ObjectiveDiscriminating atypical hepatocellular carcinoma (HCC) from other malignancies in liver nodules classified as Liver Imaging Reporting and Data System category M (LR-M) remains a significant diagnostic challenge on conventional ultrasound examination. The LR-M category, originally intended to capture non-HCC malignancies, paradoxically contains up to 63% of atypical HCCs that deviate from classic enhancement patterns, leading to potential misdiagnosis and suboptimal treatment planning. While deep learning has shown promise in HCC diagnosis, most existing models rely exclusively on single-modality ultrasound, overlooking the diagnostic benefits of integrating complementary information from multiple imaging sources. To address this gap, we propose a novel attention-weighted tri-modal ultrasound network (TUS-Net) that integrates contrast-enhanced ultrasound (CEUS), B-mode ultrasound (BUS), and time-intensity curves (TICs) to improve diagnostic accuracy for these clinically challenging lesions. MethodsOur framework incorporates a three-dimensional convolutional neural network (C3D) backbone to extract spatiotemporal features from CEUS videos, capturing dynamic vascular patterns critical for lesion characterization. To effectively fuse complementary modalities, we introduce a dual-channel feature fusion module (DCFFM) that adaptively combines features from CEUS and BUS through channel-wise attention mechanisms, allowing the model to dynamically weigh the contribution of each modality based on diagnostic relevance. Additionally, we propose a temporal intensity feature fusion module (TIFFM) that leverages quantitative hemodynamic information from TICs to guide the model’s attention toward diagnostically critical temporal phases, such as arterial wash-in and portal venous washout. The model is further enhanced by automated lesion localization using YOLOX and class activation mapping for interpretability, ensuring that predictions align with clinically meaningful imaging features. ResultsEvaluated on a tri-modal ultrasound dataset comprising 161 patients with pathologically confirmed LR-M nodules (131 atypical HCC and 30 non-HCC malignancies), our model achieved an accuracy of 86.83%, a sensitivity of 92.50%, a specificity of 75.50%, and an AUC of 89.32% in screening atypical HCC. Compared to single-modality baselines, TUS-Net demonstrated superior specificity, a clinically critical metric given the higher risk associated with misclassifying non-HCC malignancies. Ablation studies confirmed the contribution of each module, with the full model outperforming both standard C3D and 3D ResNet backbones integrated with attention mechanisms. A reader study involving junior and senior radiologists further validated the clinical utility of AI assistance, showing consistent improvements in specificity and inter-reader consistency, particularly for less experienced clinicians. ConclusionThese results surpass existing benchmark models and demonstrate the potential of our approach to enhance diagnostic precision in clinically specific cases. By intelligently fusing multi-modal ultrasound data with attention-guided mechanisms, TUS-Net offers a reliable and interpretable tool that holds promise for improving the non-invasive diagnosis of atypical HCC in challenging LR-M liver nodules.

ObjectiveDiscriminating atypical hepatocellular carcinoma (HCC) from other malignancies in liver nodules classified as Liver Imaging Reporting and Data System category M (LR-M) remains a significant diagnostic challenge on conventional ultrasound examination. The LR-M category, originally intended to capture non-HCC malignancies, paradoxically contains up to 63% of atypical HCCs that deviate from classic enhancement patterns, leading to potential misdiagnosis and suboptimal treatment planning. While deep learning has shown promise in HCC diagnosis, most existing models rely exclusively on single-modality ultrasound, overlooking the diagnostic benefits of integrating complementary information from multiple imaging sources. To address this gap, we propose a novel attention-weighted tri-modal ultrasound network (TUS-Net) that integrates contrast-enhanced ultrasound (CEUS), B-mode ultrasound (BUS), and time-intensity curves (TICs) to improve diagnostic accuracy for these clinically challenging lesions. MethodsOur framework incorporates a three-dimensional convolutional neural network (C3D) backbone to extract spatiotemporal features from CEUS videos, capturing dynamic vascular patterns critical for lesion characterization. To effectively fuse complementary modalities, we introduce a dual-channel feature fusion module (DCFFM) that adaptively combines features from CEUS and BUS through channel-wise attention mechanisms, allowing the model to dynamically weigh the contribution of each modality based on diagnostic relevance. Additionally, we propose a temporal intensity feature fusion module (TIFFM) that leverages quantitative hemodynamic information from TICs to guide the model’s attention toward diagnostically critical temporal phases, such as arterial wash-in and portal venous washout. The model is further enhanced by automated lesion localization using YOLOX and class activation mapping for interpretability, ensuring that predictions align with clinically meaningful imaging features. ResultsEvaluated on a tri-modal ultrasound dataset comprising 161 patients with pathologically confirmed LR-M nodules (131 atypical HCC and 30 non-HCC malignancies), our model achieved an accuracy of 86.83%, a sensitivity of 92.50%, a specificity of 75.50%, and an AUC of 89.32% in screening atypical HCC. Compared to single-modality baselines, TUS-Net demonstrated superior specificity, a clinically critical metric given the higher risk associated with misclassifying non-HCC malignancies. Ablation studies confirmed the contribution of each module, with the full model outperforming both standard C3D and 3D ResNet backbones integrated with attention mechanisms. A reader study involving junior and senior radiologists further validated the clinical utility of AI assistance, showing consistent improvements in specificity and inter-reader consistency, particularly for less experienced clinicians. ConclusionThese results surpass existing benchmark models and demonstrate the potential of our approach to enhance diagnostic precision in clinically specific cases. By intelligently fusing multi-modal ultrasound data with attention-guided mechanisms, TUS-Net offers a reliable and interpretable tool that holds promise for improving the non-invasive diagnosis of atypical HCC in challenging LR-M liver nodules.

ObjectiveTo screen key genes associated with neuroblastoma （NB） diagnosis and prognosis using the Gene Expression Omnibus （GEO） database， and to investigate the expression and clinical significance of nucleoporin 93 （NUP93） in NB tissues. MethodsNB gene chip data （GSE73517， GSE49710， GSE19274） were retrieved from the GEO database. Differentially expressed genes （DEGs） commonly upregulated in high-risk groups were screened. The R2 database was then used to assess the prognostic value of DEGs that were commonly upregulated in the MYCN amplification group. Finally， NUP93 expression levels in the tissues from 60 NB， 25 ganglioneuroblastoma （GNB）， and 26 ganglioneuroma （GN） cases were measured by immunohistochemistry . ResultsTwenty-five DEGs were identified as commonly upregulated in high-risk groups. Among these， 10 genes （SIVA1， NUP93， STIP1， LSM4， RAI14， MYOZ3， KNTC1， TNFRSF10B， TACC3 and CEP152） showed significantly higher expression in MYCN-amplified subgroups （P＜0.05）. Survival analysis revealed that high NUP93 expression was associated with shorter overall survival （HR = 4.0， 95% CI： 3.0，5.3， P = 1.80 × 10⁻³⁴）. Immunohistochemistry results revealed that NUP93 expression in NB tissues was significantly higher than in GNB and GN tissues （P＜0.001）. NUP93 expression was positively correlated with high mitosis-karyorrhexis index （MKI； P=0.040）， poor differentiation （P＜0.001）， and MYCN expression （r_s = 0.793， P ＜0.001）. ConclusionsHigh expression of NUP93 is associated with high MKI and poor differentiation， and predicts unfavorable prognosis in patients with NB， suggesting it may promote tumor progression by regulating MYCN. NUP93 has the potential to be a novel diagnostic biomarker and therapeutic target for NB.

