Abstract: K. Li, X. Zhou, Y. Jia, R. Wang, J. Pang, R. Shang, Y. Zhang, Y. Cui, D. Xu, M. Xiang (2027)

Myocardial ischemia may produce subtle abnormalities in cardiac electrophysiology that are not always fully captured by conventional surface electrocardiography, motivating the development of non-invasive magnetic-field-based signal-analysis approaches. Objective: Non-invasive characterization of myocardial-ischemia-associated magnetocardiographic patterns remains limited by the complexity and non-stationarity of cardiac magnetic signals. This study developed a multiscale nonlinear feature-integration framework based on the concatenation of global phase-space occupancy, operational nonlinear signal descriptors, and ST-T-localized representations. Methods: A total of 3182 one-second, 36-channel Spin-Exchange Relaxation-Free magnetocardiography recordings were available, including 1360 healthy-reference recordings and 1822 recordings from patients with a preexisting hospital-recorded clinical diagnosis of myocardial ischemia. Nine near-flat healthy-reference recordings were excluded by signal-quality control, leaving 3173 recordings for analysis. Global and ST-T-localized phase-space density features, recurrence-quantification indices, entropy, complexity, spectral descriptors, and related nonlinear signal features were extracted. The primary model was evaluated using five-fold nested stratified cross-validation with three-fold inner-loop hyperparameter optimization. All training-dependent preprocessing, resampling, model selection, and threshold selection were restricted to the corresponding training data. Results: The prespecified concatenated adaptive nonlinear feature configuration, evaluated using XGBoost and training-fold-only Synthetic Minority Over-sampling Technique, achieved a nested out-of-fold area under the receiver-operating-characteristic curve of 0.771 (95% confidence interval: 0.753-0.787) and an average precision of 0.796. Using thresholds selected within the corresponding training data, the model achieved an accuracy of 0.704, balanced accuracy of 0.702, sensitivity of 0.718, specificity of 0.686, precision of 0.755, and F1 score of 0.736. ST-T-localized features alone achieved an area under the curve of 0.750, compared with 0.687 for phase-space occupancy features and 0.668 for operational nonlinear signal descriptors. An amplitude-augmented sensitivity analysis yielded an area under the curve of 0.783. Conclusion: The proposed framework identified complementary global and ST-T-localized nonlinear MCG patterns that differed between stringently screened healthy-reference individuals and patients with a hospital-recorded clinical diagnosis of myocardial ischemia. The nested-validation analysis provides a conservative estimate of internal performance. Prospective external validation using clinically matched non-ischemic comparison groups and standardized reference criteria is required.

Please note: The abstracts of the bibliography database may underly other copyrights.

Ihr Browser versucht gerade eine Seite aus dem sogenannten Internet auszudrucken. Das Internet ist ein weltweites Netzwerk von Computern, das den Menschen ganz neue Möglichkeiten der Kommunikation bietet.

Da Politiker im Regelfall von neuen Dingen nichts verstehen, halten wir es für notwendig, sie davor zu schützen. Dies ist im beidseitigen Interesse, da unnötige Angstzustände bei Ihnen verhindert werden, ebenso wie es uns vor profilierungs- und machtsüchtigen Politikern schützt.

Sollten Sie der Meinung sein, dass Sie diese Internetseite dennoch sehen sollten, so können Sie jederzeit durch normalen Gebrauch eines Internetbrowsers darauf zugreifen. Dazu sind aber minimale Computerkenntnisse erforderlich. Sollten Sie diese nicht haben, vergessen Sie einfach dieses Internet und lassen uns in Ruhe.

Die Umgehung dieser Ausdrucksperre ist nach §95a UrhG verboten.

Mehr Informationen unter www.politiker-stopp.de.

Multiscale nonlinear feature integration for SERF-MCG-based ischemia-cohort discrimination