C66-11 Noninvasive Gas Exchange Determinants of Exercise Hemodynamics in Non-Severe Pulmonary Hypertension Associated With Interstitial Lung Disease

Abstract Background Level 3 cardiopulmonary exercise testing (CPET) can unmask the dynamic burden of pulmonary vascular disease (PVD), while submaximal exercise testing offers a simpler, noninvasive alternative. Emerging noninvasive markers such as GXCAP and VE/VCO2 may reflect PVD and correlate with invasive hemodynamics. Parikh et al. used Level 3 CPET to stratify PH-ILD patients with less severe resting hemodynamics into two groups: those with exercise-induced increases in pulmonary vascular resistance (PVRD) and those with static PVR (PVRS) during exercise. They further demonstrated that the PVRD group responded to inhaled treprostinil despite having a resting PVR below 4 WU, highlighting a potential role for exercise-based assessment in predicting treatment-responsiveness. This study aimed to evaluate whether GXCAP and VE/VCO2, measured during submaximal exercise testing, can reflect invasive findings and help identify non-severe PH-ILD patients with exercise-induced increases in PVR. Methods GXCAP and VE/VCO2 were retrospectively analyzed in 16 patients with non-severe PH-ILD who underwent both submaximal exercise testing and invasive Level 3 CPET at a tertiary pulmonary hypertension center. Patients had been previously classified by Parikh et al. as having dynamic (PVRD) or stable (PVRS) PVR responses based on invasive hemodynamics. We then evaluated whether GXCAP and VE/VCO2 could differentiate these two groups. Between-group differences were assessed using unpaired t-tests, and receiver operating characteristic analysis examined the ability of each variable to identify the PVRD phenotype associated with greater hemodynamic burden during exercise. Results Of the 16 patients, 6 were classified as PVRS and 10 as PVRD. Both groups had similar baseline characteristics. Mean GXCAP was significantly higher in the PVRS group (471 ± 85 mL·mmHg) compared with the PVRD group (311 ± 38 mL·mmHg, p 0.01). Similarly, mean VE/VCO2 was significantly lower in the PVRS group (33 ± 2) compared with the PVRD group (37 ± 2.5, p 0.01). ROC analysis confirmed excellent discriminatory ability for GXCAP (AUC 0.983), with an optimal cutoff of ≤ 349 identifying PVRD with 90% sensitivity and 100% specificity. VE/VCO2 also demonstrated good discrimination (AUC 0.900), with an optimal cutoff of ≥ 37. Notably, the GXCAP/VE/VCO2 ratio performed equivalently to GXCAP alone (AUC 0.983), with a threshold of ≤ 10 achieving 90% sensitivity and 100% specificity, underscoring its potential as a simple, intuitive composite marker. Conclusion Our study suggests that GXCAP and VE/VCO2 may serve as practical, noninvasive tools to detect exercise-induced pulmonary vascular impairment and help guide risk stratification and potential treatment responsiveness in patients with non-severe PH-ILD. This abstract is funded by: None