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.

Objectives: . Idiopathic sudden sensorineural hearing loss (ISSNHL) is a significant cause of hearing loss. Intratympanic steroid injection (ITSI) is commonly used as an initial or salvage treatment; however, the lack of a standardized treatment protocol has resulted in variability in clinical practice. In addition, no efficient prediction model currently exists to support personalized management. Therefore, this study aimed to develop tailored management strategies for ISSNHL using a machine-learning model. Methods: . This retrospective multicenter cohort study was conducted between January 2015 and December 2020, with data analysis performed between January 2021 and March 2024. Patients were selected based on the International Classification of Diseases, 10th Revision criteria for ISSNHL, along with relevant medication and procedure codes. Patients with pure-tone audiogram results not meeting ISSNHL criteria, better initial hearing in the affected ear, an identifiable etiology, no post-treatment audiogram, or delayed treatment (>6 weeks) were excluded. We included 770 patients diagnosed with ISSNHL who received ITSI. The primary outcome was the area under the receiver operating characteristic curve for prediction performance. Recovery status was determined using the last pure-tone audiogram. Modeling was conducted on the Quanta for Medical Care AI platform using five machine-learning algorithms and a nested cross-validation framework, in which feature selection and hyperparameter tuning were performed in the inner folds and model performance was evaluated in the outer folds. Results: . A random forest classifier outperformed the other models in predicting hearing outcomes, achieving an area under the receiver operating characteristic curve of 0.788. Time to ITSI was the most influential treatment-related factor, with ITSI administered within 10 days of hearing loss being associated with better outcomes. This model can be used to provide personalized prognostic estimates under different treatment protocols. Conclusion . The machine-learning-based prediction model facilitates personalized treatment strategies and timely treatment adjustments for ISSNHL, thereby optimizing the likelihood of complete recovery.

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.

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 investigate the predictive value of temperature gradients on the mortality of sepsis patients and their correlation with fluid input.Methods By means of a prospective observational method,154 patients with sepsis or septic shock admitted to the Department of Critical Care Medicine at Nanfang Hospital,Southern Medical University from November 2019 to November 2021 were included as research subjects.They were divided into a survivor group(n=118)and a non-survivor group(n=36)according to whether they survived within 28 days.The core-to-toe temperature gradient(CTTG)and toe-to-room temperature gradient(TRTG)were monitored and calculated immediately upon admission to the intensive care unit(ICU)and 6 hours after admission.Receiver operating characteristic(ROC)curve was used to explore the predictive value of temperature gradients on mortality,and multivariate Cox regression analysis was performed to explore the risk factors of 28-day mortality in sepsis patients.The results were verified through survival analysis.Correlation analysis and multivariate analysis of variance were used to explore the correlation between temperature gradients and fluid input,as well as noradrenaline doses.Results Among the 154 patients,118 survived within 28 days(survivor group),and 36 died(non-survivor group).ROC curve and multivariate Cox regression analysis showed that a toe-to-room temperature gradient of≤5.35℃within 6 hours after admission was a risk factor for 28-day mortality.Compared with patients with a high toe-to-room temperature gradient(>5.35℃),patients with a low toe-to-room temperature gradient(≤5.35℃)had a 2.74-fold increase in the risk of 28-day mortality(P=0.004,95%CI 1.54,9.12).The CTTG and TRTG upon admission to the ICU and 6 hours after admission were not significantly associated with fluid input or noradrenaline doses(P>0.05).Conclusions A toe-to-room temperature gradient of less than or equal to 5.35℃within 6 hours after ICU admission is a risk factor for 28-day mortality in sepsis patients.The improvement of temperature gradients at different time points is not associated with fluid input.

Objective To investigate the status of population dynamics and distribution changes of Aedes aegypti in Guangdong Province.Methods Continuous monitoring was conducted from May 2018 to July 2024 in Wushi Town and Qishui Town,Leizhou City,Zhanjiang City,Guangdong Province.Additionally,a survey of the distribution of Ae.aegypti along the Leizhou Peninsula coast was carried out.Results The density of Ae.aegypti in Zhanjiang showed a gradual decline from 2018 to 2024.The last detection of adult Ae.aegypti in Wushi Town was in September 2021,and the last larva was found in October 2023.No Ae.aegypti was detected in Qishui Town during surveys from 2021 to 2024.A survey of 18 coastal villages in the Leizhou Peninsula revealed no detections of Ae.aegypti.Conclusions This study provides a basis for understanding the distribution and population density fluctuations of Ae.aegypti,assessing its invasion risk,and scientifically conducting relevant prevention and control efforts.

