ObjectiveThe widespread adoption of portable fundus cameras for primary care and community screening is hindered by limitations in current autofocus(AF) technologies. Image-based methods relying on sharpness evaluation require iterative searches, resulting in slow convergence, while projection-based techniques are susceptible to optical artifacts and calibration errors. To address these challenges, this study introduces a novel AF system based on direct wavefront sensing, designed to deliver simultaneous high speed, high precision, and operational robustness within the compact form factor essential for portable ophthalmic devices. MethodsOur approach fundamentally reimagines the AF process by directly measuring the ocular wavefront aberration. We developed a custom portable fundus camera integrating a miniaturized Shack-Hartmann wavefront sensor (SHWS) into the optical path. An 850 nm laser diode projects a point source onto the retina via oblique illumination to minimize corneal reflections. Light scattered from this spot carries the eye’s refractive error through the imaging optics and is directed to the SHWS, positioned at a plane optically conjugate to the primary color CMOS imaging sensor. A microlens array within the SHWS samples the incident wavefront, generating a pattern of focal spots on a CCD. Real-time centroid analysis of these spots provides a map of local wavefront slopes. These measurements are processed through a singular value decomposition (SVD) algorithm to fit a Zernike polynomial basis set, enabling real-time reconstruction of the wavefront phase. The defocus component (S) is extracted from the second-order Zernike coefficients, providing a direct, quantitative measure of the refractive error in diopters. This value serves as a precise error signal in a closed-loop control system, which commands a voice-coil actuated focusing lens to its null position in a single, deterministic step, eliminating the need for iterative search algorithms. ResultsComprehensive evaluation demonstrated the system’s high performance. Testing on a calibrated model eye (OEMI-7) established a highly linear relationship between the computed defocus S and the focusing lens position across a ±20 Diopter (D) compensation range, achievable within a 5 mm mechanical travel. The system achieved a focusing precision of 0.08 D, corresponding to an 18-fold improvement over a conventional projection spot-size method tested under identical conditions. The total focus acquisition time, encompassing wavefront measurement, computation, and lens actuation, averaged under 0.5 s. Clinical validation with 25 human volunteers (50 eyes, refractive range -15 D to +10 D) confirmed practical efficacy. The wavefront-sensing AF succeeded in 92% of attempts with a mean time of 0.5 s, substantially outperforming a projection-based benchmark which achieved only a 32% success rate with an average time of 4.25 s. The system provided instantaneous directional guidance and maintained stability during minor ocular movements. Objective assessment of image quality, via amplitude contrast of retinal vasculature, showed consistent and significant enhancement following AF correction across the entire tested diopter range. ConclusionThis work successfully implements and validates a direct wavefront-sensing autofocus paradigm for portable fundus cameras. By directly quantifying and compensating for the optical defocus aberration, this method bypasses the fundamental limitations of image-processing and projection-based techniques, enabling rapid, precise, and deterministic diopter compensation. The developed system delivers an exceptional combination of a wide operational range (±20 D), high accuracy (0.08 D), fast convergence (0.5 s), and a compact physical footprint. This technology provides a practical and high-performance focusing solution capable of enhancing the reliability, throughput, and diagnostic utility of portable retinal imaging in large-scale screening applications. Future efforts will be directed towards system cost optimization and performance adaptation for diverse ocular conditions.

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.

background Current assessment of noise-induced hearing loss relies on the hearing threshold statistical distribution table of ISO 7029-2017 standard (ISO 7029), which is based on foreign population data and lacks a hearing threshold distribution table derived from pure-tone audiometry data of the Chinese population, hindering accurate evaluation of hearing loss in this group. Objective To establish a statistical distribution table of hearing threshold level (HTL) for otologically normal Chinese adults and to provide a scientific basis for revising the diagnostic criteria of occupational noise-induced deafness in China. Methods A total of 1271 otologically normal Chinese adults aged 18-60 years were divided into four groups with 10-year age intervals. Based on the pure tone audiometry data and power function operation formula, the relevant parameters of the median and percentile calculation formula of ISO 7029 were adjusted. Based on the adjusted parameters, a statistical distribution table of HTL of normal Chinese adults in otology was preliminarily constructed according to the calculation formula of ISO 7029. A nonparametric rank sum test was used to compare the difference between the adjusted median deviation value and the ISO 7029 calculation. Results Except for a few frequencies (2000, 3000, and 8000 Hz), the value of parameter α adjusted for the auditory threshold deviation of the Chinese population (α_CHN-md) increased with the increase of frequency, while the value of parameter β adjusted for the auditory threshold deviation of the Chinese population (β_CHN-md) decreased with the increase of frequency. The preliminary distribution table of otologically normal Chinese adults' median hearing threshold deviation (ΔH_CHN-md) showed that the hearing threshold deviation increased with age and frequency. At 8000 Hz, the increase in ΔH_CHN-md with age was the greatest for both men and women, from 0 dB to 23 dB and 21 dB respectively. The minimal threshold elevation was noted at 500, 1000, and 2000 Hz, which increased from 0 dB to 2 dB. In otologically normal Chinese adults, the maximum increase of ΔH_CHN-md with frequency was observed at age of 50 years, which increased from 2 dB at 500 Hz to 23 dB at 8000 Hz in men and from 2 dB at 500 Hz to 21 dB at 8000 Hz in women. The comparison results with the ISO 7029 calculation showed that the trend was consistent. The median hearing threshold deviation value of the adjusted Chinese population was lower than that of the ISO 7029 calculation (ΔH_md). However, they had no significant statistical difference (P>0.05). For otologically normal normal Chinese men and women, the maximum difference of ΔH_md between the 50-year-old group and the ISO calculation at 8000 Hz was 7 dB and 9 dB, respectively. Conclusion This preliminary analysis of hearing threshold deviations in otologically normal Chinese adults suggests that current standards may potentially underestimate hearing loss at frequencies below 8000 Hz while possibly overestimating it at 8000 Hz. These findings require further validation through expanded sample sizes to more accurately assess the characteristics of hearing thresholds in the Chinese population and their discrepancies with existing standards.

