Predictive Power Research Articles (Page 1)

Endometriosis is a long-term health problem that affects a significant number of women globally. Among the various forms of endometriosis, ovarian endometriosis (OEM) is the most prevalent. This research aimed to investigate the factors contributing to the recurrence of OEM after laparoscopic conservative surgery and develop a predictive model utilizing machine learning techniques. The clinical data of 338 patients diagnosed with OEM who underwent laparoscopic conservative surgery at Wuhan University Renmin Hospital between January 2020 and January 2023 were retrospectively analyzed. During a 2-year follow-up period, patients were categorized into either the recurrence group or the non-recurrence group based on the incidence of disease recurrence. Chi-square and Spearman analysis were implemented to identify the factors related to postoperative recurrence in patients with OEM. Statistically significant factors were selected to construct the correlation models. Four algorithms were used in model construction: Random Forest, Gaussian Process, Extreme Gradient Boosting, and Multilayer Perceptron. The primary metric for evaluating model performance was the area under the receiver operating characteristic curve. Sixteen variables were associated with postoperative recurrences. The Gaussian Process had the best predictive power and the area under the receiver operating characteristic curve of the test set was 0.90. The test dataset for the Gaussian Process revealed a sensitivity of 0.75, specificity of 0.90, positive predictive value of 0.46, negative predictive value of 0.97, and accuracy rate of 0.88. The predictive model for the Gaussian Process developed in this study effectively assessed the risk of postoperative recurrence in patients with OEM.

Estimation of stature from hand length and hand breadth in undergraduate medical students: An anthropometric study.

Ferroptosis, an iron-dependent form of regulated cell death, has been linked to the occurrence and progression of ischemic stroke (IS). This study aims to uncover the key ferroptosis-related genes in IS and their correlations with immunoinflammatory responses. Key differentially expressed ferroptosis-related genes were screened by integrating differential analysis, weighted gene co-expression network analysis (WGCNA), and protein-protein interaction analysis. Machine learning algorithms, LASSO regression, Random Forest, RGF, and LightGBM were employed to identify potential diagnostic biomarkers, and diagnostic model was then established. Oxygen-glucose deprivation/reoxygenation (OGD/R)-stimulated HT-22 cells were established to validate the expression of biomarkers by RT-qPCR and western blot. Fifteen key ferroptosis-related genes were identified by integrated analyses, and ATM, DUSP1, SRC, and STAT3 were further screened as biomarkers by four algorithms. The diagnostic model established based on these four biomarkers exhibited well predictive power for IS, with AUC over 0.8 in both training and validation sets. Expression of DUSP and STATS positively correlated with neuroinflammation pathway, and positively correlated with abundance of neutrophils and macrophages. SRC positively correlated with abundance of monocytes, whereas ATM positively correlated with CD8 T cells and resting memory CD4 T cells. Both mRNA and protein levels of DUSP1, SRC, and STATS3 were significantly enhanced, while the level of ATM was reduced in OGD/R-stimulated HT-22 cells than control cells. In conclusion, dysregulation of key ferroptosis-related genes, ATM, DUSP1, SRC, and STAT3 might be implicated in the progression of IS, which could be biomarkers or targets for the diagnosis and therapy of IS.

Accurately measuring the quality of stroke care based on claims data alone is challenging. Traditional outcome metrics, e.g. mortality rates, do not capture the effectiveness of critical stroke care processes. We aimed to develop hybrid quality indicators (QIs) by integrating clinical stroke severity data with claims data. Claims data were linked to patient-level clinical data from 15 hospitals (2017-2020) and harmonized in the Observational Medical Outcome Partnership (OMOP) data model. Inclusion criteria, outcomes and risk factors were developed by medical expert panels. We applied machine learning for modeling the outcomes 30-day-mortality, reinfarction within 90 days, and care degree increase within 180 days. We compared extreme gradient boosting (XGBoost) models with and without the National Institutes of Health Stroke Scale (NIHSS) using Receiver-Operating-Characteristic-Area-Under-the-Curve (ROC-AUC) and Brier Score (BS). Hospitals were ranked according to the impact of each QI using Standardized Mortality Ratios (SMRs). The study included 9,348 ischemic (I63) and 1,554 hemorrhagic (I61) strokes, with NIHSS available for 5,012 patients. For all three outcomes, disease severity as measured by NIHSS was the most important determinant. The predictive power of the hybrid models was higher than that of models based on claims data alone. For SMR, the influence of NIHSS was greater than that of age, the most important variable in the claims data model. The results were consistent between the two entities, different outcomes, and sensitivity analyses. Including NIHSS information alongside claims data improves the risk adjustment of quality indicators.

