A protocol for validation of novel artificial intelligence-based framework for dyspnoea investigation with cardiopulmonary exercise testing.

Background: Exertional dyspnoea largely represents the sensory translation of an ever-growing dynamic mismatch between ventilatory demand and capacity as exercise intensifies. This fundamental tenet, however, has not been formally incorporated into data display and clinical interpretation of incremental cardiopulmonary exercise testing (CPET). The objectives of the present study were to validate a novel framework (Dynamic Assessment of Dyspnoea and Ventilation on Exercise (DyVe-X)) to quantify the severity of exertional dyspnoea while exposing its mechanical-ventilatory underpinnings across progressively higher exercise intensities in subjects showing varied pre-test likelihood of abnormality. Methods: Incremental CPET data from three large cohorts of tobacco-exposed subjects: smokers and ex-smokers at risk for, or at different stages of, COPD (n=1161). Based on recently published normative data, a homonymous software (DyVe-X 1.0.0 Kingston, ON, Canada) uses an algorithm based on artificial intelligence techniques to establish the burden of 1) exertional dyspnoea Implications: This study may validate an original framework for investigating dyspnoea using CPET, characterised by continuous assessment of symptom intensity and time course, in light of the rate of dynamic depletion of mechanical-ventilatory reserves.