ObjectivePrimary liver cancer, predominantly hepatocellular carcinoma (HCC), is a significant global health issue, ranking as the sixth most diagnosed cancer and the third leading cause of cancer-related mortality. Accurate and early diagnosis of HCC is crucial for effective treatment, as HCC and non-HCC malignancies like intrahepatic cholangiocarcinoma (ICC) exhibit different prognoses and treatment responses. Traditional diagnostic methods, including liver biopsy and contrast-enhanced ultrasound (CEUS), face limitations in applicability and objectivity. The primary objective of this study was to develop an advanced, light-weighted classification network capable of distinguishing HCC from other non-HCC malignancies by leveraging the automatic analysis of brightness changes in CEUS images. The ultimate goal was to create a user-friendly and cost-efficient computer-aided diagnostic tool that could assist radiologists in making more accurate and efficient clinical decisions. MethodsThis retrospective study encompassed a total of 161 patients, comprising 131 diagnosed with HCC and 30 with non-HCC malignancies. To achieve accurate tumor detection, the YOLOX network was employed to identify the region of interest (ROI) on both B-mode ultrasound and CEUS images. A custom-developed algorithm was then utilized to extract brightness change curves from the tumor and adjacent liver parenchyma regions within the CEUS images. These curves provided critical data for the subsequent analysis and classification process. To analyze the extracted brightness change curves and classify the malignancies, we developed and compared several models. These included one-dimensional convolutional neural networks (1D-ResNet, 1D-ConvNeXt, and 1D-CNN), as well as traditional machine-learning methods such as support vector machine (SVM), ensemble learning (EL), k-nearest neighbor (KNN), and decision tree (DT). The diagnostic performance of each method in distinguishing HCC from non-HCC malignancies was rigorously evaluated using four key metrics: area under the receiver operating characteristic (AUC), accuracy (ACC), sensitivity (SE), and specificity (SP). ResultsThe evaluation of the machine-learning methods revealed AUC values of 0.70 for SVM, 0.56 for ensemble learning, 0.63 for KNN, and 0.72 for the decision tree. These results indicated moderate to fair performance in classifying the malignancies based on the brightness change curves. In contrast, the deep learning models demonstrated significantly higher AUCs, with 1D-ResNet achieving an AUC of 0.72, 1D-ConvNeXt reaching 0.82, and 1D-CNN obtaining the highest AUC of 0.84. Moreover, under the five-fold cross-validation scheme, the 1D-CNN model outperformed other models in both accuracy and specificity. Specifically, it achieved accuracy improvements of 3.8% to 10.0% and specificity enhancements of 6.6% to 43.3% over competing approaches. The superior performance of the 1D-CNN model highlighted its potential as a powerful tool for accurate classification. ConclusionThe 1D-CNN model proved to be the most effective in differentiating HCC from non-HCC malignancies, surpassing both traditional machine-learning methods and other deep learning models. This study successfully developed a user-friendly and cost-efficient computer-aided diagnostic solution that would significantly enhances radiologists’ diagnostic capabilities. By improving the accuracy and efficiency of clinical decision-making, this tool has the potential to positively impact patient care and outcomes. Future work may focus on further refining the model and exploring its integration with multimodal ultrasound data to maximize its accuracy and applicability.

OBJECTIVE To promote the application of drones in emergency rescue and related fields， expand “low-altitude+ medical” rescue services， and advance the standardization of “low-altitude+medical” distribution services. METHODS The Consensus on Low-altitude Transport and Delivery Services for Emergency Medicines via Drones （2025 Edition） （hereinafter referred to as the Consensus） was jointly initiated by the Division of Therapeutic Drug Monitoring， Chinese Pharmacological Society and the Expert Committee on Precision Medication of the Guangdong Pharmaceutical Association. Guangzhou Red Cross Hospital served as the leading unit， organizing 53 multidisciplinary experts nationwide to participate in drafting and reviewing. A nominal group technique was employed to discuss and finalize the consensus outline， resulting in a preliminary draft. Delphi method was employed， and 11 external review experts were invited to conduct the evaluation. After the experts’ opinions were analyzed and integrated， the Consensus was finalized. RESULTS & CONCLUSIONS The finalized Consensus includes its purpose， principles， and applicable scenarios， basic requirements， and operational procedures for low-altitude transport and delivery of emergency medications； distribution requirements and precautions for controlled substances， fragile medications， and temperature-sensitive medications； and recommendations for emergency medications supplies suitable for the low-altitude transportation and distribution. The release of this Consensus is expected to provide guidance and support for the standardization of “low-altitude+medical” distribution services and the application of low-altitude economy in the healthcare sector.