Effects of ancestry, agriculture, and lactase persistence on the stature of prehistoric Europeans.

Introduction Diseases of plants remain one of the greatest threats to sustainable agriculture, with a direct adverse effect on crop productivity and threatening food security worldwide. Conventional detection methods rely heavily on manual detection and laboratory analysis, which are time-consuming, subjective, and unsuitable for large-scale monitoring. The use of the most recent progress in computer vision and artificial intelligence has opened up a prospect of automated, scalable, and precise disease diagnosis. Methods This paper introduces a feature-efficient hybrid model that trains classical Machie Learning (ML) classifiers with Deep Neural Network (DNN) using ResNet-based feature extraction and Principal Component Analysis (PCA). The PlantVillage dataset with mixed crop-disease pairs is used to implement and thoroughly test five hybrid models. Results Wide-ranging experiments proved that the Logistic Regression (LR)+DNN hybrid resulted in the best classification accuracy of 96.22% as compared to other models and available benchmarks. Besides being able to outperform other techniques in terms of predictive power, the framework displayed good training stability and robustness to class imbalance as well as a higher degree of interpretability based on LIME-based analysis. Discussion The obtained results confirm the hybrid ML+DNN paradigm as a safe, transparent, scalable disease recognition framework when applied to plant diseases. Providing opportunities for timely and accurate disease detection, the proposed framework can help with precision agriculture, where pesticide use can be reduced, consequently, and a significant contribution to sustainable farming can be achieved.

Combining keystroke logging, screen recordings, interviews, and text quality assessment in two mixed-methods studies with technical writers, this research (1) identifies defining variables of technical writing processes and (2) examines their correlations with and predictive power for text quality. Study 1, an exploratory investigation with 10 participants, identified 22 distinct writing behaviors under six categories of information searching, information reusing, content shaping, organization structuring, language styling, and layout designing during planning, translating, and reviewing sessions. These behavioral variables, together with time-related variables, were subsequently analyzed as “process indicators” in a comparative experiment with 43 participants across experience levels. Results of Study 2 revealed significant differences among experience levels in writing speed, planning duration, pause, search, reuse, content shaping, and structuring. Detailed planning and systematic content/structure editing were strongly associated with higher-quality texts. Building on these findings, we propose a process model of technical writing, explain its correlations with writing score, and depict process profiles of different experience levels. We also highlight the importance of information processing skills in enhancing writing efficiency, offering empirical guidance for technical writing instruction and professional training.

