Atrial fibrillation (AF) represents a major risk factor of ischemic stroke recurrence with serious management implications. However, it often remains undiagnosed due to lack of standard or prolonged cardiac rhythm monitoring. We aim to create a novel end-to-end artificial intelligence (AI) model that uses MRI data to rapidly identify high AF risk in patients who suffer from an acute ischemic stroke. This study comprises an internal retrospective cohort and a prospective cohort from Shanghai sixth people's hospital to train and validate an MRI-based AI model. Between January 1, 2018 and December 31, 2021, 510 patients were retrospectively enrolled for algorithm development and performance was measured using fivefold cross-validation. Patients from this trial were registered with http://www.chictr.org.cn, ChiCTR2200056385. Between September 1, 2022 and July 31, 2023, 73 patients were prospectively enrolled for algorithm test. An external cohort of 175 patients from Huashan Hospital, Minhang Hospital, and Shanghai Tenth People's Hospital was also enrolled retrospectively for further model validation. A combined classifier leveraging pre-defined radiomics features and de novo features extracted by convolutional neural network (CNN) was proposed to identify underlying AF in acute ischemic stroke patients. Area under the curve (AUC), sensitivity, specificity, accuracy, positive predictive value, and negative predictive value were calculated for model evaluation. The top-performing combined classifier achieved an AUC of 0.94 (95% CI, 0.90–0.98) in the internal retrospective validation group, 0.85 (95% CI, 0.79–0.91) in the external validation group, and 0.87 (95% CI, 0.90–0.98) in the prospective test group. Based on subgroup analysis, the AI model performed well in female patients, patients with NIHSS > 4 or CHA₂DS₂-VASc ≤ 3, with the AUC of 0.91, 0.94, and 0.90, respectively. More importantly, our proposed model identified all the AF patients that were diagnosed with Holter monitoring during index stroke admission. Our work suggested a potential association between brain ischemic lesion pattern on MR images and underlying AF. Furthermore, with additional validation, the AI model we developed may serve as a rapid screening tool for AF in clinical practice of stroke units. This work was supported by grants from the 10.13039/501100001809 National Natural Science Foundation of China (NSFC, Grant Number: 81871102 and 82172068); 10.13039/501100008233 Shanghai Jiao Tong University School of Medicine , Two-Hundred Talent Program as Research Doctor (Grant Number: SBR202204); Municipal Science and Technology Commission Medical Innovation Project of Shanghai, (Grant/Award Number: 20Y11910200); 10.13039/501100010167 Research Physician Program of Shanghai Shen Kang Hospital Development Center (Grant Number: SHD2022CRD039) to Dr. Dong Huang and the SJTU Trans-med Awards Research (No. 20220101) to Dahong Qian.

Programmed cell death (PCD) pathways with druggable potential represent a promising but still underexplored frontier in heart failure (HF) research for diagnosis, prognosis, and therapy. To address this gap, we developed a Druggable Programmed Cell Death Index (DPCDI) through an integrative machine learning framework. An optimal combination of Lasso and Random Forest algorithms identified 15 pivotal genes (CALCOCO2, VPS13D, CLU, STAT3, OPTN, UBB, CXCL12, PPP1R15A, ATF4, IVNS1ABP, HMGB2, JAK2, EXOC7, ENO1, TPCN1) for DPCDI construction. Non-negative matrix factorization (NMF) analysis stratified HF patients into two distinct subtypes, with Subtype 2 exhibiting elevated apoptosis and mitophagy activity. Single-cell RNA sequencing revealed dynamic JAK2 and IVNS1ABP expression during cardiomyocyte state transitions, while CXCL12 showed stage-specific regulation in endothelial cells. Mendelian randomization analysis indicated that genetic predisposition to elevated JAK2 and STAT3 expression was associated with reduced HF risk, whereas CXCL12 overexpression increased susceptibility. Experimental validation in HF mouse models confirmed increased Cxcl12 and Jak2 expression and decreased Stat3 levels. Furthermore, knockout of Cxcl12, Jak2, and Stat3 induced HF phenotypes. Molecular docking identified pifithrin-α as a potent ligand for CXCL12 and strophanthidin for STAT3. Collectively, DPCDI provides a comprehensive framework for HF diagnosis, risk stratification, and targeted therapeutic development.

Background: It remains unclear whether long-term preoperative pharmacological therapy impacts the predictive value of left atrial appendage (LAA) radiomics for post-ablation atrial fibrillation (AF) recurrence. This study aims to explore the impact of preoperative beta-blocker therapy on an LAA radiomics and clinical features model for predicting post-ablation AF recurrence. Methods: Among 707 screened AF patients, 555 met eligibility criteria for this retrospective study. Patients were stratified by preoperative beta-blocker therapy. Baseline clinical and echocardiographic parameters were analyzed. A radiomics score (Rad-score) was calculated based on selected LAA features in the without beta-blocker group and integrated with clinical parameters into a nomogram model. The model performance was compared between groups. To address baseline clinical imbalances, inverse probability of treatment weighting (IPW) was applied. Following IPW adjustment, rad-score and beta-blocker interactions were evaluated via multivariable Cox regression, and Kaplan-Meier analysis assessed 12-month AF recurrence-free survival. Results: Overall, 91 (16.4%) patients experienced recurrence. In the without beta-blocker group ( n = 295, including 50 with recurrence), patients with recurrence exhibited reduced lymphocyte counts, elevated body mass index (BMI), and a lower prevalence of diabetes. These three parameters, along with a Rad-score derived from 11 LAA radiomic features, were integrated to develop a prediction nomogram. The model discrimination (Area Under the ROC Curve (AUC) = 0.742) dropped significantly in the with beta-blocker group (AUC = 0.501, P < 0.001). In the IPW-adjusted multivariable Cox analysis, Rad-score predicted recurrence in patients without beta-blocker therapy (Hazard Ratio (HR) = 2.73, P < 0.001), but not in users (HR = 0.91, P = 0.338). Rad-score significantly interacted with beta-blocker use ( P for Interaction < 0.001), successfully stratifying IPW-adjusted recurrence-free survival only in the without beta-blocker cohort (Log-rank P < 0.001). Conclusions: To ensure prognostic accuracy, preoperative medication history must be explicitly integrated as a core variable within comprehensive risk stratification models. However, the model requires multicenter external validation to confirm its broader applicability. Trial registration: Not applicable.