Abbreviated MRI (AMRI) protocols may offer an alternative to ultrasound for hepatocellular carcinoma (HCC) surveillance in high-risk patients, but standardized assessment and interreader agreement are essential for clinical implementation. As surveillance programs expand, these examinations will increasingly be interpreted by early-to-mid career radiologists, making it critical to evaluate agreement within this workforce segment. To evaluate interreader agreement for image quality assessment, observation detection, and proposed AMRI-LI-RADS classification in non-contrast (NC) AMRI for HCC surveillance among early-to-mid career radiologists, and to assess the influence of experience within this cohort on interpretation patterns. In this retrospective analysis of 614 NC-AMRI examinations (HASTE and DWI sequences), four early-to-mid career radiologists (2, 4.5, 5.5, and 5.5 years of post-training experience in abdominal MRI) independently evaluated image quality (5-point scale), presence of observations (≥ 10mm), and proposed AMRI-LI-RADS classification (benign, equivocal, or malignant). Agreement was assessed using percentage agreement and prevalence-adjusted bias-adjusted kappa (PABAK) for image quality, and Cohen's kappa for observation detection and proposed AMRI-LI-RADS classification. Friedman test, Cochran's Q test, and Bhapkar test for marginal homogeneity were used to evaluate differences across experience levels, with subsequent pairwise comparisons and regression analyses to quantify experience-related effects. Most examinations received excellent image quality ratings across all readers (HASTE: 73.3%-97.1%; DWI: 66.3%-93.7%). Interreader agreement was substantial for dichotomized image quality (PABAK: 0.807-0.995), observation detection (kappa: 0.742-0.993), and proposed AMRI-LI-RADS classification (kappa: 0.641-0.915). Reader experience significantly influenced assessments , with less experienced readers assigning higher quality scores and detecting more observations. There was no significant influence of reader experience on AMRI-LIRADS classification. NC-AMRI demonstrates high interreader agreement for image quality assessment, observation detection, and proposed AMRI-LI-RADS classification among early-to-mid career radiologists, supporting reproducibility of NC-AMRI interpretation. However, significant reader-dependent differences in quality ratings and observation detection indicate systematic variations in reader thresholds, highlighting the importance of standardized interpretation criteria and reader calibration when implementing NC-AMRI surveillance programs.
Rapid, accurate detection of early ischemic changes (EIC) on non-contrast computed tomography (NCCT) is critical for the triage and treatment of acute ischemic stroke (AIS) patients, yet NCCT interpretation remains challenging due to low soft-tissue contrast and inter-reader variability. This study validates StrokeSENS ASPECTS, a fully automated AI-based software tool designed to aid clinicians in grading region-level EIC. A fully crossed multi-reader multi-case (MRMC) clinical reader study was conducted with eight clinicians who independently scored 100 NCCT scans from patients with confirmed middle cerebral artery (MCA)/internal carotid artery (ICA) occlusion, unaided and aided by StrokeSENS ASPECTS, with a three-expert-neuroradiologist consensus serving as the reference standard. Reader performance was assessed using binary classification metrics, including balanced accuracy, overall accuracy, sensitivity, and specificity; inter-reader agreement was assessed using Fleiss's kappa. The use of StrokeSENS ASPECTS improved readers' balanced accuracy by 5.7 percentage points, whereas overall accuracy, sensitivity, and specificity improved by 2.6, 9.7, and 1.6 percentage points, respectively, when compared to the unaided baseline (p < 0.001). Inter-reader agreement showed a significant increase in Fleiss's Kappa of 0.285 from 0.323% (unaided) to 0.608% (aided). StrokeSENS ASPECTS have shown to improve clinicians' ability to detect EIC on NCCT and reduce inter-reader variability. This demonstrates StrokeSENS ASPECTS's safety and effectiveness as an aid in the evaluation of AIS.