Purpose This study aims to investigate the key drivers of bank stability in Vietnam's emerging economy, offering a robust, data-driven framework that integrates advanced machine learning (ML), regularization techniques and explainable artificial intelligence (XAI) to address challenges in financial risk modeling and regulatory transparency. Design/methodology/approach Using bank-level data from 2010 to 2023, this study employs Ridge, Lasso and Elastic Net regression to manage multicollinearity and identify relevant predictors. Gradient boosting with ridge regularization, optimized via particle swarm optimization, achieves superior predictive accuracy (R2 = 96.03%). SHapley Additive exPlanations (SHAP) and Local Interpretable Model-agnostic Explanations (LIME) are applied to interpret the model's outputs, revealing both global and local effects of explanatory variables. Findings The model attains a strong explanatory power (R2 = 96.03%), affirming the validity of the hybrid ML-XAI approach. Core drivers of banking stability – lag_ZSCORE, foreign bank presence, return on equity, equity ratio and macroeconomic factors – are consistently identified by SHAP and LIME. The integration of interpretable artificial intelligence (AI) not only enhances predictive accuracy but also delivers actionable insights, offering meaningful guidance for financial stability in emerging markets. Research limitations/implications While the model demonstrates strong performance within Vietnam's banking sector, its applicability may vary in different regulatory or macroeconomic environments. The study focuses on supervised learning and structured data; future research could explore unsupervised methods or unstructured sources like textual financial disclosures. Additionally, while SHAP and LIME provide interpretability, they do not guarantee causal inference. Nevertheless, the research lays a solid methodological foundation for future cross-country comparisons, particularly in the Association of Southeast Asian Nations, and encourages the integration of XAI into financial stability frameworks globally. Practical implications This study equips financial regulators, central banks and policymakers with interpretable and high-performing tools to monitor and enhance banking stability. By clearly identifying and quantifying key stability drivers, it enables targeted, data-informed interventions and policy adjustments. Banks can also utilize these insights to optimize risk management, capital allocation and strategic planning. The transparent AI methods ensure trust in the decision-support system, promoting broader adoption in regulatory environments that demand both accuracy and explainability. Originality/value This is the first study in Vietnam to fuse ML, regularization and XAI for bank stability analysis. It not only enhances predictive power but also ensures model transparency, crucial for policymaking and Basel III compliance. The methodological innovation offers a replicable blueprint for financial risk assessment in emerging markets.

Little is known about how the Public Company Accounting Oversight Board (PCAOB) selects audits for inspections. We exploit public data that track PCAOB searches of issuer Securities and Exchange Commission (SEC) filings and the trial transcripts of United States of America v. David Middendorf to provide the first large-sample evidence on confidential PCAOB monitoring activities. We initially validate that PCAOB searches spike during actual inspections of triennially inspected auditors and for a list of KPMG audits inspected in 2016. Importantly, we find that PCAOB searches vary predictably with ongoing audit-flagged events, client corporate events, and client characteristics, and these screening and inspection activities extend beyond enforcement actions. Using KPMG data, we introduce a PCAOB inspection prediction model for Big 4 clients based solely on public data, with predicted incidence closely aligning with actual inspections. One implication of our study is that, perhaps to mitigate political costs, the PCAOB relies overly on conspicuous trigger events that already signal low audit quality (i.e., restatements, auditor changes, chief financial officer turnovers, bankruptcies, and SEC comment letters), raising the question of how its risk-based inspection program is designed to improve overall audit quality or minimize large audit failures. In addition, although model-inferred PCAOB inspections are not associated with future restatements, broader PCAOB monitoring activities, as proxied by search intensity, exhibit modest predictive power. Overall, our study provides a more nuanced understanding of the PCAOB inspection program and the factors driving its revealed preferences. This paper was accepted by Ranjani Krishnan, accounting. Funding: This work was supported by the Jones Graduate School of Business at Rice University, Baruch College, and the Penn State Smeal College of Business. Supplemental Material: The online appendix and data files are available at https://doi.org/10.1287/mnsc.2024.08084 .

The activity of hippocampal place cells is essential for various processes of spatial memory, including encoding, consolidation, and retrieval. However, whether they also participate in spatial memory updating remains elusive. This study demonstrates that during the process of long-term spatial associative memory updating, triggered by changes in reward magnitudes in a maze over consecutive days, there is a distinct period characterized by loss of preference for different reward magnitudes before adapting to new magnitude ratios. During this period, sequential replays in the hippocampus during post-training sleep are significantly biased toward the region associated with the larger reward. Additionally, replays between trials show elevated predictive power for the upcoming choices. Our results suggest that hippocampal replays may play a key role in updating long-term spatial associative memory.