Neurogastroenterology and motility disorders of the upper gastrointestinal (GI) tract represent a complex and heterogeneous group of conditions that involve the interaction between the GI tract and the central nervous system. They constitute a significant number of outpatient gastroenterology visits, resulting in a high healthcare burden. These disorders often occur in the absence of identifiable structural causes on routine endoscopy and radiological imaging. A more targeted approach in the assessment of functional GI disorders is increasingly being integrated into routine clinical practice, given the recent advancements in technology and physiologic testing. When used in the appropriate clinical context, these tests not only elucidate the physiological basis for the patient's symptoms, but also prevent inappropriate treatment and repeated investigations. This review aims to summarise the advances in clinically available diagnostic tools for the evaluation of upper GI functional disorders.
Humans ; Gastrointestinal Diseases/physiopathology* ; Upper Gastrointestinal Tract/physiopathology* ; Gastrointestinal Motility ; Endoscopy, Gastrointestinal

Objective To investigate the correlation between G protein-coupled receptor family C group 6 member A(GPRC6A)gene polymorphisms and their expression and pulmonary infections in elderly patients with chronic heart failure(CHF).Methods 138 elderly CHF patients admitted to the Xianyang First People's Hospital from January 2021 to January 2024 were selected as the research subjects,and were divided into an infected group(n=42)and an uninfected group(n=96)based on their lung infection status.Polymerase chain reaction(PCR)was used to detect polymorphisms at the rs6901250 and rs1606365 loci of the GPRC6A gene.The allele and genotype frequency distributions of the infected and uninfected groups were compared.Logistic regression modeling was used to analyze the s6901250 and rs1606365 loci under three genetic models(co-dominant,dominant and reces-sive)and lung infections in elderly patients with CHF.Real-time fluorescence quantitative PCR(RT-qPCR)method was used to detect the expression level of GPRC6A gene.The predictive value of the mRNA expression level of the GPRC6A gene for the development of pulmonary infections in elderly patients with CHF was analyzed by applying the receiver operator characteristic(ROC)curve.Results The distribution of genotypes at loci rs6901250 and rs1606365 of the GPRC6A gene in both the infected and uninfected groups of the lungs of elderly CHF patients conformed to the Hardy-Weinberg equilibrium law(χ2=0.199～0.376,all P>0.05),which was representative of the population.Compared with the uninfected group,the frequency of allele A at locus rs6901250(57.14%vs 41.67%)was significantly higher in the infected group,Allele G(54.76%vs.37.50%)and genotype GG(14.06%vs 29.99%)frequencies were significantly higher at locus rs1606365,and the differences were statistically significant(χ2=5.628,7.114,6.849,all P<0.05).At locus rs6901250,in the co-dominant model(GG vs AA)and the dominant model(GA+AA vs GG),the elderly CHF patients with AA genotype the risk of lung infection was higher than that of GG genotype(OR=1.753,1.546,all P<0.05);.rs1606365 locus showed that the risk of lung infection was higher than that of CC genotype in el-derly CHF patients with GG genotype under all three genetic models of co-dominant model(CC vs GG),dominant model(CG+GG vs CC)and recessive model(CG+CC vs GG)(OR=1.833,1.741,0.695,all P<0.05).The mRNA expression level of GPR-C6A gene in the lung-infected group of elderly CHF patients(1.43±0.35)was significantly higher than that in the uninfected group(1.02±0.21),and the difference was statistically significant(t=8.515,P<0.001).The results of the ROC curve analysis showed that the GPRC6A gene expression level predicted lung infection in elderly CHF patients with an AUC value of 0.895,a cut-offvalue of 1.37,and sensitivity and specificity of 85.7%and 66.7%,respectively.Conclusion The AA genotype at the rs6901250 locus and the GG genotype at the rs1606365 locus of the GPRC6A gene increased the risk of developing lung infec-tions in elderly patients with CHF.MRNA expression levels of the GPRC6A gene were elevated in the infected group,and its ex-pression level could be used as a predictive indicator for the development of lung infections in elderly patients with CHF.