To compare biparametric MRI (bpMRI) and multiparametric MRI (mpMRI) for detecting clinically significant prostate cancer (csPCa), and to assess the impact of artificial intelligence (AI)-assisted bpMRI on diagnostic performance and biopsy-related outcomes in readers with different expertise. In this retrospective multi-reader study, 173 men referred for prostate mpMRI were evaluated by five radiologists (two experts, three basic readers) who scored bpMRI and mpMRI using PI-RADS v2.1. After a 45-day wash-out, the same readers reinterpreted bpMRI with concurrent AI decision support (available for 127 cases). Diagnostic performance for csPCa (ISUP ≥ 2) and benefit-to-harm ratios were compared across protocols and reader groups. Mean area under the receiver operating characteristic curve was 0.820 for bpMRI and 0.819 for mpMRI (difference 0.001), indicating comparable diagnostic performance. In the AI subset, AI increased mean specificity from 59.7% to 67.8% while reducing sensitivity from 90.2% to 84.5%. Biopsy selectivity and efficiency improved (3.9 to 4.6 and 1.7-1.9), particularly in basic readers. The effect was significant in basic readers (McNemar p = 0.012 for sensitivity, p < 0.001 for specificity) but not in experts. bpMRI showed comparable performance to mpMRI for csPCa detection in routine practice. AI-assisted bpMRI improved biopsy efficiency in less-experienced readers at the cost of reduced sensitivity, warranting careful consideration of the trade-off between missed cancers and avoided unproductive biopsies.
While the current Vesical Imaging Reporting and Data System (VI-RADS) provides a valuable tool for evaluating bladder cancer, it is not tailored to the ureteral orifices and primarily relies on subjective evaluation. To identify significant parameters from MRI for diagnosing muscle-invasive bladder cancer (MIBC) at the ureteral orifice, and to evaluate their incremental value to VI-RADS for readers of varying experience. Retrospective. Development cohort: 81 patients with ureteral orifice bladder tumors (mean age, 68 ± 9; 60 men). External validation cohort: 34 patients (mean age, 63 ± 5; 22 men). 3.0 T; fast spin-echo T2-weighted imaging, single-shot echo planar diffusion-weighted imaging, 3D spoiled gradient echo T1-weighted dynamic contrast-enhanced imaging. Four radiologists of varying experience independently assigned VI-RADS scores. Quantitative parameters (tumor length, tumor contact length [TCL], stalk width [SW], and Ktrans) and qualitative parameters were assessed. Pathology was the reference standard. Intraclass correlation coefficient, weighted κ analysis, independent samples t-test, Mann-Whitney U test, chi-square test, receiver operating characteristic curve analysis, univariate and multivariate logistic regression, 1000 bootstrap resamples, DeLong's test, and McNemar's test with Bonferroni correction. Significance was defined as p < 0.05, with p < 0.008 for multiple comparisons. Multivariate analysis identified TCL (odds ratio [OR] = 1.12) and stalk width (OR = 1.74) as independent predictors. In both cohorts, the TCL + SW-modified VI-RADS significantly improved sensitivity over the original VI-RADS for diagnosing MIBC at the ureteral orifice in two junior radiologists. The TCL + SW-modified VI-RADS did not significantly improve sensitivity compared with the TCL-modified VI-RADS in either cohort (p = 0.25-0.50). The integration of TCL and stalk width into the VI-RADS improves diagnostic performance for MIBC at the ureteral orifice, particularly for less-experienced radiologists. Nevertheless, external validation in larger, multicenter cohorts is required. 3. Stage 2. This study aimed to identify significant parameters from MRI for diagnosing muscle‐invasive bladder cancer (MIBC) at the ureteral orifice, and to evaluate their incremental value to the original VI‐RADS among readers with differing experience levels. Tumor contact length (TCL) and stalk width (SW) were found to be independent predictors of muscle invasion. Integrating TCL and SW into VI‐RADS improved diagnostic sensitivity for MIBC at the ureteral orifice, particularly for less‐experienced readers. The TCL + SW‐modified VI‐RADS offers a practical tool to assess bladder cancers at the ureteral orifice, particularly aiding less‐experienced radiologists in achieving more reliable preoperative assessment of this challenging subsite.