Neurosymbolic AI is an emerging paradigm that combines neural network learning capabilities with the structured reasoning capacity of symbolic systems. Although machine learning has achieved cutting-edge outcomes in diverse fields, including healthcare, agriculture, and environmental science, it has potential limitations. Machine learning and neural models excel at identifying intricate data patterns, yet they often lack transparency, depend on large labelled datasets, and face challenges with logical reasoning and tasks that require explainability. These challenges reduce their reliability in high-stakes applications such as healthcare. To address these limitations, we propose a hybrid framework that integrates symbolic knowledge expressed in First-Order Logic into neural learning via a Logic Tensor Network (LTN). In this framework, expert-defined medical rules are embedded as logical axioms with learnable thresholds. As a result, the model gains predictive power, interpretability, and explainability through reasoning over the logical rules. We have utilized this neurosymbolic method for predicting diabetes by employing the Pima Indians Diabetes Dataset. Our experimental setup evaluates the LTN-based model against several conventional methods, including Support Vector Machines (SVM), Logistic Regression (LR), K-Nearest Neighbors (K-NN), Random Forest Classifiers (RF), Naive Bayes (NB), and a Standalone Neural Network (NN). The findings demonstrate that the neurosymbolic framework not only surpasses traditional models in predictive accuracy but also offers improved explainability and robustness. Notably, the LTN-based neurosymbolic framework achieves an excellent balance between recall and precision, along with a higher AUC-ROC score. These results underscore its potential for trustworthy medical diagnostics. This work highlights how integrating symbolic reasoning with data-driven models can bridge the gap between explainability, interpretability, and performance, offering a promising direction for AI systems in domains where both accuracy and explainability are critical.

ABSTRACT Objective: Intensive efforts are made to discharge patients successfully after mental health hospitalisation. However, there is limited knowledge about the usefulness of patients’ subjective assessments of their ability to adjust after discharge, and especially, the potential benefits of combining scales. Method: One week before discharge, 107 inpatients (aged 19–65 years), mostly hospitalized for schizophrenia (63%), were individually interviewed and completed the following scales: Sense of Coherence, Adaptability, Adjustment, Patients’ perceptions of hospitalisation, and background data. Three months later, discharged patients reported on (a) attendance at the regional outpatient clinic for follow-up and (b) re-hospitalisation in a mental health (MH) hospital. Results: The study findings indicate that the number of prior hospitalisations and the patient's subjective satisfaction with the hospitalisation significantly contributed to predicting later maladjustment. Additionally, patients’ responses to the Adjustment Scale retrospectively predicted their satisfaction with hospitalisation. Discussion: Patients’ subjective evaluation of their hospitalisation experience appears to be partly influenced by their prospective assessment of their ability to re-adjust after discharge. Therefore, it is valuable to explore the utility of a combined measure that includes ongoing data collection of patients’ hospitalisation experiences and their feelings of readiness for discharge, which could help reduce re-hospitalisation rates.

This study examines the transition from digital entrepreneurial alertness to digital startup intent in connection with perceived desirability, feasibility, and intentions. The theory of planned behavior (TPB) and the entrepreneurial event/potential model (EPM) form the foundation for a mediation model, which is examined by structural equation modeling (SEM) using AMOS on data gathered from 571 Saudi youth engaged in digital entrepreneurship. The results show that digital entrepreneurial alertness has a strong predictive power in relation to intent to start digital ventures, and that this is partly mediated by perceived desirability and feasibility. Intentions, however, fully mediate the relationship between alertness, desirability, feasibility, and actual digital entrepreneurial behavior. This study adds to digital entrepreneurship scholarship by de-mystifying the thought processes bridging opportunity recognition and action, particularly in emerging economies. This study validates the EPM framework and confirms its applicability to include digital entrepreneurial alertness (DEA) as a key antecedent of digital entrepreneurial intentions (DEI) and other factors. This study also highlights the theoretical relevance of the EPM by illustrating its utility in understanding youth decisions to pursue digital entrepreneurship, particularly in transitional countries such as Saudi Arabia. Policymakers and educators in Saudi Arabia should promote attention and amplify desirability/feasibility perceptions to stimulate youth engagement in digital ventures. This work highlights intentions as the determinative gateway between entrepreneurial cognition and concrete digital startup success.

