To evaluate the value of super-resolution ultrasound (SRUS) in identifying tumor blood flow characteristics for differential diagnosis of neoplastic versus non-neoplastic gallbladder polyps.

A total of 86 patients with gallbladder polypoid lesions confirmed by surgical pathology at Zhongshan Hospital of Fudan University from March 2024 to April 2025 were enrolled. According to the pathology results, they were classified into 46 cases of neoplastic polyps and 40 cases of non-neoplastic polyps. All patients underwent preoperative conventional ultrasound and contrast-enhanced ultrasound (CEUS). CEUS was used to observe the enhancement of gallbladder polypoid lesions in the arterial phase and their washout in the venous phase. SRUS was applied to evaluate the blood flow pattern and quantitative parameters of the lesions. The receiver operating characteristic curve was drawn to compare the diagnostic efficacy of single and combined modalities (conventional ultrasound, CEUS, and SRUS) for identifying neoplastic gallbladder polyps.

SRUS successfully detected intralesional blood flow in all 86 cases of gallbladder polypoid lesions (86/86, 100%). Neoplastic gallbladder polyps were predominantly characterized by grades Ⅱ and Ⅲ blood flow patterns (40/46, 86.96%). Significant differences were observed between neoplastic and non-neoplastic polyps in both blood flow pattern grade and quantitative parameters, including blood flow velocity, fractal dimension, vascular density, blood flow volume, and perfusion index (all P≤0.001). SRUS yielded an area under the curve (AUC) value of 0.82 in diagnosing neoplastic gallbladder polyps usingblood flow patterns of grade Ⅱ or higher, which was significantly higher than that of conventional ultrasound (polyp maximum diameter≥14.5 mm, AUC=0.71) and CEUS (early venous phase washout, AUC=0.70) (P=0.001 and 0.008, respectively). Three features derived from conventional ultrasound, CEUS, and SRUS were selected to construct three combination diagnostic models: conventional ultrasound plus CEUS, conventional ultrasound plus SRUS, and conventional ultrasound plus CEUS plus SRUS. Among these combination models, the diagnostic model integrating SRUS, conventional ultrasound, and CEUS showed the optimal performance (AUC=0.84).

The microvascular morphology and quantitative parameters derived from SRUS enable effective differentiation between neoplastic and non-neoplastic gallbladder polyps. Furthermore, the SRUS-based risk prediction model demonstrates significantly superior diagnostic performance compared with conventional ultrasound–based approaches, thereby substantially enhancing the diagnostic value of ultrasound in the evaluation of gallbladder polypoid lesions.

To validate the effectiveness of the 2023 American Association for the Study of Liver Diseases (AASLD) guidance for hepatocellular carcinoma (HCC) surveillance in high-risk populations, and to compare it with ultrasound (US) alone and US combined with alpha-fetoprotein (AFP) monitoring, in order to assess their clinical utility.

This prospective study retrospectively included 451 high-risk individuals who underwent HCC surveillance at the First Affiliated Hospital of Sun Yat-sen University between January 2018 and August 2022. All patients underwent US combined with AFP screening. The performance of US alone, US combined with AFP, and the 2023 AASLD guidance for HCC surveillance was evaluated and compared using the area under the receiver operating characteristic curve (AUC), sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV).

Among the 451 participants, 38 (38/451, 8%) were confirmed as having HCC. The 2023 AASLD guidance achieved an AUC of 0.87 for HCC detection, with a sensitivity of 95% and an NPV of 99%, which were significantly higher than those of US surveillance alone (AUC, 0.72, P＜0.001; sensitivity, 55%, P＜0.001; NPV, 96%, P=0.002) and US combined with AFP surveillance (AUC, 0.77, P=0.007; sensitivity, 68%, P=0.006; NPV, 97%, P=0.014). However, the specificity of the 2023 AASLD guidance (80%) was slightly lower than that of US surveillance alone and US combined with AFP surveillance (90%, P＜0.001; 86%, P=0.015), indicating a potential risk of over-referral among non-HCC individuals.