The purpose of this study is to evaluate whether the reasoning model DeepSeek-R1 can function as a second-reader quality control (QC) tool for Chinese-language ultrasound reports. In this retrospective diagnostic-accuracy study with a parallel blinded review design, 500 deidentified finalized ultrasound reports were randomly sampled from 9711 eligible reports finalized in 2024 at a tertiary cancer center. DeepSeek-R1 and physician reviewer groups independently evaluated the same reports, and none of the review conditions had access to the outputs of the others or to the consensus reference standard. DeepSeek-R1 achieved 69.1% sensitivity and 98.1% specificity. DeepSeek-R1 showed numerically higher sensitivity than senior physicians (69.1% vs 47.1%), though this difference did not reach significance after Bonferroni correction (adjusted p = 0.147); specificity was identical at 98.1% for both. The model performed best for findings-impression discordance (36/42) and more modestly for completeness/template/indicator violations (7/21). In a post hoc exploratory OR-rule simulation, the combined workflow yielded 95.6% sensitivity (95% CI 87.8-98.5) and 96.3% specificity (95% CI 94.1-97.7). This retrospective single-center study provides workflow-level feasibility evidence that a reasoning model can serve as a high-specificity second-reader control for finalized ultrasound report text, with human review retained for local rules, exceptions, and final sign-off.
Preeclampsia (PE) is a placenta-driven hypertensive disorder characterized by oxidative stress, mitochondrial dysfunction, and abnormal autophagy. However, how these stress-response programs are coordinated at the post-transcriptional level remains incompletely understood. We integrated transcriptomic, proteomic, and untargeted metabolomic profiling of placentas from PE patients and normotensive controls, followed by validation in an independent clinical cohort. Mechanistic roles were interrogated using hypoxia-treated HTR-8/SVneo trophoblasts and an L-NAME-induced PE-like mouse model. Autophagy dynamics were assessed using dual-fluorescent autophagic flux reporters, transmission electron microscopy, and molecular markers. Multi-omics analyses converged on oxidative stress-related pathways, oxidative phosphorylation, and autophagy. The m6A reader HNRNPC was selectively upregulated at the protein level in PE placentas and positively correlated with the autophagy-related SNARE protein SEC22B. Clinically, PE placentas exhibited elevated global m6A methylation, increased HNRNPC and SEC22B expression, enhanced autophagic activity, and excessive oxidative stress. In trophoblasts, hypoxia induced ROS and autophagy, while HNRNPC overexpression further amplified intracellular and mitochondrial ROS, enhanced autophagic flux, and promoted apoptosis. In vivo, placental overexpression of HNRNPC aggravated hypertension, proteinuria, placental structural damage, oxidative stress, and autophagy dysregulation in PE-like mice. Our findings implicate a potential HNRNPC-SEC22B-autophagy regulatory network that amplifies placental stress maladaptation in preeclampsia. This study links epitranscriptomic remodeling to redox and autophagy dysregulation, suggesting HNRNPC as a potential mechanistic hub and therapeutic target in PE.
Previous research has identified cerebellar involvement in reading, but it remains unclear whether specific cerebellar regions are selectively engaged in lexical reading. To address this issue, the present study focused on the activation of the right cerebellar lobules VI, VII, and VIII across reading and non-reading tasks. By using functional magnetic resonance imaging (fMRI) with Chinese adult readers, we conducted two experiments. Experiment 1 identified reading-selective areas by comparing activation during implicit lexical reading against pseudo-character/string and picture-viewing tasks. The results pinpointed the right cerebellar lobules VI and VIII as critically involved in reading, which showed significantly greater activation for real words than for pseudo-characters/strings and picture-viewing tasks. Lobule VII did not show significant task effect. Experiment 2 delineated the functional roles of these areas during orthographic, phonological, and semantic processing. Lobule VI exhibited high inter-participant consistency during phonological tasks. These findings highlight the functional heterogeneity of the cerebellum, demonstrating that one specific subregion (i.e. right lobules VI) is selectively engaged in lexical reading and phonological decoding.