Introduction Emotions play a pivotal role in learning, particularly in the post-COVID-19 era where online instruction has become prevalent. Despite this, emotional engagement in online mathematics remains underexplored. Methods The study introduces the Assessment Online Emotions Questionnaire (AOEQ) , developed to assess emotional dynamics among 406 10th-grade students in online mathematics. Guided by Russell’s Circumplex Model of Affect , data were analyzed using EFA and CFA to validate the instrument Results The analyses confirmed a robust three-factor structure (Useful, Unuseful, Neutral Emotions) with high reliability and significant correlations between emotional states and academic performance. Discussion Findings highlight the predictive power of positive emotions for academic success and propose AOEQ as an innovative tool for evaluating emotional engagement in virtual learning.

Abstract This study aimed to understand the relations between English as a foreign language (EFL) students’ classroom relationships, emotions, and self-regulated learning (SRL) strategy use. Data was collected via a questionnaire completed by 436 4-5th graders in Mainland China. Results of structural equation modelling (SEM) revealed that both teacher-student relationships and peer relationships contributed to students’ SRL strategy use. In addition, some of these associations were fully mediated by emotions, and some were not. Specifically, both teacher-student relationships and peer relationships were related to students’ emotions (i.e., enjoyment, hope, boredom, empathy, and gratitude) and in turn contributed to their use of SRL strategies (i.e., using prior knowledge and monitoring). However, both teacher-student relationships and peer relationships were directly related to students’ use of SRL strategies (i.e., contextual regulation and evaluation), without the mediation of emotions. Although it was found that the predictive power of teacher-student relationships was more substantial than peer relationships in emotions and SRL strategy use, both types of classroom relationships demonstrated unique and significant functions. Implications for constructing a caring and supportive classroom culture and enhancing EFL students' SRL strategy use are discussed.

Background/Objectives: Secondary hyperparathyroidism (SHPT) affects 30–50% of end-stage renal disease patients. Parathyroidectomy (PTX), while effective for medication-refractory SHPT, carries 20–70% risk of hungry bone syndrome (HBS)—severe sustained hypocalcemia requiring intensive care and prolonged hospitalization. Accurate preoperative risk stratification using biochemical markers and validated prediction tools is critical for optimal preventive management. Methods: We conducted a comprehensive narrative review synthesizing evidence on HBS predictors after PTX in SHPT, evaluating traditional and novel bone turnover markers, clinical risk factors, and multivariate prediction models, through a structured literature search and analysis. Results: Preoperative bone turnover status represents the strongest contributor to HBS risk. Traditional biomarkers—particularly parathyroid hormone (PTH > 1000–2400 pg/mL) and alkaline phosphatase (ALP > 150–300 U/L)—demonstrate moderate-to-strong individual predictive power. Novel bone turnover markers (bone-specific ALP, P1NP, TRAP-5b) offer incremental value, especially in CKD populations where renal clearance affects traditional markers. Combined risk prediction models substantially outperform single biomarkers, achieving area under curve values of 0.87–0.95. The simple NYU 2-point score (ALP > 150 U/L + PTH > 1000 pg/mL) showed 96.8% accuracy, with 100% negative predictive value. More complex tools like nomograms (C-index 0.92–0.94) and machine-learning algorithms (AUC 0.88) provide enhanced discrimination by integrating multiple continuous parameters. Additional clinical factors—younger age (<48 years), prolonged dialysis (≥5 years), low preoperative calcium, high gland weight, and absence of autotransplantation—further refine risk assessment. Postoperative calcium typically reaches nadir at 48–72 h, defining the critical monitoring window. Conclusions: High-turnover bone biomarkers and combined risk models effectively identify high-risk SHPT patients. Risk-stratified protocols (i.e., prophylactic supplementation, intensive monitoring, and selective ICU admission) can substantially reduce HBS-related morbidity. Ongoing efforts should focus on validating these predictive tools across diverse populations and integrating them into clinical practice, thereby facilitating real-time HBS risk assessment and protocol-driven care.

