccard score of（69.22 ± 0.03）%，a Dice score of（79.16 ± 0.18）%，a Precision score of（86.70 ± 0.17）%，a Recall score of（77.82 ± 0.04）%，an Sα score of（85.26 ± 0.01）%，and an Eθmn score of（90.58 ± 0.17）%. Compared to other models，this model demonstrated significant improvements across all evaluation metrics，achieving the highest values in each metric with increments of over 8% and 1%，respectively. Conclusions:The collaborative diffusion model with a dynamic controller，constructed based on dynamic ultrasound images of thyroid nodules，demonstrates excellent performance in ultrasound image segmentation. It improves the accuracy of thyroid nodule screening，thereby providing a valuable auxiliary diagnostic tool for clinical practice.

Objective:To propose a deep learning（DL）-based ultrasound imaging auxiliary tool for automatic segmentation and recognition of the brachial plexus（BP），and to enhance the accuracy and safety of clinical procedures.Methods:It was a multicenter study that collected 773 healthy subjects from Peking University Third Hospital and its branch campuses，the Third Medical Center of the Chinese PLA General Hospital，and Shanghai Eighth People's Hospital between August 2024 and February 2025. Brachial plexus（BP）images in the interscalene groove were captured used high-frequency ultrasound by senior sonographers，a dataset comprising 1 289 standardized images were constructed and the improved model（CHA-TransUNet）was trained. The test set was input into 6 different models（CHA-TransUNet，R50-Unet，TransUnet，SegFormer，SwinUnet，MISSFormer）for segmentation. Segmentation accuracy was evaluated using metrics including the Dice similarity coefficient（DSC），95% Hausdorff distance（HD95）and mean intersection over union（mIoU），and was compared with the segmentation results of 3 ultrasound physicians with varying experience levels（junior physicians and senior physicians）to validate the model's segmentation efficacy.Results:The CHA-TransUNet model established based on a dataset of 1 289 standardized images achieved segmentation results for the BP with a DSC of 90.15%，mIoU of 91.02%，and HD95 of 8.08. Its accuracy was higher than other mainstream models（DSC：90.15% vs. 87.60%，87.77%，81.35%，84.78%，84.55%），significantly better than junior physicians（DSC：90.15% vs. 68.73%， Z=-127.76， P<0.001），and approached the level of senior physician（DSC：90.15% vs. 86.15%， Z=-31.33， P=0.549）. The model demonstrated superior boundary recognition in complex anatomical structures（e.g.，C6/C7 nerve roots）compared to ultrasound physicians（junior and senior）（HD95：8.08 vs. 26.34，17.44，56.80）. Conclusions:This study proposes an analysis model for BP ultrasound images，CHA-TransUNet. This model achieves segmentation and recognition of the BP with relatively complex pathways and structures. The model exhibits high accuracy and stability，outperforming current mainstream network models and junior physicians while approaching the performance level of senior physicians. It assists junior physicians or trainees in more accurately identifying and localizing the BP.

Objective:To investigate the clinical value of a fully autonomous ultrasound robot in splenic ultrasound imaging.Methods:A retrospective study was conducted by enrolling 56 adult volunteers from the Third Medical Center of the Chinese PLA General Hospital between February 1-8，2024 as research subjects.A senior physician sequentially performed splenic ultrasound examinations using both the fully autonomous ultrasound robot and a matched portable ultrasound device. The acquired images were randomly coded and scored via a double-blind method by 3 physicians. The differences of the image quality scores and high-quality image proportions between the two groups were compared. Examination durations were recorded and compared between the two groups.Results:Both modalities successfully acquired splenic images in all 56 volunteers. No statistically significant differences were observed in image quality scores among the 3 physicians：（3.52 ± 1.31）points vs.（3.83 ± 1.23）points，（2.77 ± 1.23）points vs.（3.17 ± 1.17）points，and（3.48 ± 0.97）points vs.（3.79 ± 0.94）points（all P>0.05）. The numbers of images scoring ≥ 3 points showed no significant differences：45（80.36%） vs. 50（89.29%），30（53.57%） vs. 38（67.86%），and 48（85.71%） vs. 52（92.86%）（all P>0.05）. The fully autonomous ultrasound robot required significantly longer examination time［（60.86 ± 50.55）s vs.（7.95 ± 4.35）s， t=6.88， P＜0.01］. Conclusions:The fully autonomous ultrasound robot demonstrates comparable image quality and clinically acceptable image proportions to conventional portable ultrasound in splenic examinations. These findings suggest its potential equivalence to operator-dependent ultrasound for splenic imaging，supporting its feasibility as an alternative ultrasound modality despite longer procedural duration.