Although the specificity of the AASLD 2023 guidance is slightly reduced, its sensitivity and NPV for HCC detection are significantly higher than those of US alone and US combined with AFP. These findings suggest that the visualization score and recall strategy proposed in the 2023 AASLD guidance may have clinical utility for the early detection and management of HCC in high-risk populations. However, in clinical implementation, the risk of false-positive results should be carefully considered.

To investigate the utility of the DeepSeek large language model (LLM) in the structured generation of ultrasound reports and the automatic assessment of T-staging for gastric and rectal cancer.

A total of 121 ultrasound examination reports for gastric and rectal cancer, collected from Sichuan Cancer Hospital between January 2023 and December 2024, were included in this study. A structured template for gastric and rectal cancer ultrasound reports was developed by a team of senior sonographers. The DeepSeek R1 and V3 models were employed to extract structured information and assess T-staging. The performance of structured report generation was evaluated using recall, precision, and F1 score, while T-staging performance was assessed based on accuracy. Three physicians were invited to compare the reports generated by DeepSeek with the original reports to evaluate review efficiency and clinical usability.

Regarding structured information extraction, both DeepSeek R1 and V3 models achieved recall, precision, and F1 scores exceeding 0.9, with no statistically significant differences between the two (P＞0.05). In T-staging assessment, the DeepSeek R1 model (utilizing reasoning mode) achieved the highest accuracy of 76.86%, which was significantly superior to the 59.50% achieved by the DeepSeek V3 model (χ²=8.51, P＜0.05). Compared to the average time required to review original reports [(60.96±6.11) s/report], the review time for structured reports generated by DeepSeek R1 [(18.12±4.52) s/report] (t=60.38; P＜0.001) and DeepSeek V3 [(17.15±2.60) s/report] (t=71.98; P＜0.001) was significantly shortened. The 5-point Likert scale evaluation showed that the score for the original reports was 3 (3, 3), while the scores for the DeepSeek R1 and V3 reports were 1 (1, 2) (Z=-9.72; P＜0.001) and 1 (1, 2) (Z=-9.95; P＜0.001), respectively, indicating a statistically significant difference.

The DeepSeek large language models, particularly the R1 version, can effectively extract structured information from gastric and rectal cancer ultrasound reports and demonstrates high accuracy in T-staging assessment. The generated reports contribute to improved review efficiency and possess the potential to assist in clinical decision-making.

To evaluate the diagnostic value and inter-observer agreement of the 2020 adapted Bosniak cyst categorization proposed based on ultrasound for cystic renal masses (CRMs).

This retrospective study enrolled patients with surgically resected CRMs and postoperative pathology confirmation from the First Medical Center of the Chinese PLA General Hospital between October 2018 and October 2024. All patients underwent both conventional renal ultrasound and contrast enhanced ultrasound (CEUS) within one month before surgery. A total of 187 patients with 188 lesions were included in the final analysis. Four sonographers—two senior and two junior—independently reviewed the conventional ultrasound and CEUS images and classified each lesion according to the 2020 Bosniak criteria. Interobserver agreement was evaluated using Fleiss' kappa (κ) and weighted Cohen's κ, with subgroup analyses performed based on observer experience. Using pathological findings as the reference standard, receiver operating characteristic (ROC) curve analysis was conducted to assess and compare the diagnostic performance of the Bosniak classification based on conventional ultrasound versus CEUS.

Among the 188 lesions, 125 were benign and 63 were malignant. The area under the ROC curve (AUC) of conventional ultrasound and CEUS for differentiating benign and malignant CRMs was 0.872 and 0.928, respectively. The diagnostic specificity of CEUS was significantly higher than that of conventional ultrasound (90.40% vs 82.40%, P＜0.05). After excluding 88 cystic lesions (Bosniak categories Ⅰ and Ⅱ renal cysts), the diagnostic specificity of both conventional ultrasound and CEUS decreased markedly; however, CEUS still demonstrated significantly higher specificity than conventional ultrasound (67.57% vs 40.54%, P＜0.05). Interobserver agreement for CEUS classification among the four sonographers was significantly higher than that for conventional ultrasound (κ: 0.790 vs 0.565, P＜0.001). For conventional ultrasound, interobserver agreement was higher between the two senior sonographers than between the two junior sonographers (weighted κ=0.865 vs 0.746, P＜0.001). In contrast, for CEUS, there was no statistically significant difference in interobserver agreement between senior and junior sonographers (weighted κ=0.951 vs 0.921, P＞0.05). In the subgroup of CRMs including only categories Ⅱ, ⅡF, and Ⅲ, the interobserver agreement for conventional ultrasound and CEUS was κ=0.188 and κ=0.600, respectively (P＜0.001).