Objective:To investigate the clinical and neonatal outcomes of nonobstructive azoospermia (NOA) patients undergoing a second microdissection testicular sperm extraction (microTESE) combined with intracytoplasmic sperm injection (ICSI).Methods:A retrospective self-controll study and a cohort study were conducted in 29 NOA patients who underwent the second microTESE-ICSI treatments at the Assisted Reproduction Center of Northwest Women's and Children's Hospital between October 2013 and December 2024, with successful sperm retrieval in the first microTESE for all included patients. The sperm retrieval rate (SRR), the two-pronuclei (2PN) rate, the high-quality embryo rate, the transferable embryo rate, the blastocyst formation rate, the pregnancy and neonatal outcomes between the two microTESE-ICSI cycles were analyzed. According to whether sperm was detected during the second microTESE, the patients were divided into detected sperm group ( n=22) and without detected sperm group ( n=7). The surgical interval, preoperative hormone levels, testicular volume, and patient pathological type were compared between the two groups. Patients who had achieved at least one biochemical pregnancy in the first cycle were classified as the pregnancy group, while those who had not achieved a biochemical pregnancy were classified as the non-pregnancy group, and the pregnancy outcomes of patients during the second microTESE-ICSI treatment were compared between the two groups. Previous surgical recordings were also analyzed. Results:The SRR of the second microTESE was 75.9% (22/29). There were no statistically significant differencs in surgical interval, preoperative hormones level, testicular volume, and patient pathological type between detected sperm group and without detected group during the second examination (all P>0.05). SRR showed no difference between idiopathic and non-idiopathic NOA ( P>0.05). 2PN rate, high-quality embryo rate, transferable embryo rate, blastocyst formation rate did not differ statistically between the two cycles (all P>0.05). Among patients completing follow-up in the second microTESE-ICSI cycle, cumulative pregnancy and live birth rates after all embryos transfer were 52.4% (11/21) and 42.9% (9/21), respectively. Further analysis revealed that patients in pregnancy group [75.0% (6/8)] had significantly higher cumulative live birth rate than that in non-pregnancy group [23.1% (3/13), P=0.032] in the second cycle. No birth defects were observed in neonates. Surgical video review suggested that excessive microdissection and low-magnification electrocautery hemostasis might increase intratesticular scarring. Conclusion:For NOA patients with failed initial microTESE-ICSI, the second treatment can effectively retrieve sperm and achieve favorable pregnancy and neonatal outcomes. Patients with prior biochemical pregnancy may benefit from higher live birth rate in subsequent cycles.

Objective To investigate the correlation between G protein-coupled receptor family C group 6 member A(GPRC6A)gene polymorphisms and their expression and pulmonary infections in elderly patients with chronic heart failure(CHF).Methods 138 elderly CHF patients admitted to the Xianyang First People's Hospital from January 2021 to January 2024 were selected as the research subjects,and were divided into an infected group(n=42)and an uninfected group(n=96)based on their lung infection status.Polymerase chain reaction(PCR)was used to detect polymorphisms at the rs6901250 and rs1606365 loci of the GPRC6A gene.The allele and genotype frequency distributions of the infected and uninfected groups were compared.Logistic regression modeling was used to analyze the s6901250 and rs1606365 loci under three genetic models(co-dominant,dominant and reces-sive)and lung infections in elderly patients with CHF.Real-time fluorescence quantitative PCR(RT-qPCR)method was used to detect the expression level of GPRC6A gene.The predictive value of the mRNA expression level of the GPRC6A gene for the development of pulmonary infections in elderly patients with CHF was analyzed by applying the receiver operator characteristic(ROC)curve.Results The distribution of genotypes at loci rs6901250 and rs1606365 of the GPRC6A gene in both the infected and uninfected groups of the lungs of elderly CHF patients conformed to the Hardy-Weinberg equilibrium law(χ2=0.199～0.376,all P>0.05),which was representative of the population.Compared with the uninfected group,the frequency of allele A at locus rs6901250(57.14%vs 41.67%)was significantly higher in the infected group,Allele G(54.76%vs.37.50%)and genotype GG(14.06%vs 29.99%)frequencies were significantly higher at locus rs1606365,and the differences were statistically significant(χ2=5.628,7.114,6.849,all P<0.05).At locus rs6901250,in the co-dominant model(GG vs AA)and the dominant model(GA+AA vs GG),the elderly CHF patients with AA genotype the risk of lung infection was higher than that of GG genotype(OR=1.753,1.546,all P<0.05);.rs1606365 locus showed that the risk of lung infection was higher than that of CC genotype in el-derly CHF patients with GG genotype under all three genetic models of co-dominant model(CC vs GG),dominant model(CG+GG vs CC)and recessive model(CG+CC vs GG)(OR=1.833,1.741,0.695,all P<0.05).The mRNA expression level of GPR-C6A gene in the lung-infected group of elderly CHF patients(1.43±0.35)was significantly higher than that in the uninfected group(1.02±0.21),and the difference was statistically significant(t=8.515,P<0.001).The results of the ROC curve analysis showed that the GPRC6A gene expression level predicted lung infection in elderly CHF patients with an AUC value of 0.895,a cut-offvalue of 1.37,and sensitivity and specificity of 85.7%and 66.7%,respectively.Conclusion The AA genotype at the rs6901250 locus and the GG genotype at the rs1606365 locus of the GPRC6A gene increased the risk of developing lung infec-tions in elderly patients with CHF.MRNA expression levels of the GPRC6A gene were elevated in the infected group,and its ex-pression level could be used as a predictive indicator for the development of lung infections in elderly patients with CHF.

