Pulmonary arterial hypertension (PAH) is characterized by a progressive increase in pulmonary vascular resistance leading to right ventricular failure and eventually to death. The therapeutic strategy has become complex and needs to perform recurring follow up evaluations including right heart catheterizations (RHC). Cardiac magnetic resonance imaging (cMRI) has the advantage to accurately assess right ventricular volumes and important prognostic predictors such as cardiac index, stroke volume and right ventricular ejection fraction. The main objective of EVITA is to assess the hemodynamic diagnosis performances at baseline and at follow up visits of cMRI in comparison with the results of the RHC (current guidelines) to detect an unfavorable hemodynamic status. The primary endpoint is sensitivity and specificity of cMRI for the diagnosis of an unfavorable status defined by the current RHC criteria (with 95% confidence interval). The secondary objectives are 1) to identify clinical and hemodynamic variables independently contributing to prognosis, 2) to describe complications due to cMRI and to RHC, 3) to compare acceptability and tolerability of cMRI over RHC for the patient, 4) to constitute biological collection of blood samples to determine diagnostic and prognostic PAH biomarkers, 5) To compare the measurements of indexed stroke volume performed by RHC and by cMRI, 6) To evaluate the prognostic value to predict an unfavourable hemodynamic status of cMRI variables (including indexed stroke volume) after taking into account NYHA functional class, 6-minute walk test distance and BNP or NT-proBNP after 4 months of PAH treatment and 7) To evaluate the prognostic value to predict the first occurrence of morbimortality events of cMRI variables (including indexed stroke volume) after taking into account NYHA functional class, 6-minute walk test distance and BNP or NT-proBNP after 4 months of PAH treatment. PAH patients will be recruited in centers of the French network of severe pulmonary hypertension in a prospective cohort study. 180 subjects will be enrolled in the study: that size will give the study 90% power to find significant at the 5%-level. If the primary endpoint were achieved, since first, strategies and procedures planed in this project are consistent with those currently used in routine and second, inclusion criteria are not limited to a sub-population of PAH patients, positive results could allow to broadly extend our findings. Therefore, it will be possible to decrease the number of RHC, an invasive and cumbersome procedure without altering the prognosis. Moreover, all clinical procedures would be performed in outpatient clinics and thereby would reduce the cost to assess the severity of the disease. Current recommendations for evaluation of severity and follow-up being mainly derived from consensus of opinion of the experts, positive results will also improve the level evidence of severity assessment of PAH patients. According to secondary objectives we expect to better predict morbimortality events with cMRI compared to RHC.