The integrity of human plasma and serum (P&S) is crucial for reliable biomarker discovery and validation, but it is affected by preanalytical exposure to thawed conditions (temperatures > -30 °C). The limited number of existing quality control methods for assessing plasma or serum (P/S) integrity often require resource-intensive techniques like mass spectrometry. We have developed and validated an accessible, cost-effective absorbance-based plate reader assay based on the oxidative ex vivo consumption of most small molecule thiols & disulfides (SMT&D) by albumin. The assay separates SMT&D from proteins, then uses tris(2-carboxyethyl)phosphine to reduce SMT&D, with subsequent quenching by 4-azidobenzoic acid, followed by reaction of thiols with Ellman's reagent. The method was optimized for yield, including steps to remove trace metals and minimize thiol re-oxidation. Due to inter-individual matrix effects, it was determined that measurements of each specimen before and after intentional expiration (which produce a single Δ-absorbance value, calibrated to Δ-SMT&D) are required to forensically evaluate prior exposure to thawed conditions. The assay demonstrated robust linearity, precision, and accuracy across diverse inter-individual matrices. Stability assessments at 23 °C, 4 °C and -20 °C revealed consistent exponential decay patterns in SMT&D, similar to those observed in the established LC/MS-based ΔS-Cys-Albumin assay. Estimated population reference ranges for P&S demonstrated sufficient dynamic range for the analysis of single samples with unknown histories. This potential was verified with blinded challenges. The assay provides a practical, readily accessible tool for quantifying P&S exposure to thawed conditions, facilitating improved quality control for investigators working with P&S for clinical or research purposes.
Nitrogen is an essential nutrient vital for plant health and productivity. How plants integrate nutrient signals and epigenome dynamics to modulate transcription and developmental transition remains largely unknown. Here, we uncover the crucial role of EARLY BOLTING IN SHORT DAYS (EBS) homeostasis in controlling floral transitions in response to nitrogen deficiency. EBS, a bivalent histone reader capable of recognizing both H3K27me3 and H3K4me3 histone marks, can switch its binding preference to regulate the vegetative-to-reproductive transition. We demonstrate that nitrogen and Target of Rapamycin (TOR) signaling regulate EBS protein abundance through a direct TOR-EBS interaction. TOR phosphorylates EBS at the S195 and S196 residues, which promotes EBS stability and represses the transcription of FT and other flowering genes, thereby preventing premature floral transition. Collectively, this study identifies EBS as a direct substrate of TOR and reveals a mechanistic link between nutrient signaling, epigenome dynamics, and plant developmental transition. Our findings provide important insights into complex nutrient-TOR-chromatin interplays and highlight the intricate mechanisms by which plants adapt their growth and developmental processes based on nutrient availability.
Democratic challenges are often attributed to the spread of misleading, untrustworthy, or biased information, leading scholars to focus on minimizing exposure to such "bad" content online. Instead, we introduce a scalable intervention to put factual and verified public affairs information in users' social media feeds to make them better informed and more resilient to various online threats. We conducted 48 field quasi-experiments using Instagram ads targeting news non-users to enhance their belief accuracy, democratic attitudes, and behavioral intentions related to climate change, COVID-19 vaccines, media literacy, and election integrity. The treatment videos reached 2,496,878 Instagram accounts, 690,470 users watched at least 50% of the video, and 40,584 of those users completed post-test assessment. The intervention was effective: 46 out of 48 of the quasi-experiments had positive effect sizes and 40 out of 48 achieved statistical significance. The intervention predicted not only belief accuracy but also attitudes, media literacy, and - to some extent - behavioral intentions related to vaccination. These patterns emerged across topics, did not dissipate with time (two of three climate change quasi-experiments show continued effects), and were not contingent on persuasive appeals and format features presented in the ads.