Background: Cancer survivors have a substantial risk of cardiovascular diseases (CVD) due to shared risk factors and cancer treatment-induced cardiac toxicity. However, the predictive power of cardiovascular risk prediction models in this population remains unclear. Research Question: This systematic review and meta-analysis aims to synthesize available data on the performance of various CVD risk scores in adult cancer survivors. Methods: PubMed, Embase, Scopus, and Web of Science databases were systematically searched from inception until May 2025 for studies that validated the performance of CVD risk prediction models in cancer survivors. Risk scores were evaluated across 4 outcome categories: vascular events (myocardial infarction, peripheral arterial disease, and stroke), cancer therapy-related cardiac dysfunction (CTRCD) or heart failure, arrhythmias, and composite CV events. For each score, discrimination metrics (area under the curve [AUC] and C-index) were pooled using random-effects models. Calibration metrics (sensitivity, specificity, and accuracy) were pooled using a bivariate random-effects model. Results: A total of 31 observational studies involving 901,664 patients (pooled mean age 61.1±14.4 years; 50.8% female) diagnosed with cancer were included. The most common cancer types were breast cancer (18.1%) and colon cancer (11.6%). In these studies, 27 unique risk scores (20 targeted to the general population and 7 to patients with cancer) were assessed. Good discriminatory power (AUC ≥0.70) was observed in 3 out of 8 risk scores predicting vascular events, 3 out of 9 for CTRCD or heart failure, 3 out of 10 for arrhythmia, and 5 out of 8 for composite cardiovascular events. Among all risk scores, the CORE (AUC: 0.74) and SCORE (5 studies; pooled AUC: 0.74, sensitivity: 52%, specificity: 85%, accuracy: 65%) risk scores exhibited the highest predictive performance for vascular events. For predicting CTRCD or heart failure outcome, the ARIC-HF and FRESCO risk scores showed the highest predictive performance, with both achieving an AUC of 0.76. Conclusion: Available cardiovascular risk scores developed in both general and cardio-oncology populations demonstrate moderate to good predictive power in cancer survivors. However, their calibration and generalizability remain limited. Future studies are needed to recalibrate and optimize scores in cardio-oncology settings to identify individuals at higher risk of developing CVD.

Background and Aims: Carotid plaque is an early manifestation of atherosclerosis and is closely associated with the risk of myocardial ischemia, ischemic stroke, and other atherosclerotic cardiovascular diseases (ASCVDs). This study aims to identify new protein biomarkers associated with carotid plaque, which will enhance early warning of ASCVD. Methods: We launched a nested case-control study based on the blood samples at baseline and repeated carotid ultrasound measurements during follow-ups in the ChinaHEART cohort. Among participants without carotid plaque at baseline, 145 with incident carotid plaque within two-year follow-up were selected as cases, and 147 without incident carotid plaque during follow-up were matched for demographic characteristics and traditional risk factors as controls. After the Meso Scale Discovery test for 28 biomarkers, Least absolute shrinkage and selection operator (LASSO) regression was used to select potential predictors and constructed a logistic regression model for predicting carotid plaque. Furthermore, the incremental predictive value was validated in the UK Biobank of 30,800 subjects. Results: A total of 11 biomarkers, including thrombomodulin, ICAM-3, P-Selectin, GDF-15, adiponectin, MCP-1, IL-10, PlGF, Tie-2, VEGF-D, and VCAM-1 were selected by LASSO regression and used to construct a prediction model for the carotid plaque. The area under the ROC curve (AUC) of the eventual model is 0.778 and it showed good calibration capability graphically with a Brier score of 0.192. In the UK Biobank cohort, when these biomarkers were added to a traditional predictive model, a better predictive power was generated, with an AUC improvement of 0.021 (P <0.001, Delong test), Brier score of 0.093, a continuous NRI of 0.259 (0.223-0.294, P <0.001), IDI of 0.017 (0.015-0.019, P <0.001) reference to the traditional model. Conclusions: We found and validated the biomarkers, including thrombomodulin, ICAM-3, P-Selectin, GDF-15, adiponectin, MCP-1, IL-10, PlGF, Tie-2, VEGF-D, and VCAM-1, can predict the incidence of carotid plaque in ChinaHEART, and except for Tie-2, these biomarkers have additional value for the prediction of incident ASCVD in UK Biobank.