For the 2020 adapted Bosniak cyst categorization proposed based on ultrasound, the CEUS version demonstrates superior diagnostic performance compared to the conventional ultrasound approach, evidenced by higher interobserver agreement and improved specificity in differentiating benign from malignant CRMs.

To evaluate the clinical value of intraplaque contrast agent diffusion direction and intraplaque neovascularization (IPN) grading on contrast-enhanced ultrasound (CEUS) in assessing carotid plaque vulnerability and predicting the risk of symptomatic stroke, and to construct a multiparametric predictive model to improve the identification of high-risk individuals.

A total of 84 patients who underwent carotid endarterectomy (CEA) at Suzhou Municipal Hospital between January 2023 and December 2024 were prospectively enrolled. According to the occurrence of symptomatic stroke events or not, the patients were divided into a symptomatic stroke group (n=48) and an asymptomatic group (n=36). Preoperative conventional ultrasound and CEUS were performed to assess plaque morphology, contrast agent diffusion direction, and IPN grade. Postoperative pathological specimens were collected to evaluate microvessel density (MVD) and the proportion of immature microvessels (MVR). Spearman rank correlation analysis was used to evaluate the correlation between CEUS features and pathological characteristics. Multivariate logistic regression was used to identify independent risk factors for symptomatic stroke, and receiver operating characteristic (ROC) curves were plotted to assess the discriminative performance of different predictive models.

IPN grade was significantly positively correlated with both MVD (r=0.782, P＜0.001) and MVR (r=0.508, P＜0.001). Plaques showing an "inside-out" contrast agent diffusion pattern exhibited significantly higher incidences of fibrous cap rupture and intraplaque hemorrhage compared with non-inside-out diffusion plaques (87.50% vs 55.56% and 87.50% vs 66.67%, respectively; P＜0.05). Logistic regression revealed that IPN grade 3 (odds ratio [OR]= 4.36, 95% confidence interval [CI]: 1.50-12.67, P＜0.01) and "inside-out" diffusion pattern (OR=3.88, 95%CI: 1.43-10.57, P＜0.01) were independent risk factors for symptomatic stroke events. The area under the ROC curve (AUC) of the combined predictive model was 0.806, which was higher than that of either IPN grade 3 (AUC=0.751) or "inside-out" diffusion pattern alone (AUC=0.673).

CEUS-based assessment of contrast agent diffusion direction and IPN grading can effectively reflect carotid plaque vulnerability and demonstrates an appreciated ability to predict the risk of symptomatic stroke. The multiparametric model based on these two indicators shows good discriminative performance in preoperative risk assessment and screening of high-risk stroke populations, indicating potential clinical applicability.

To summarize the echocardiographic characteristics of congenital coronary artery fistula (CAF) complicated with giant coronary artery aneurysm in children and assess its utility in diagnosis and follow-up.

A retrospective study was conducted on 127 children with congenital CAF involving a single coronary artery and draining into a single cardiac chamber, who were diagnosed by echocardiography at Beijing Children's Hospital Affiliated to Capital Medical University from April 2009 to April 2025. According to the presence of giant coronary artery aneurysm or not, they were divided into a giant coronary artery aneurysm group and a non-giant coronary artery aneurysm group. Univariate analysis was used to compare baseline data between the two groups, and multivariate Logistic regression analysis was performed to identify factors associated with the development of giant coronary artery aneurysm. Postoperative echocardiographic follow-up was performed for children in the giant coronary artery aneurysm group who underwent surgery to dynamically observe the surgical outcomes.