暂无摘要（点击查看详情）
Academic publishing in surgery has undergone profound change during the past several decades. Expansion of medical schools, residency programs, international academic centers, and digital publishing platforms has produced unprecedented growth in manuscript submissions and intensified competition for professional attention. Journals are judged both by readership, as measured by article downloads, and by scientific influence, as reflected in scholarly citation. At The American Surgeon, these changes prompted development of editorial frameworks designed to identify contributions most likely to matter to practicing surgeons and subsequent investigators. Many manuscripts contained observations whose significance was underrecognized by their authors. This observation led to the Hidden Publishable Idea (HPI), a framework for identifying contributions most useful to readers. Once identified, the HPI often revealed methodological limitations that imposed an evidentiary ceiling, preventing definitive conclusions while suggesting new hypotheses for future investigation. Analysis of downloads and citations suggested that readership and scholarly adoption are related but distinct outcomes. This observation led to development of the CitDL matrix, a two-by-two framework based on high and low download and citation performance. The editorial objective was not simply manuscript acceptance, but identification and development of contributions that could move manuscripts toward greater readership, greater scholarly engagement, or both. These concepts represent adaptive responses to the contemporary challenge of helping useful ideas find their audience and contribute to the advancement of surgical practice and science.
Accurate assessment of remaining growth is crucial for risk stratification and treatment planning in adolescent idiopathic scoliosis (AIS). Risser staging is widely used but demonstrates only moderate reproducibility in routine practice, particularly on standing full-spine radiographs where the iliac apophysis occupies a small area. This study aimed to develop and evaluate an interpretable, decision-oriented deep-learning model for automatic stratification of skeletal maturity as Risser 0-2 vs. 3-5 using routine full-spine radiographs. This single-center retrospective study included 875 standing posteroanterior full-spine radiographs from AIS patients aged 10-18 years. An expert consensus reference standard was established using a three-surgeon adjudication process. A predefined, rule-based pelvic region of interest was automatically extracted for model training. The data were split at the patient level into training, validation, and test sets (70%/10%/20%, stratified). Multiple convolutional neural network architectures were evaluated, and ResNet-18 was selected based on performance stability. Model interpretability was assessed using gradient-weighted class activation mapping (Grad-CAM). An independent clinical reader study was conducted on 50 sampled test cases, in which three spine surgeons completed two reading sessions (without and with model assistance) to assess agreement with the expert reference and reading time. For binary stratification (Risser 0-2 vs. 3-5), the selected model achieved an area under the receiver operating characteristic curve (AUC) of 0.938, an accuracy of 0.875, and a Cohen's kappa of 0.748 on the independent test set. Grad-CAM visualizations indicated that model predictions focused on the iliac apophysis and adjacent ossification regions. In the reader study, compared with unaided reading, model assistance reduced the mean reading time by 9.7-11.8 s per case across the three readers, from 34.7-43.2 s to 25.0-31.8 s (all P < 0.001) and improved agreement with the expert reference, with the most pronounced gains observed among junior surgeons. Exploratory six-class Risser staging showed substantially lower performance, particularly for intermediate stages. Using routine standing full-spine radiographs, an interpretable deep-learning model enabled clinically actionable Risser stratification and improved reading efficiency in an independent reader study, without the need for additional imaging. This approach may support the standardized assessment of growth potential in routine AIS management.