Objective:To construct a thyroid nodule segmentation model using ultrasound dynamic images and explore its potential for assisting in the screening of thyroid nodules.Methods:A total of 126 patients with thyroid nodules（comprising 150 nodules）who were diagnosed and treated at Xuzhou Cancer Hospital from April 2024 to December 2024 were prospectively enrolled. Two-dimensional ultrasound was performed to capture short-axis and long-axis video images of thyroid nodules，forming a dynamic ultrasound image dataset. The dataset was divided into training，validation，and test sets in a ratio of 6∶1∶3. After the training loss curve converged，the model that performed well on the validation set was selected for testing. Three-fold cross-validation was employed for training and testing. All 300 ultrasound videos were divided into three subsets. In each experiment，two subsets were used as the training set，and one subset was used as the test set to evaluate the model's generalization ability. A collaborative spatiotemporal diffusion model was established based on the dynamic trends and tissue texture details of thyroid nodules. Six widely used segmentation metrics were employed to evaluate the model's application capabilities.Results:The study included 126 patients with 150 thyroid nodules，300 dynamic ultrasound images，and video lengths of 3-4 seconds per nodule，resulting in 12 312 segmented images. The size of the thyroid nodules was（10.7 ± 10.6）mm（transverse diameter）×（8.4 ± 6.3）mm（anteroposterior diameter）. Among the nodules，62（41.3%）had clear boundaries，while 88（58.7%）had indistinct boundaries；61（40.7%）exhibited regular shapes，while 89（59.3%）were irregular；66（44.0%）had a taller-than-wide aspect ratio；and 70（46.7%）showed microcalcifications. The collaborative diffusion model based on dynamic ultrasound image segmentation achieved the following scores：a Jaccard score of（69.22 ± 0.03）%，a Dice score of（79.16 ± 0.18）%，a Precision score of（86.70 ± 0.17）%，a Recall score of（77.82 ± 0.04）%，an Sα score of（85.26 ± 0.01）%，and an Eθmn score of（90.58 ± 0.17）%. Compared to other models，this model demonstrated significant improvements across all evaluation metrics，achieving the highest values in each metric with increments of over 8% and 1%，respectively. Conclusions:The collaborative diffusion model with a dynamic controller，constructed based on dynamic ultrasound images of thyroid nodules，demonstrates excellent performance in ultrasound image segmentation. It improves the accuracy of thyroid nodule screening，thereby providing a valuable auxiliary diagnostic tool for clinical practice.

Objective:To propose a deep learning（DL）-based ultrasound imaging auxiliary tool for automatic segmentation and recognition of the brachial plexus（BP），and to enhance the accuracy and safety of clinical procedures.Methods:It was a multicenter study that collected 773 healthy subjects from Peking University Third Hospital and its branch campuses，the Third Medical Center of the Chinese PLA General Hospital，and Shanghai Eighth People's Hospital between August 2024 and February 2025. Brachial plexus（BP）images in the interscalene groove were captured used high-frequency ultrasound by senior sonographers，a dataset comprising 1 289 standardized images were constructed and the improved model（CHA-TransUNet）was trained. The test set was input into 6 different models（CHA-TransUNet，R50-Unet，TransUnet，SegFormer，SwinUnet，MISSFormer）for segmentation. Segmentation accuracy was evaluated using metrics including the Dice similarity coefficient（DSC），95% Hausdorff distance（HD95）and mean intersection over union（mIoU），and was compared with the segmentation results of 3 ultrasound physicians with varying experience levels（junior physicians and senior physicians）to validate the model's segmentation efficacy.Results:The CHA-TransUNet model established based on a dataset of 1 289 standardized images achieved segmentation results for the BP with a DSC of 90.15%，mIoU of 91.02%，and HD95 of 8.08. Its accuracy was higher than other mainstream models（DSC：90.15% vs. 87.60%，87.77%，81.35%，84.78%，84.55%），significantly better than junior physicians（DSC：90.15% vs. 68.73%， Z=-127.76， P<0.001），and approached the level of senior physician（DSC：90.15% vs. 86.15%， Z=-31.33， P=0.549）. The model demonstrated superior boundary recognition in complex anatomical structures（e.g.，C6/C7 nerve roots）compared to ultrasound physicians（junior and senior）（HD95：8.08 vs. 26.34，17.44，56.80）. Conclusions:This study proposes an analysis model for BP ultrasound images，CHA-TransUNet. This model achieves segmentation and recognition of the BP with relatively complex pathways and structures. The model exhibits high accuracy and stability，outperforming current mainstream network models and junior physicians while approaching the performance level of senior physicians. It assists junior physicians or trainees in more accurately identifying and localizing the BP.

Objective:To study the ultrasonographic manifestations of intercostal muscle and diaphragm involvement in Duchenne muscular dystrophy (DMD) and their correlations with functional status, and to explore the pattern of muscle damage in patients with DMD and the potential role of ultrasonography in assessing disease progression.Methods:A total of 28 patients with DMD who received treatment in the Third Medical Centre of PLA General Hospital from May to December 2023 were prospectively collected as DMD group, and 28 healthy children matched in age and sex were included as controls for a prospective study.Diaphragm thickening fraction (DTF) and intercostal muscle thickening fraction (ICMTF) were measured by B-mode and M-mode ultrasonography, and the muscle gray values were recorded. The differences between groups were compared, and the values of DTF and ICMTF in evaluating the structural and functional changes of respiratory muscle were analyzed.Results:Compared with the control group, the gray value of respiratory muscle was significantly decreased in DMD group, the diaphragm and intercostal muscle were significantly thickened at the end of inspiratory and expiratory periods, DTF was significantly decreased, and ICMTF was significantly increased (all P<0.001). Conclusions:Ultrasound can evaluate the structural changes of respiratory muscle in DMD, so as to clarify the relationship between the structure and function of respiratory muscle in DMD patients.

