Old criteria, new intelligence: The evolution of ECG in pulmonary hypertension diagnosis

Background: Pulmonary hypertension (PH) carries a significant mortality risk, highlighting the need for improved early detection strategies. This review examines the evolution of electrocardiographic assessment in PH diagnosis, from traditional criteria to artificial intelligence (AI) algorithms. Methods: We conducted a literature review analyzing 24 studies published between 1986 and 2025, including 18 traditional ECG validation studies and 6 AI-based investigations. Methodological quality was assessed using Quality Assessment of Diagnostic Accuracy Studies-2 (QUADAS-2) and QUADAS-AI frameworks. Studies were categorized by approach and diagnostic performance metrics were systematically evaluated. Results: Traditional ECG criteria demonstrated consistently high specificity (71–100 %) but low sensitivity (0–66 %) for detection of PH, limiting screening utility while maintaining confirmatory value. AI-based algorithms achieved superior balanced diagnostic performance with sensitivity of 74–85 % and specificity of 85 %. AI algorithms also demonstrated early detection capability, identifying PH up to 2–5 years before conventional clinical diagnosis. Conclusions: Traditional criteria retain value for diagnosis and response to therapy, while AI can be leveraged for early detection of PH. Implementation requires addressing computational infrastructure, healthcare provider training, and regulatory approval.