Among the 127 children included, 49 (49/127, 38.6%) were in the giant coronary artery aneurysm group, including 23 cases (23/49, 46.9%) of right CAF. The drainage sites were the right ventricle in 28 cases (28/49, 57.1%), right atrium in 17 (17/49, 34.7%), left ventricle in 3 (3/49, 6.1%), and left atrium in 1 (1/49, 2.0%). Fourteen children (14/49, 28.6%) had other cardiac malformations. Sixteen children (16/49, 32.7%) underwent transcatheter closure or surgical treatment. The non-giant coronary artery aneurysm group consisted of 78 children (78/127, 61.4%), including 21 cases (21/78, 26.9%) of right CAF. The drainage sites were the right ventricle in 51 cases (51/78, 65.4%), left ventricle in 12 (12/78, 15.4%), right atrium in 10 (10/78, 12.8%), and left atrium in 5 (5/78, 6.4%). Twenty-nine children (29/78, 37.2%) had other cardiac malformations. None of the children in this group underwent surgical treatment. The rate of missed diagnosis or misdiagnosis in the 49 children with giant coronary artery aneurysm was approximately 12.2% (6/49). Logistic regression analysis showed that right CAF (odds ratio [OR]=0.362, 95% confidence interval [CI]: 0.162-0.810) and fistula draining into the atrium (OR=3.166, 95%CI: 1.320-7.594) were independent risk factors for the formation of giant coronary artery aneurysm. Echocardiographic findings of CAF complicated with giant coronary artery aneurysm mainly included giant aneurysmal dilation of the affected coronary artery, with no significant dilation of the contralateral coronary artery. All 16 surgically treated children showed good corrective outcomes, and no severe complications such as coronary artery stenosis were found during follow-up.

Right CAF and CAF draining into the atrium may be independent risk factors for the formation of giant coronary artery aneurysm. CAF complicated with giant coronary artery aneurysm has typical echocardiographic manifestations, and familiarity with these imaging features can reduce missed and misdiagnosis. Although the short-term benefits of surgical intervention in children are not very significant, compared with adults, it may reduce the risk of serious complications.

To evaluate the value of lung ultrasonography (LUS), right heart echocardiographic parameters, and related clinical indicators in evaluating severe pulmonary fibrosis (PF).

This is a cross-sectional study that included 107 patients with PF treated at Tangdu Hospital of Air Force Medical University from March 2024 to January 2025, among whom 49 were divided into a non-severe PF group and 58 into a severe PF group. All patients underwent LUS assessment and transthoracic echocardiography, and relevant clinical indicators were also collected. The clinical data and ultrasonic characteristic parameters of the two groups were compared. Logistic regression analysis was used to identify the independent risk factors for severe PF. The receiver operating characteristic (ROC) curve was used to assess the diagnostic efficacy of each independent risk factor and their combination for severe PF.

There were statistically significant differences in age, proportion of males, and smoking history between the two groups (all P＜0.05). Compared with the non-severe group, erythrocyte sedimentation rate, cytokeratin 19 fragment level, LUS score, and pleural line thickness were significantly increased in patients with severe PF (all P＜0.05), while the proportion of patients with impaired right ventricular relaxation function, tricuspid annular plane systolic excursion/systolic pulmonary arterial pressure ratio＜0.55 mm/mmHg, and tricuspid regurgitation velocity＞2.8 m/s was higher in the severe group (P＜0.05). In addition, the severe group also showed characteristics of greater pulmonary artery diameter and increased pulmonary vascular resistance. Multivariate Logistic regression analysis showed that smoking (odds ratio [OR]=17.48, 95% confidence interval [CI]: 1.47-207.84, P=0.024), impaired right heart relaxation function (OR=9.52, 95%CI: 1.25-72.80, P=0.03), pulmonary artery diameter (OR=1.72, 95%CI: 1.11-2.68, P=0.016), LUS score (OR=1.17, 1.01-1.35, P=0.034), and pleural line thickness (OR=21.22, 95%CI: 4.28-105.21, P<0.001) were independent risk factors for severe PF. The area under the ROC curve (AUC) of the combined model incorporating the above five factors was 0.97, and its diagnostic efficiency was better than that of single indicators.

This study confirms that smoking, impaired right heart relaxation function, greater pulmonary artery dilation, high LUS score, and pleural line thickening are closely related to severe PF. The combination of LUS and right heart echocardiography can be used as a new method for the assessment and dynamic monitoring of PF severity.