Objective:To explore the value of conventional ultrasonography combined with cervical compression in the diagnosis of orbital venous malformation (OVM).Methods:A total of 43 patients with suspected OVM were admitted in sequentially from January 2019 to July 2022 in the Third Medical Center of PLA General Hospital. All patients were examined by ultrasonography combined with cervical compression and demonstrated by operation or digital subtraction angiography (DSA). The conventional ultrasound features of OVM were summarized, and the value of conventional ultrasonography combined with cervical compression in the diagnosis of OVM was discussed.Results:The features of the conventional ultrasound combined with cervical compression for diagnosis of OVM were as follows: the interior of lesion was mainly tubular structure, and the compression test was positive. After cervical compression, the lesion enlarged and the inner diameter of the internal tubular structure widened. Doppler flow imaging showed that the interior of lesion was mainly venous blood flow. Compared with the results of postoperative pathology or DSA, the sensitivity, specificity, accuracy and positive predictive value of ultrasonography combined with cervical compression were 0.952, 1.000, 95.3% and 100%, respectively. The results of Fisher exact diagnosis showed that there was no significant difference between ultrasonography and operation or DSA of OVM( P>0.05). Conclusions:Conventional ultrasound combined with cervical compression can be used as an effective method for the diagnosis of OVM.

Objective:To explore the value of 5G robotic remote ultrasound in the diagnosis of plateau pulmonary edema(HAPE).Methods:A total of 27 patients who quickly entered Nagqu, Tibet at an altitude of 4 600 m-5 600 m from March to December 2021 and developed one of the clinical symptoms of HAPE were collected. All patients were examined by 5G remote robotic ultrasound and lung CT respectively. Kappa test was used to analyze the consistency of the two diagnostic results, and McNemar test was used to compare the difference in diagnostic results. The ROC curve was used to analyze the sensitivity and specificity of remote lung ultrasound scores in the diagnosis of HAPE.Results:Among the 27 patients, 16 showed thickening of pleural line, increasing of B line, lung consolidation, pleural effusion, etc. Meanwhile, 11 showed no abnormality. Additionally, 8 cases had diffuse pulmonary fluid in both lungs, and 8 cases had localized pulmonary fluid. ROC curve showed that the area under the curve of lung ultrasound score for the diagnosis of HAPE was 0.947 (95% CI=0.78-0.99, P<0.001). The sensitivity and specificity were 0.933 and 0.917, respectively. Lung CT diagnosis was positive in 15 cases. Lung CT showed thickening of lung texture, ground glass, small nodular shadow, fine reticulate shadow, etc. The diagnostic results of the two techniques were in good agreement (Kappa=0.924, P<0.001), and there was no significant difference between the two methods ( P>0.05). Conclusions:5G remote robotic ultrasound has high consistency with CT in the diagnosis of HAPE and is an alternative early diagnosis method for HAPE. It may have clinical application value in scattered medical resources and remote plateau areas.

Objective:To explore the value of 5G-based robotic remote ultrasound diagnosis system in musculoskeletal joint injuries.Methods:From March to December 2020, 58 volunteers at a training base who felt musculoskeletal pain or paresthesia were selected and performed both robotic remote ultrasound (remote ultrasound group) and conventional ultrasound (portable ultrasound group). The two types of examinations were compared, the consistency of the two diagnosis results was analyzed by the Kappa test, and the the difference of the diagnosis results was compared by McNemar test.Results:Among the 58 volunteers, 40 cases were positive by both methods and 11 volunteers had 2-3 positive results. There were 59 positive results in the remote ultrasound group and 64 positive results in the portable ultrasound group. The positive rate of the examination sites from high to low was knee joint>foot and ankle joint >hand and wrist joint >shoulder joint>elbow joint, calf and hip. The diagnosis results of the two groups were in good consistency (Kappa=0.782, P<0.001), and there was no statistically significant difference in the diagnosis results between the two groups (χ 2=3.2, P=0.063). Five more diseases with positive results were detected in the portable ultrasound group: 1 meniscus injury, 1 medial collateral ligament injury, 1 soft tissue injury around the metatarsal, 1 biceps tendinitis with effusion and 1 cubital ulnar nerve subluxation. Conclusions:The 5G-based robotic remote ultrasound system has good consistency with conventional ultrasound in the diagnosis of musculoskeletal injures. It can be applied to the ultrasound diagnosis of musculoskeletal joint injuries in remote areas.