Optimizing reliability of RBC signal oscillation measures from hyperpolarized 129Xe MRI

Recent advances to gas exchange hyperpolarized 129Xe MRI (Xe-MRI) have demonstrated that cardiogenic oscillations within the xenon red blood cell (RBC) signal are sensitive to pulmonary disease. Moreover, by implementing keyhole image reconstruction with gas exchange images collected using standard methodology, maps of regional oscillation amplitude can be generated. While such mapping has been demonstrated on a limited basis, validating these maps remains challenging due to the absence of easily measured biomarkers of pulmonary microvascular health. Moreover, as this is a very new technique, each of the previous implementations has used different methodology; it is unclear which of these methods provides optimal measures of regional oscillation amplitude. In this study, we evaluated oscillation mapping using the different published methods to determine which has the best same-day reliability. Because there are no easily obtainable measures of pulmonary vascular health, reliability serves as a valuable endpoint for validating that these maps are sensitive to real pulmonary physiology. We evaluated the same-day reliability of RBC oscillation measures in patients with systemic sclerosis (N = 6) and pulmonary arterial hypertension (PAH; N = 10), using single-time-point data from healthy volunteers (N = 9) to demonstrate "healthy" oscillation amplitude maps. Global measures of RBC oscillation amplitude (intraclass correlation coefficient; ICC = 0.88) had comparable reliability to standard xenon MRI measures (ICC ≥ 0.82). When examining oscillation mapping, some regional features showed disagreement across scans, but reliability of overall means was strong (ICC ≥ 0.86). Moreover, we show that recent advances in oscillation amplitude mapping for generating both amplitude and phase can provide equivalent maps to those methods that only provide amplitude. Overall, our findings demonstrate that Xe-MRI oscillation mapping has strong reliability when using optimized methods, even in participants with pulmonary disease.