{"doi":"10.1016/j.jjcc.2025.05.014","title":"Performance of convolutional neural network-enhanced electrocardiography in detecting acute coronary syndrome: focusing on subtypes and reduced leads","abstract":null,"journal":"Journal of Cardiology","year":2025,"id":629745,"datarank":0.16479184330021646,"base_score":1.0986122886681096,"endowment":1.0986122886681096,"self_citation_contribution":0.16479184330021646,"citation_network_contribution":0.0,"self_endowment_contribution":0.16479184330021646,"citer_contribution":0.0,"corpus_percentile":null,"corpus_rank":null,"citation_count":2,"citer_count":0,"citers_with_citation_signal":0,"citers_with_endowment":0,"datacite_reuse_total":0,"is_dataset":false,"is_dataset_confidence":null,"is_data_producer":false,"deposit_databanks":null,"is_oa":false,"file_count":0,"downloads":0,"has_version_chain":false,"published_date":null,"fair_score":null,"fair_percentile":null,"algorithm_id":"datarank_citation_only_1hop_v6","ranking_scope":"data_only","authors":[{"id":19128,"name":"Shinya Suzuki","orcid":"0000-0002-1816-4260","position":1,"is_corresponding":false},{"id":1631079,"name":"Naomi Hirota","orcid":null,"position":2,"is_corresponding":false},{"id":1631081,"name":"Jun Motogi","orcid":null,"position":3,"is_corresponding":false},{"id":1631082,"name":"Takuya Umemoto","orcid":null,"position":4,"is_corresponding":false},{"id":1631084,"name":"Hiroshi Nakai","orcid":null,"position":5,"is_corresponding":false},{"id":1631085,"name":"Wataru Matsuzawa","orcid":null,"position":6,"is_corresponding":false},{"id":1631087,"name":"Tsuneo Takayanagi","orcid":null,"position":7,"is_corresponding":false},{"id":1631089,"name":"Akira Hyodo","orcid":null,"position":8,"is_corresponding":false},{"id":1631090,"name":"Keiichi Satoh","orcid":null,"position":9,"is_corresponding":false},{"id":1631092,"name":"Takuto Arita","orcid":null,"position":10,"is_corresponding":false},{"id":1631093,"name":"Naoharu Yagi","orcid":null,"position":11,"is_corresponding":false},{"id":1631094,"name":"Mikio Kishi","orcid":null,"position":12,"is_corresponding":false},{"id":1631095,"name":"Hiroto Kano","orcid":null,"position":13,"is_corresponding":false},{"id":1631097,"name":"Shunsuke Matsuno","orcid":null,"position":14,"is_corresponding":false},{"id":1631098,"name":"Yuko Kato","orcid":null,"position":15,"is_corresponding":false},{"id":1631099,"name":"Takayuki Otsuka","orcid":null,"position":16,"is_corresponding":false},{"id":1631100,"name":"Tokuhisa Uejima","orcid":null,"position":17,"is_corresponding":false},{"id":1631102,"name":"Junji Yajima","orcid":null,"position":18,"is_corresponding":false},{"id":1631103,"name":"Yasuo Okumura","orcid":null,"position":19,"is_corresponding":false},{"id":1631105,"name":"Yuji Oikawa","orcid":null,"position":20,"is_corresponding":false},{"id":1625444,"name":"Takeshi Yamashita","orcid":null,"position":21,"is_corresponding":false},{"id":1631077,"name":"Koichiro Hori","orcid":null,"position":0,"is_corresponding":false}],"reference_count":0,"raw_metadata":{"has_enrichment":true,"resolved":true,"title":"Performance of convolutional neural network-enhanced electrocardiography in detecting acute coronary syndrome: focusing on subtypes and reduced leads","abstract":"<h4>Background</h4>Early and accurate diagnosis of acute coronary syndrome (ACS), particularly non-ST-elevation ACS (NSTE-ACS), remains a critical challenge in emergency settings. Despite advancements in diagnostic modalities, conventional electrographic (ECG) interpretation often fails to detect subtle ischemic changes, particularly in NSTE-ACS, highlighting the need for artificial intelligence (AI)-driven approaches.<h4>Methods</h4>This study retrospectively analyzed data from a single-center cohort (Shinken Database 2010-2022, n = 32,167) to develop AI-driven ECG models for ACS detection. A convolutional neural network (CNN) model and an integrated neural network (INN) model, which incorporated diagnostic probabilities for ACS subtypes and target vessels, were evaluated using area under the receiver operating characteristics curve (AUROC), area under the precision-recall curve (AUPRC), sensitivity, and F1 scores for all‑lead ECG and reduced‑lead ECG models.