To evaluate the diagnostic performance of a regulatory-approved (CE-marked) artificial intelligence system (RetCAD) applied to nonmydriatic color fundus photographs for diabetic retinopathy (DR) screening in routine clinical care. This was a prospective single-center observational diagnostic accuracy study including adults with diabetes who underwent nonmydriatic fundus imaging between January 9, 2023 and August 6, 2024. Nonmydriatic true-color confocal fundus photographs were obtained using the iCare DRSplus camera. RetCAD generated a 5-category DR grade and a continuous severity score according to the International Clinical Diabetic Retinopathy scale. Two board-certified ophthalmologists independently graded images; one served as the reference standard, the second reader's grades were used to assess inter-reader agreement using Cohen's kappa. Diagnostic accuracy was evaluated at 3 prespecified thresholds: any DR; moderate DR or worse (referable DR, defined as International Clinical Diabetic Retinopathy grade 2 or above); and severe DR or worse. Receiver operating characteristic curves and area under the curve were derived from the artificial intelligence severity score. Subgroup analyses included age, diabetes duration, estimated glomerular filtration rate, and glycated hemoglobin. 609 participants (1218 eyes) were screened; 533 participants (1040 eyes) were included (median age 56 years [interquartile range 42-67], 51% female). For detection of referable DR at the eye level, sensitivity was 0.85 (95% confidence interval [CI] 0.79-0.91) and specificity was 0.97 (95% CI 0.96-0.98). The referral rate was 17.0%, with a reference prevalence of 17.2%. The areas under the curves were 0.85 for any DR, 0.96 for referable DR (moderate DR or worse), and 0.98 for severe DR or worse. At the patient level, sensitivity and specificity for referable DR were 0.89 and 0.95, with a referral rate of 22.0%. Sensitivity was significantly lower in participants aged ≥65 years and in those with estimated glomerular filtration rate <60 mL/min/1.73 m². Inter-reader agreement was high (κ = 0.877 unweighted; κ = 0.954 squared-weighted). RetCAD demonstrated high accuracy and strong discriminative ability for identifying referable DR using nonmydriatic fundus imaging, supporting its use as a triage tool for ophthalmic referral in routine clinical practice.
Psychologists' evaluation reports communicate diagnoses, findings, and recommendations that guide educational and health decisions for children and adolescents. Research has long shown that many of these reports are technical, test-focused, and hard for families, educators, and clinicians to use. This systematic review and meta-analysis pooled experiments in which readers compared accessible and traditional versions of the same, or closely matched, evaluation information. It included 14 experiments (N = 1,283 readers) across educational and clinical settings. Accessible reports used plain language, clearer organization, concrete examples, and reduced technical terminology, whereas traditional reports used technical language and conventional test-by-test organization. Accessible reports improved reader outcomes, with a moderate-to-large benefit (Hedges' g = 0.83, 95% confidence interval [0.60, 1.06]). Every experiment favored the accessible version. Accessible report writing is an evidence-supported communication practice for psychological evaluation reports.
Nonlinear narrative picture books represent an emerging literacy form in children's literature, yet their cognitive processing mechanisms remain understudied. This study investigated visual guidance patterns and comprehension mechanisms in 6-7-year-old children reading nonlinear narrative picture books using eye-tracking technology. Sixty-four children (72-95 months) read two Chinese nonlinear narrative picture books while their eye movements were recorded using Tobii Pro Spectrum (1,200 Hz). Four eye-tracking metrics were analyzed: fixation duration, path consistency index, cross-AOI scanning frequency, and image-first reading proportion. Comprehension was assessed through structured interviews evaluating factual understanding, inference, global coherence, and narrative structure understanding. Children exhibited a predominant "image-first" pattern, with 58.6% of total fixation time on images versus 26.3% on text. However, mean fixation duration was longer for text (348.6 ms) than images (256.2 ms), indicating deeper text processing. Path consistency (r = 0.48, p < 0.001) and cross-AOI scanning (r = 0.45, p < 0.001) positively correlated with age. Eye movement measures explained an additional 26.3% variance in comprehension beyond baseline ability, with cross-AOI scanning as the strongest predictor (β = 0.45, p < 0.001). Cluster analysis identified three reader types: integrative readers achieved highest comprehension (M = 21.8), followed by text-dominant (M = 18.7) and image-dominant readers (M = 16.2). Findings extend text-image integration models to nonlinear contexts, supporting age-adaptive design principles and differentiated reading instruction strategies for children's picture book reading.