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.

Aim To investigate the effect of linarin on improving cognitive behavior of APP/PS1 mice,and to explore the therapeutic effect of linarin on A β deposi-tion and neuroinflammation and its correlation.Meth-ods APP/PS1 transgenic mice were randomly divid-ed into the model group,high-dose group,medium-dose group,low-dose group and positive control group.C57BL/6J mice were set as the normal group.Morris water maze was used to evaluate the learning and mem-ory abilities of mice.TUNEL staining was used to de-tect the apoptosis of neurons in the CA1 region of mice.IHC was used to detect the expression levels of Aβ42 and GFAP.Western blot was used to detect the expression levels of BACE1 and PS-1.Results Com-pared with the normal group,mice of the model group showed lower NCP,shorter target quadrant travel,less target quadrant residence time percentage(all P＜0.01),higher apoptosis rate of neurons in the CA1 re-gion(P＜0.01),significantly higher protein expres-sion levels of A β42 and GFAP(all P＜0.01),and significantly higher protein expression levels of BACE1 and PS-1(all P＜0.01).Compared with the model group,the medium-dose group,high-dose group and positive control group showed higher NCP,longer tar-get quadrant travel,more target quadrant residence time percentage(all P＜0.05),lower apoptosis rate of neurons in the CA1 region(P＜0.01),significantly lower protein expression levels of A β42 and GFAP(all P＜0.01),and significantly lower protein expression levels of BACE1 and PS-1(all P＜0.01).Conclu-sions Linarin can inhibit two key enzymes to reduce the decomposition of APP and the generation of A β42,thereby inhibiting the activation of astrocytes,allevia-ting neuroinflammation,improving the core pathologi-cal features of AD,and thus significantly improving learning and memory impairment in APP/PS1 mice.

OBJECTIVE To explore the association of the expressions of tissue forkhead box transcription factor(FOX)F2,angiopoietin-like protein 4(ANGPTL4)and FOXP3 with the pathological features and recurrence of the cervical cancer patients with human papillomavirus(HPV)infection undergoing radical hysterectomy.METHODS Totally 69 cervical cancer patients with HPV infection who underwent radical hysterectomy in Danzhou People's Hospital from Nov.2018 to Jul.2021 were randomly assigned as the cervical cancer group,meanwhile,72 patients who had cervical intraepithelial neoplasia(CIN)were chosen as the CIN group.The patients of the cer-vical cancer group were divided into the recurrence group with 25 cases and the no recurrence group with 44 cases according to the prognosis.The levels of tissue FOXF2,ANGPTL4 and FOXP3 were compared among the groups.The levels of FOXF2,ANGPTL4 and FOXP3 were compared among the cervical cancer patients with dif-ferent clinical pathological features,and the pathological features were compared among the patients with different treatment outcomes.The values of the three indexes in diagnosis of the postoperative recurrence in the cervical cancer patients with HPV infection undergoing radical resection were analyzed.RESULTS There were significant differences in the levels of tissue FOXF2 mRNA,ANGPTL4 mRNA and FOXP3 mRNA between the cervical cancer group and the CIN group(P＜0.05).There were significant differences in the levels of FOXF2 mRNA,ANGPTL4 mR-NA and FOXP3 mRNA among the cervical cancer patients with HPV infection who varied in federation international of gynecology and obstetrics(FIGO)stage,lymphatic metastasis and infiltration depth(P＜0.05).There were significant differences in the FIGO stage,lymphatic metastasis,infiltration depth,FOXF2 mRNA level,ANGPTL4 mRNA level and FOXP3 mRNA level between the recurrence group and the no recurrence group(P＜0.05).The area under the curve(AUC)value of the joint detection of FOXF2 mRNA,ANGPTL4 mRNA and FOXP3 mRNA was higher than that of the single detection in diagnosis of the postoperative recurrence in the cervical cancer patients with HPV infection under-going radical resection(P＜0.05).CONCLUSIONS The cervical cancer patients with HPV infection undergoing radi-cal resection show the abnormal expressions of tissue FOXF2 mRNA,ANGPTL4 mRNA and FOXP3 mRNA.There is association between the three indexes and the FIGO stage,lymphatic metastasis and infiltration depth.The joint detection of the three indexes has high value in diagnosis of the postoperative recurrence.

