This thesis examined the utility of the ultra-fast compressed sense (CS) magnetic resonance imaging (MRI) sequence in assessing cardiac contractile reserve in pulmonary arterial hypertension (PAH) patients during exercise. Additionally, the accuracy and reproducibility of the CS-MRI sequence against the current gold standard invasive catheterization technique in evaluating cardiac contractile reserve was assessed. The findings of four experiments conducted to achieve these aims are presented in this thesis. Experiment #1 was designed to examine the clinical feasibility and reliability of the CS-MRI sequence in comparison to the conventional balanced steady-state free precession cine (bSSFP) sequence for quantification of left ventricular (LV) volumes and mass. 50 patients with clinical indications for cardiac MRI were included in the study. Segmented MRI cine images were acquired on a 1.5 Tesla scanner in the LV short axis stack orientation using a retrospectively gated conventional bSSFP sequence followed by a prospectively triggered CS_bSSFP sequence with net acceleration factor of 8. Comparison of sequences was made in LV volumes and mass, image quality score, quantitative regional myocardial wall motion, and imaging time. Image quality was also assessed using published criteria. Differences (bSSFP minus CS_bSSFP, mean ± SD) and Pearson’s correlations were 14.8 ± 16.3 (P = 0.31) and r = 0.98 (P < 0.0001) for end-diastolic volume (EDV), 8.4 ±11.3 (P = 0.54) and r = 0.99 (P < 0.0001) for end-systolic volume (ESV), –0.4 ± 2.5 (P = 0.87) and r = 0.97 (P < 0.0001) for ejection fraction (EF), and –0.9 ± 11.8 (P = 0.92) and r = 0.97 (P < 0.0001) for LV mass. Bland–Altman analyses [bias and (limits of agreement)] revealed strong agreement in LVEDV [8.7 ml, (–12.1, 29.6)], LVESV [4.3 ml, (–11.9, 20.6)], LVEF [-0.02%, (–5.37, 5.33)], and myocardial mass [-6.1 g, (–14.7, 26.9)]. Image quality was comparable with a similar mean score between bSSFP and CS_bSSFP sequences (P = 0.42), with a good correlation in image quality observed (r = 0.68, P < 0.0001). Quantitative regional myocardial wall motion demonstrated strong correlation between the sequences (r = 0.87, P < 0.0001). Imaging time was significantly shorter for the CS_bSSFP sequence (1.1 ± 0.5 versus 5.6 ± 1.6 min, P < 0.0001). Collectively, these data suggest that the novel CS_bSSFP accurately and reliably quantitates LV volumes and mass, shortens acquisition times, and is clinically feasible. Experiment #2 was designed to examine the accuracy of cardiac output (CO) measured by CSMRI and to determine the inter-observer reproducibility on ventricular stroke volumes (SV) against the invasive gold standard right heart catheterization (RHC) at rest and during exercise. Ten PAH patients participated in the RHC and exercise CS-MRI protocols and five healthy volunteers participated only in the CS-MRI protocol. Before the cardiac MRI and RHC examinations, a graded maximal cardiopulmonary exercise test (CPET) was performed in supine position on an MRI cycle ergometer (Lode, Groningen, Netherlands) with an incremental protocol beginning at 0W and increasing progressively until exhaustion using a ramp protocol. Exercise cardiac MRI was then performed with subjects lying within the MRI bore using the same MRI cycle ergometer. Images were obtained in the resting state and then at two submaximal exercise workloads (low- and high-intensity exercise). Cardiac MRI exercise protocol consisted of a 3-min period of cycling in a supine position on an electromagnetically braked MRI cycle at the same submaximal workloads. Exercise RHC was performed in PAH patients within 24-hours of cardiac MRI. Right heart catheterisation was performed utilizing a right internal jugular vein approach. Haemodynamic measurements were recorded at rest and following 3 minutes of supine exercise at the same submaximal workloads as corresponding cardiac MRI. LVEDV and LVESV at high-intensity exercise decreased relative to baseline values, whereas right ventricular (RV) EDV and RVESV increased in PAH patients. There was excellent agreement (intraclass correlation coefficient, R = 0.92) and modest variability (coefficient of variation = 13%) between CO derived by CS-MRI and CO derived by the thermodilution method. There was a slight bias, with slightly larger CO measures being obtained by the thermodilution method. At rest, this bias was minor (0.45 L/min, representing 4.7% of mean volumes) and the differences increased marginally at highintensity exercise (0.79 L/min, 7.8% of mean volumes). The coefficient of variation for interobservational reliability remained low from rest to low- and high-intensity exercise for LV (2.5%, 4.0% and 3.9%, respectively) and RV (3.0%, 3.8% and 4.4%, respectively). RV contractility