Objective:To compare bowel function 12 months after surgery between side-to-end anastomosis (SEA) and end-to-end anastomosis (EEA) groups of patients who had undergone rectal cancer resection.Methods:This single-center, prospective, open-label, phase III randomized controlled trial was approved by the Ethics Committee of Peking University People's Hospital (2018PHB040-01) and registered at ClinicalTrials. org (NCT03669237). Inclusion criteria were as follows: (1) histologically confirmed rectal adenocarcinoma; (2) tumor located 0 to 12 cm from the anal verge; (3) age≥18 years; and (4) planned R0 resection with primary reconstruction. Exclusion criteria included: (1) emergency surgery; (2) cognitive impairment; (3) non-primary anastomosis; (4) history of left-sided colonic or anorectal surgery; and (5) preexisting chronic defecation dysfunction. Eligible rectal cancer patients scheduled for elective sphincter-preserving surgery at Peking University People's Hospital were prospectively enrolled between October 2018 and March 2021 and randomly assigned to either the EEA group or the SEA group via computer-generated numbers prior to entering the operating room. All patients underwent standard radical tumor resection. Bowel function was evaluated by the low anterior resection syndrome (LARS) questionnaire. It consists of five single-choice questions and yields a total score ranging from 0 to 42. Defecation function is categorized into three levels: no LARS (0-20 points), minor LARS (21-29 points), and major LARS (30-42 points). The primary endpoint was the LARS score 12 months after surgery. Secondary endpoints included LARS scores from 1 to 11 months and during long-term follow-up(>12 months). The final follow-up was completed in July 2022. All randomized patients were included in the intention-to-treat set (ITTS). The full analysis set (FAS) was defined as ITTS patients with valid outcome data. All primary statistical analyses were performed in the FAS, and results were further compared in the per-protocol set (PPS) based on the actual treatment received.Results:A total of 323 patients underwent eligibility assessment, of whom 71 did not meet the inclusion criteria and 52 declined to participate. Ultimately, 200 patients were randomized. Median age was 64 years and 85 were women. The SEA and EEA groups comprised 102 and 98 patients, respectively. A total of 181 patients (90.5%) were included in the FAS, and 170 (85.0%) were included in the PPS. Among these, the 12-month LARS score was evaluated in 178 patients (98.3%) in the FAS and in 167 (98.2%) in the PPS. Median LARS score at 1–12 months were significantly lower in the SEA group in both the FAS dataset [12 months:8 (interquartile range [IQR], 0–22) vs. 14 (IQR, 8–29); Z=2.687, P=0.007] and the PPS dataset [12 months: 8 (IQR, 0–22) vs. 14 (IQR, 6–29); Z=2.543, P=0.011]. During long-term follow-up, the median LARS score was also significantly lower in the SEA group in the FAS dataset [2 (IQR, 0–4) vs. 11 (IQR, 2–23); Z=2.968, P=0.003] and the PPS dataset [2 (IQR, 0–14) vs. 11 (2, 27); Z=2.687, P=0.007]. Conclusion:Compared with the EEA group, bowel function was superior in the SEA group 1 year after surgery and during long-term follow-up.