AIM To study the chemical constituents from Commelina communis L.METHODS The 95％ethanol extract from C.Communis was isolated and purified by activated charcoal,silica gel,Sephadex LH-20,and HPLC,then the structures of obtained compounds were identified by physicochemical properties and spectral data.RESULTS Seventeen compounds were isolated and identified as p-hydroxyl ethyl cinnamate(1),p-hydroxybenzaldehyde(2),vanillin(3),4-hydroxy-2,3-dimethyl-2-nonen-4-olide(4),hemeratrol A(5),chakyunglupulin B(6),chakyunglupulin A(7),2-(2-hydroxyethyl)-3-methylfumaric acid(8),N-cis-feruloyl tyramine(9),N-trans-coumaroyltyramine(10),5,6,7,3',4',5'-hexamethoxyflavone(11),N-trans-sinapoyltyramine(12),dihydro-feruloyltyramine(13),N-trans-feruloyltyramine(14),2-phenylethanol-β-D-glucoside(15),quercetin-3-O-β-D-glucoside(16),and isorhamnetin-3-O-β-D-glucopyranoside(17).CONCLUSION Compounds 4-8,10 and 11 are isolated from Commelina genus for the first time,and 1,9,12-15 are first isolated from this plant.

Objective To explore the injury types,associated injuries,and correlations of the lateral meniscus posterior root(LMPR),and to improve the comprehensive understanding of LMPR and its associated injuries.Methods The patients with LMPR who underwent knee MRI examination were retrospectively selected.A total of 223 patients with LMPR injury were classified into 4 types.The integ-rity of the meniscofemoral ligament and the grading of the cartilage injuries in the lateral tibiofemoral compartment were recorded.The relationship between the types of LMPR injury and the lateral meniscus tear locations,as well as lateral meniscus extrusion were analyzed.The relationship between the integrity of the meniscofemoral ligament and lateral meniscus extrusion was analyzed.The correlation between the time to clinical presentation after injury and the grading of the cartilage injuries was analyzed in patients with anterior cruciate ligament(ACL)ruptures.Results The incidence of LMPR injury was 1.02％,with males affected 2.19 times more frequently than females.Among patients with ACL ruptures,the incidence was 13.17％.Both type Ⅰ and type Ⅱ LMPR inju-ries predominantly involved only the posterior root,while type Ⅲ injuries mainly affected the posterior root extending to the posterior horn and body.The incidence of the lateral meniscus extrusion was higher in type Ⅲ LMPR injury than in type Ⅱ.When the menis-cofemoral ligament was not intact,the incidence of the lateral meniscus extrusion increased.In patients with ACL ruptures,a longer time to clinical presentation after injury was associated with more severe cartilage injuries grading in the lateral tibiofemoral compart-ment.Conclusion Males are more susceptible to LMPR injuries than females.The classification of LMPR injuries is correlated with the location of the lateral meniscus tears and the incidence of the lateral meniscus extrusion.The integrity of the meniscofemoral lig-ament is related to the incidence of the lateral meniscus extrusion.The time to clinical presentation after injury is related to the sever-ity of cartilage injury in ACL rupture patients.

Objective To design an electric ice blanket system for prehospital emergency care of patients of heat sroke disease.Methods The electric ice blanket system consisted of a cooling host and cooling accessories.The cooling host had its chassis made of acrylonitrile-butadiene-styrene(ABS),which was equipped externally with an lithium battery and a DC power adapter,and integrated internally a cooling system,an internal circulation pump,an external circulation pump and a main control system;the cooling accessories included a cooling blanket,a cooling cap and a cooling vest,which had the inner layer made of thermoplastic polyurethane(TPU)elastomer and the outer layer made of Oxford cloth or polyester fiber.The system was compared with the existing subcooling therapeutic apparatus on the market in terms of cooling effect with a water bag simulation cooling experiment.Results The cooling experiment showed that the system was comparable to the existing subcooling therapeutic apparatus on the market in terms of cooling effect while behaved well in size and weight.Conclusion The system developed has a high cooling effect and advantages in portability and compatibility to the environment without power supply,which can be used for the early treatment of patients of heat stroke disease.[Chinese Medical Equipment Journal,2025,46(9):16-21]