<h4>Results</h4>The CNN model using all‑lead ECG achieved an AUROC of 0.877, an AUPRC of 0.391, and an F1 score of 0.184, while the INN model showed similar results (AUROC 0.889, AUPRC 0.356, and F1 score 0.188). For subtypes related to NSTE-ACS, the CNN model using all‑lead ECG (CNN model using double‑lead ECG) model achieved an AUROC of 0.785 (0.783), sensitivity of 0.723 (0.672), and specificity of 0.699 (0.768) for unstable angina, and an AUROC of 0.795 (0.786), sensitivity of 0.527 (0.567), and specificity of 0.878 (0.849) for NSTE-myocardial infarction. Among patients with troponin testing (n = 4169), the CNN model achieved a sensitivity of 76 %, a positive predictive rate (PPR) of 32 %, and an F1 score of 0.452, while the INN model achieved 78 %, 35 %, and 0.483, respectively. The leads I and II model demonstrated the highest AUROC among reduced‑lead models (0.866), with F1 scores in patients with troponin testing of 0.395 and 0.390 for the CNN and INN models, respectively.<h4>Conclusion</h4>Both CNN and INN-enhanced ECGs demonstrated good performance in detecting ACS including NSTE-ACS with subtle ischemic ECG changes. However, low PPR limit these models' standalone diagnostic utility. Instead, they hold promise as supportive tools, especially in resource-limited settings where reduced‑lead ECGs may be beneficial.","is_dataset_classified":null,"base_score":1.0986122886681096,"endowment":1.0986122886681096,"datacite_reuse_total":0,"file_count":0,"downloads":0,"views":0,"has_version_chain":false,"is_dataset":false,"is_oa":false,"pmid":"40447041","pmcid":null,"openalex_id":"https://openalex.org/W4410871657","authors":[],"funders":[{"funder_name":"AMED","grant_id":"JP17ek0210082","title":null}],"total_grants":1,"fwci":1.6146,"citation_percentile":0.83942309,"influential_citations":0,"citation_trend":[{"year":2025,"count":1},{"year":2026,"count":1}],"oa_status":"closed","license":"https://www.elsevier.com/legal/tdmrep-license","oa_locations":[{"url":"https://api.elsevier.com/content/article/PII:S0914508725001376?httpAccept=text/xml","host_type":"publisher"},{"url":"https://api.elsevier.com/content/article/PII:S0914508725001376?httpAccept=text/plain","host_type":"publisher"},{"url":"https://doi.org/10.1016/j.jjcc.2025.05.014","host_type":"journal"},{"url":"https://pubmed.ncbi.nlm.nih.gov/40447041","host_type":"repository"}],"fields_of_study":["ECG Monitoring and Analysis","Acute Myocardial Infarction Research","Cardiac electrophysiology and arrhythmias"],"mesh_terms":["Convolutional Neural Networks","Aged","Electrocardiography","Female","Humans","Male","Middle Aged","Retrospective Studies","ROC Curve","Neural Networks, Computer","Acute Coronary Syndrome"],"keywords":["Acute coronary syndrome","Medicine","Receiver operating characteristic","Internal medicine","Cardiology","Convolutional neural network","Myocardial infarction","ST elevation","Unstable angina","Artificial intelligence","Electrocardiography","Area under the curve","F1 score","Cutoff","Computer science"],"sdg_mappings":[{"sdg_number":0,"sdg_label":"Good health and well-being"}],"linked_datasets":[],"clinical_trials":[],"software_tools":[],"database_accessions":[],"source":"live","citation_network_status":"fetched"},"created_at":"2026-08-05T19:33:57.410014Z","pmid":null,"pmcid":null,"fwci":null,"citation_percentile":null,"influential_citations":0,"oa_status":null,"license":null,"views":0,"total_file_size_bytes":0,"version_count":0,"fair_f":null,"fair_a":null,"fair_i":null,"fair_r":null,"fair_zscore":null,"fair_rationale":null,"fair_model":null,"fair_agent_version":null,"fair_fulltext_source":null,"fair_has_llm":null,"fair_computed_at":null,"clinical_trials":[],"software_tools":[],"db_accessions":[],"linked_datasets":[],"topics":[]}