Objective:To observe the effects of acupuncture on the motor function of Parkinson disease(PD)model mice and to investigate the neuroprotective effects of acupuncture on PD from the perspective of cellular senescence.Methods:C57BL/6J mice were randomly divided into a normal control(NC)group,a 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine(MPTP)group,an acupuncture(ACU)group,and a rasagiline(RAS)group,with 6 mice in each group.Except for the mice in the NC group,all mice were injected intraperitoneally with MPTP[30 mg/(kg·bw)]to establish a PD mouse model.After the models were successfully established,mice in the ACU group received acupuncture at Baihui(GV20)and bilateral Yanglingquan(GB34)for 15 min,once a day for 14 consecutive days.Mice in the RAS group were treated with gavage of rasagiline mesylate[0.5 mg/(kg·bw)],once daily for 14 d.Mouse balance and motor functions were detected using the mouse fatigue rotating rod apparatus.Immunohistochemistry staining was used to detect the number of tyrosine hydroxylase(TH)-positive neurons and the protein expression levels of special AT-rich sequence-binding protein 1(SATB1),p21,and p53 in the substantia nigra(SN)region of the mouse brain in each group.The glutathione peroxidase(GSH-Px)activity of mouse brain SN tissue was detected by enzyme-linked immunosorbent assay.The protein expression levels of interleukin(IL)-6 and senescence-associated β-galactosidase(SA-β-gal)in the SN tissue of mice in each group were detected by Western blotting.The relative expression of SATB1,p21,and p53 mRNA in the SN of each group was detected by real-time quantitative polymerase chain reaction.Results:Compared to the NC group,the overall rod performance(ORP)score,the number of TH-positive neurons,and GSH-Px activity in the SN region were significantly lower in the mice in the MPTP group(P<0.01);compared to the MPTP group,the ORP score,the number of TH-positive neurons,and GSH-Px activity were significantly increased in the ACU group and the RAS group(P<0.01 or P<0.05).Compared to the NC group,the protein levels of IL-6 and SA-β-gal in the SN tissue,the protein and mRNA expression levels of p21 and p53 were significantly increased(P<0.01);compared to the MPTP group,the protein levels of IL-6 and SA-β-gal in the SN tissue,the protein and mRNA expression levels of p21 and p53 were significantly decreased in the ACU group and the RAS group(P<0.01 or P<0.05).Compared to the NC group,the relative expression of SATB1 protein and mRNA in the SN of mice in the MPTP group was significantly decreased(P<0.01);compared to mice in the MPTP group,mice in the ACU group and the RAS group showed significant increases in the relative expression of SATB1 protein and mRNA(P<0.01 or P<0.05).Conclusion:Acupuncture can improve motor function and increase the number of TH-positive neurons in the SN of PD model mice.Its neuroprotective effect may relate to the regulation of the SATB1/p21 signaling pathway and the inhibition of cellular senescence-related biomarker expression in the SN.

Objective To explore the value of the combined detection of long intergenic non-coding RNA(RNA LINC)00665 and carbohydrate antigen 153(CA153)in the diagnosis and prognosis assessment of breast cancer.Methods 50 breast cancer patients admitted to Jiujiang First People's Hospital were selected to be included in breast cancer group from January to December 2023,50 breast fibroadenoma patients treated in the hospital during the same period of time were selected to be included in breast fibroadenoma group,and another 50 female healthy individuals undergoing physical examination in the hospital during the same period of time were selected to be control group.The clinical data of the enrolled patients were collected,and the levels of LINC00665 and CA153 were detected by reverse transcription quantitative polymerase chain reaction and electrochemiluminescence respectively.The differences in the expression levels of LINC00665 and CA153 among three groups were compared,and the correlation between the expression levels of serum LINC00665 and CA153 was analyzed by Spearman correlation analysis;The diagnostic value of LINC00665,CA153 and the combined detection of the two in the diagnosis of breast cancer was analyzed by receiver operating characteristic(ROC)curve,and the factors affecting the prognosis of breast cancer were analyzed by multifactorial regression.Results Compared with control group,serum LINC00665 and CA153 expression levels were higher in breast cancer group,and serum LINC00665 expression levels were higher in breast fibroadenoma group.Serum LINC00665 and CA153 expression was correlated with TNM stage,pathological type,and lymph node metastasis of breast cancer(P＜0.05).Spearman correlation analysis showed that there was a linear positive correlation between serum LINC00665 and CA153 expression in patients with breast cancer(r=0.651,P＜0.05).ROC curve analysis showed that the value of combined diagnosis of LINC00665 and CA153 was higher than that of individual diagnosis(P＜0.05).The results of multifactorial regression analysis showed that LINC00665,CA153,TNM stage,and lymph node metastasis were independent risk factors affecting the prognosis of breast cancer patients(P＜0.05).Conclusion The higher expression levels of serum LINC00665 and CA153 in breast cancer patients were correlated with TNM stage,pathologic type,and lymph node metastasis,which can be used in the diagnosis and prognosis evaluation of breast cancer.