Background: Atrial Fibrillation (AF) is a significant cause of cardiovascular morbidity and mortality risk. Age is a powerful predictor of adverse outcomes in patients with AF. In the United States, the burden of AF is disproportionately borne by men until the age of 75, after which it shifts to being more prevalent among women. Our objective was to understand the influence of sex on cardiovascular contributors to mortality in patients with AF stratified by age over 2 decades. Methods: We analyzed the earliest and latest available data using national death certificates from the Centers for Disease Control Wide Ranging Online Data for Epidemiologic Research (CDC WONDER) from 1999 to 2023 with AF listed as the primary or contributing cause of death and one of 4 subconditions (dementia, heart failure [HF], myocardial infarction, and stroke). Using demographic information included in death certificates, we aimed to stratify trends in age and sex as it relates to AF and these subconditions. Confidence intervals were supplied by CDC WONDER database directly or calculated according to CDC supplemental information. Results: A total of 3,272,015 patients were identified with AF as a primary or secondary contributing cause of death denoted as total mortality (461,321 deaths had AF as the primary cause of death). In women with AF, HF and dementia are contributors to mortality and are increasing in all age groups, despite lower contributions of stroke and myocardial infarction (Figure). In women >75 years, the incidence of stroke has decreased and been increasingly replaced by HF and dementia as leading contributors. In contrast, stroke remains a more common contributor in men until they exceed 85 years of age, at which point the trends align with those observed in women. In younger men (<75), HF, myocardial infarction, stroke, and dementia are all modestly increasing as contributors to mortality. Conclusion: As individuals age, notable sex-based differences emerge in the cardiovascular factors contributing to mortality with AF. There are opportunities to enhance outcomes in women through targeted therapies aimed at reducing the risks of dementia and HF as they age. In younger men, early intervention to lower various cardiovascular risks may improve health outcomes, with continued efforts in later years focused on mitigating the risks of dementia and HF.

Background: Myocardial fibrosis assessed by cardiac magnetic resonance (CMR) with late gadolinium enhancement (LGE) imaging has emerged as a powerful predictor for sudden cardiac death (SCD) in dilated cardiomyopathy (DCM). However, in the cohort of DCM patients with LVEF <35%, isolated LGE may not exhibit significant difference between those with and without adverse outcomes. In such circumstances, morphological parameters may provide additional prognostic value for DCM patients with left ventricular ejection fraction (LVEF) <35%. The predictive value of LV remodeling index (LVRI), a novel geometry parameter derived from CMR, for ventricular tachyarrhythmia (VTA) in this population remains unclear. Purpose: To explore the predictive value of LVRI for VTA in patients with DCM with LVEF <35%. Methods: In this retrospective single-center study, consecutive DCM patients with LVEF <35% (n=271) who underwent CMR imaging were followed up. The study endpoint was VTA, including sudden cardiac death and major ventricular arrhythmias. The newly derived LVRI was defined as the cubic root of the LV end-diastolic volume divided by the maximal LV wall thickness. Competing risk regression analysis and Kaplan-Meier analysis were used to evaluate the association of LVRI with VTA. Results: During a median follow-up of 71 months (interquartile range: 17–134 months), 35 (12.9%, mean age 46.7 years, 27 males) participants reached VTA events. The presence of late gadolinium enhancement (LGE) (62.9% vs. 60.2%, p = .761) and LVEF (23.3±6 vs. 21.9±10.3, p = .197) were not significantly different between the patients with and without endpoint. Kaplan-Meier curve analysis showed that participants with LVRI ≥7.5 were more likely to experience VTA ( p < .0001). In the multiple competing risk analysis, LVRI ≥7.5 (HR, 2.496; 95% CI: 1.213-5.138; p = .013) was observed as an independent predictor of VTA after adjusting for age, sex and left bundle branch block. Conclusions: For nonischemic DCM patients with LVEF <35%, LVRI ≥7.5 was associated with lethal VTA events and provided incremental value over conventional CMR parameters.