Exploratory multimodal artificial intelligence biomarker analysis for in-hospital mortality in adult patients with pneumonia: a proof-of-concept study – Anonymized Research Dataset (v2)

Anonymized multimodal research dataset from a single-centre retrospective cohort study (Hospital Alma Máter de Antioquia, Medellín, Colombia, 2024). Ethics approval IN82-2021. n = 27 patients (3 non-survivors, 24 survivors) with complete chest X-ray (CXR), ECG, and clinical note from admission for community-acquired pneumonia. V2 changes: Folder renamed from ecg_images to cxr_images (correct modality). Added ecg_extracted_points/ (644 anonymised ECG signal JSON files from 34 patients, E001–E034) and clinical_notes_anonymized.csv (27 patients, free-text admission notes with identifiers redacted). Three AI modules applied: (1) ResNet-18 CXR consolidation quantification (Class Activation Mapping); (2) Spanish NLP pipeline extracting IDSA/ATS severity scores; (3) NeuroKit2 HRV analysis of digitised ECG signals. Patient MRNs replaced with sequential codes P001–P027 (study cohort) and E001–E034 (ECG extracted-points cohort). No names, admission dates, or other direct identifiers present. Submitted to PLOS Digital Health (manuscript PDIG-D-25-00464).