Traditionally, mucus plugs have been evaluated using two-dimensional computed tomography (CT) images through expert visual assessment, a process that requires substantial experience and specialized knowledge of airway anatomy. In the present study, we propose a novel approach employing three-dimensional CT (3D-CT) with semi-automated quantification, which can be performed in routine clinical practice without requiring expert readers. We aimed to demonstrate its feasibility and clinical relevance for assessing mucus plug burden and exacerbation severity in asthma. This exploratory, single-center, retrospective study included 23 patients with asthma and a history of exacerbations. Using a clinically available 3D-CT application, we semi-automatically quantified bronchial endpoint count and bronchial volume in both stable and exacerbation phases, and calculated their percentage change. We further assessed correlations of these parameters with mucus plug scores and arterial blood gas measurements. The bronchial endpoint count and bronchial volume were significantly lower during exacerbations than during the stable phase; both showed a significant negative correlation with mucus plug scores during exacerbations. The percentage change of bronchial endpoint count and bronchial volume from the stable to exacerbation phase also correlated with mucus plug scores. The bronchial endpoint count and bronchial volume at exacerbation-related admission did not correlate with the PaO2/FiO2 (P/F) ratio; however, their percentage change from the stable phase to exacerbation-related admission significantly correlated with the P/F ratio. Clinically available semi-automated 3D-CT analysis enables objective and reproducible quantification of bronchial endpoint count and bronchial volume during asthma exacerbations, eliminating the need for expert readers. Their percentage change from the stable to exacerbation phase not only reflect mucus plug burden but also correlate with oxygenation status. These imaging biomarkers enable objective and quantitative assessment of structural airway abnormalities, including airway narrowing, distal airway obstruction, and mucus impaction. They are not intended to replace conventional clinical or physiological assessment, but rather to provide complementary structural information regarding the underlying causes of physiological abnormalities, such as impaired oxygenation, when detected by physiological evaluation. Furthermore, such quantitative structural assessment may contribute to a better understanding of airway structural changes and airway dynamics during asthma exacerbations.
Objective: Analyze the operation of the WeChat official account of the Chinese Journal of Epidemiology, so as to better use the WeChat official account platform to disseminate academic research results, improve the journal's influence and service ability for readers and authors. Methods: By searching all the tweets on the WeChat official account of the "Chinese Journal of Epidemiology" operated by the Chinese Journal of Epidemiology, we analyzed the tweets sent on the WeChat official account from May 2021 to March 31, 2026. Using metrics such as overall reach, average post reach, headline reach, peak reach, and WeChat communication index (WCI), we evaluated the performance of the official account. Linear regression was employed to analyze trends in key indicators, while VOSviewer 1.6.20 software was used for keyword co-occurrence analysis. Results: As of March 31, 2026, the WeChat official account of the "Chinese Journal of Epidemiology" had published 237 articles, with an average read count of 3 302.86 per article and a WCI of 154.66. Linear regression analysis revealed that the average read count, average "like" count, average "share" count, and average "star" count all showed an upward trend (all P<0.05). The "Standard-Protocol-Guideline" and "Expert Forum/Editorial" sections ranked highest in average read count (8 235.00 and 3 655.11 reads respectively), while articles on emerging infectious diseases achieved an average read count of 4 369.20. Articles from other sections also garnered significant attention through diverse presentation formats. Conclusions: The Chinese Journal of Epidemiology's WeChat official account tweets have high communication power and influence, and have become a mature and influential academic exchange platform for epidemiology. In addition to keeping an eye on the academic achievements of this journal, we should strengthen the operation and maintenance of WeChat official account, enrich the content and form of tweets, strengthen the operation and maintenance of WeChat official account, and further improve the ability to serve the readers and authors. 目的： 分析《中华流行病学杂志》微信公众号的运营情况，以更好地利用微信公众号平台传播学术研究成果、提升期刊影响力及为广大读者和作者服务的能力。 方法： 通过检索《中华流行病学杂志》微信公众号所有推文信息，分析微信公众号自2021年5月开通至2026年3月31日的推文情况。将整体传播力、篇均传播力、头条传播力、峰值传播力及微信传播指数（WCI）作为评估指标，分析该微信公众号的运营情况。采用线性回归分析主要指标的变化趋势，采用VOSviewer 1.6.20软件进行热词分析。 结果： 截至2026年3月31日，“中华流行病学杂志”微信公众号推文量为237篇，篇均阅读数为3 302.86次，WCI为154.66。线性回归分析结果显示，篇均阅读数、篇均在看数、篇均点赞数、篇均转发数均呈上升趋势（均P<0.05）。“标准·方案·指南”和“专家论坛/述评”栏目推文的篇均阅读数居前列，分别为8 235.00次和3 655.11次。新发传染病感染推文的篇均阅读数为4 369.20次，其他栏目推文通过多样化的展示方式也获得充分关注。 结论： 《中华流行病学杂志》微信公众号推文的传播力和影响力均较高，已成为较成熟且具有专业影响力的流行病学学术交流平台。应在保持关注学术成果推送展示外，加强微信公众号的运营与维护，丰富推文内容和形式，进一步提升为广大读者和作者服务的能力。.
