Supplementary material, analysis code and PRISMA 2020 checklist for: Artificial Intelligence for Detecting Elevated Left-Sided Filling Pressure Against Catheter Measurement: A Systematic Review, Meta-Analysis, and Audit of Validation Reporting

Supporting files for a systematic review, meta-analysis and audit of validation reporting of artificial intelligence models that estimate elevated left-sided filling pressure from non-invasive inputs, evaluated against catheter-measured pressure. Contents: the Supplementary Material (search strategies and executed selection, study-level evidence inventory, quantitative eligibility and uncertainty, PROBAST+AI risk of bias, source discrepancies and exact operating counts, evidence confidence, Supplementary Figures S1 and S2); the completed PRISMA 2020 checklist; the meta-analysis code with embedded aggregate inputs (analysis_v8.py) and its output (results_v8.json); and the figure code (figures_v8.py). Screening, extraction and risk-of-bias appraisal were AI-assisted (OpenAI Codex; Anthropic Claude); discordant screening records were adjudicated by the first author, and the extent of human verification is stated in the article. The screening ledger and AI screening decisions are available from the corresponding author on request. No individual participant data are included.