as measured by RVSV and ejection fraction (RVEF) augmented significantly from rest to exercise in normal controls, but not in PAH (PInteraction = 0.02 and PInteraction = 0.006, respectively). Using EF as the metric, RV contractile reserve from rest to high-intensity exercise was significantly higher in controls compared to PAH (10 ± 2.8% versus 3.7 ± 4.6%, respectively; p = 0.008). On the other hand, LV contractility as measured by LVSV, LVEF and LVESV from rest to exercise was not significantly different between controls and PAH. Experiment #3 was designed to examine and compare the effects of exercise on RV systolic function and ventriculo-arterial coupling ratio (VACR) in PAH and healthy subjects using cardiac MRI. Nine clinically stable patients with severe PAH and nine healthy controls were included in the study. Ventricular contractile reserve was defined as (ventricular function stress – ventricular function rest)/ventricular function rest and the VACR was calculated as SV/ESV. Resting RVSV and RVEF were significantly higher in controls (P < 0.05 for both). During exercise, both PAH and controls showed a similar and significant increase in heart rate (PAH versus controls, rest: 67 ± 6 versus 66 ± 16 beats·min−1, Ex1: 89 ± 10 versus 84 ± 21 beats·min−1, Ex2: 106 ± 16 versus 99 ± 23 beats·min−1; PInteraction = 0.261) and cardiac index (PAH versus controls, rest: 2.7 ± 0.6 versus 2.8 ± 0.8 L·min−1·m−2, Ex1: 4.0 ± 0.5 versus 4.1 ± 1.4 L·min−1·m−2, Ex2: 4.7 ± 0.7 versus 5.0 ± 1.6 L·min−1·m−2; PInteraction = 0.766). RV systolic function as measured by EF and ESV augmented significantly from rest to exercise in controls, but not in PAH. Using EF as the metric, RV contractile reserve from rest to Ex2 was significantly higher in controls (20 ± 13% versus −1 ± 16%; P = 0.032). At rest, VACR was similar between PAH and controls (0.9 ± 0.3 versus 1.2 ± 0.3, respectively; P = 0.075). However, during exercise VACR failed to increase in PAH patients, whereas a significant increase in VACR was observed in controls (PInteraction < 0.001). The change in VACR from rest to Ex2 was significantly higher in controls (66 ± 15% versus −7 ± 29%; P < 0.001). Collectively, these data demonstrated that patients with severe PAH were unable to significantly increase RV contractile function during exercise compared with controls and VACR assessed by exercise cardiac MRI may be a more sensitive index to identify severe PAH patients with early RV maladaptation. Experiment #4 was designed to examine the feasibility and reproducibility of biventricular myocardial deformation in PAH and controls using cardiac MRI feature tracking (cMRI-FT). Additionally, the ventriculo–ventricular interaction was compared between these subjects. Nine clinically stable patients with severe PAH and nine healthy controls were included in the study. RV and LV longitudinal strain (EllRV and EllLV) were derived using the mid-axial images. Radial (ErrLV) and circumferential strain (EccLV) were derived using the midventricular short-axis images. Relationships between strain and volumetric parameters were assessed at rest and during submaximal in-magnet exercise. Despite normal RVEF, patients with severe PAH had significantly lower EllRV at rest (−16.6 ± 2.7 versus −20.1 ± 3.6, P = 0.03). During exercise, RV systolic contractile reserve measured by EllRV was significantly reduced in PAH (PInteraction = 0.02). In PAH, indexed RV end-systolic volume (ESVi) significantly correlated with EccLV and ErrLV at rest (r = −0.65 and r = −0.70, P < 0.05) and with ErrLV during exercise (r = −0.43, P < 0.05). High observer agreement was also observed. Collectively, these data demonstrated that RV systolic function and contractile reserve as measured by EllRV was significantly reduced in patients with severe PAH and that the close relation between RVESVi with EccLV and ErrLV provides evidence of systolic ventriculo–ventricular interaction in PAH. The work presented in this thesis serves to expand our current knowledge and understanding of the importance of RV systolic function and contractile reserve in patients with severe PAH and how cardiac MRI-based exercise assessment can assist in this respect. Furthermore, the close relation between RVESVi with EccLV and ErrLV provides evidence of systolic ventriculo–ventricular interaction in PAH. These data not only indicate that CS-MRI could potentially be useful in risk stratifying patients with PAH, its utility in the prognostic assessment in patients with severe PAH warrants further investigation. Knowledge of right ventricular response to acute exercise will expand and deepen our understanding of the right ventriculo-pulmonary artery interaction and help clarify the questions surrounding treatment and intervention in patients with PAH.