To develop and validate APEX-NET for early diagnosis and severity stratification of acute pancreatitis (AP) using non-contrast CT (NCCT), by leveraging contrast-enhanced CT (CECT) feature learning. This five-center retrospective and prospective study included 3383 patients, comprising AP and Non-AP (abdominal pain patients and healthy individuals) patients. APEX-NET was trained and evaluated to perform pancreas segmentation, AP diagnosis (AP vs Non-AP), and severity prediction (mild, moderately severe, or severe per the revised Atlanta classification) using 3 internal and 2 external cohorts. A feature mapping module was employed to derive simulated CECT features from NCCT based on paired NCCT-CECT feature learning. The model was further evaluated with subgroup analyses, and a reader study was conducted by comparing its performance with six radiologists of varying experiences. Evaluation metrics included the Dice similarity coefficient, area under the receiver operating characteristic curve (AUC), and accuracy. For AP diagnosis, APEX-NET achieved AUCs of 0.949, 0.958, 0.981, and 0.955 in the validation, internal, and two external testing cohorts, respectively. For severity prediction, APEX-NET significantly outperformed the NCCT model (p < 0.05), with macro-average AUCs of 0.873 (validation) and 0.872 (internal testing). The advantage of APEX-NET had been demonstrated in almost all the age, gender, and etiology subgroups. In the reader study, APEX-NET performed comparably to senior radiologists and superior to junior radiologists (p < 0.05). APEX-NET enables accurate NCCT-based diagnosis and early severity stratification of AP, demonstrating strong potential for clinical integration to overcome the inherent delay of CECT-based assessment. Question The absence of an accurate method for predicting AP severity from early NCCT, the initial diagnostic scan, thus forgoing the critical intervention window. Findings Achieve accurate severity prediction for AP by incorporating contrast-enhanced feature learning. Demonstrate robust performance across diverse demographic groups, etiologies, and imaging parameters. Clinical relevance The APEX-NET, an integrated deep learning framework using NCCT, accelerated the diagnosis and severity stratification of AP, demonstrating performance comparable to senior radiologists and direct potential for clinical workflow integration by reducing reliance on delayed contrast-enhanced scans.
The EURO-PROBE dataset provides manually created segmentations of the prostate and intraprostatic urethra on axial T2-weighted MRI volumes, complementing the existing LUND-PROBE (LUND Prostate Radiotherapy Open Benchmarking and Evaluation) dataset for prostate radiotherapy research. Delineations were performed by delineators with varying levels of experience, including radiologists, radiology and urology residents, and medical students, which enables both inter-reader variability analysis and stratification by reader experience. Across all 467 cases in the LUND-PROBE dataset, at least one paired segmentation of the prostate and intraprostatic urethra was created for each case, yielding 1,227 paired segmentations in total. By expanding the LUND-PROBE dataset, EURO-PROBE supports the development of automated segmentation tools, including for the urethra, which is a highly relevant organ at risk in focal dose escalated prostate radiotherapy.