Comment on “Toward clinically actionable explainable AI in pulmonary arterial hypertension: Endpoints, calibration, and external validation”