PurposePulmonary hypertension (PH) is associated with a poor outcome in chronic obstructive pulmonary disease (COPD) and is diagnosed invasively. We aimed to assess the diagnostic accuracy and prognostic value of non-invasive cardiovascular magnetic resonance (CMR) models.MethodsPatients with COPD and suspected PH, who underwent CMR and right heart catheter (RHC) were identified. Three candidate models were assessed: 1, CMR-RV model, based on right ventricular (RV) mass and interventricular septal angle; 2, CMR PA/RV includes RV mass, septal angle and pulmonary artery (PA) measurements; 3, the Alpha index, based on RV ejection fraction and PA size.ResultsOf 102 COPD patients, 87 had PH. The CMR-PA/RV model had the strongest diagnostic accuracy (sensitivity 92%, specificity 80%, positive predictive value 96% and negative predictive value 63%, AUC 0.93, p<0.0001). Splitting RHC-mPAP, CMR-RV and CMR-PA/RV models by 35mmHg gave a significant difference in survival, with log-rank chi-squared 5.03, 5.47 and 7.10. RV mass and PA relative area change were the independent predictors of mortality at multivariate Cox regression (p=0.002 and 0.030).ConclusionCMR provides diagnostic and prognostic information in PH-COPD. The CMR-PA/RV model is useful for diagnosis, the RV mass index and PA relative area change are useful to assess prognosis.Key Points• Pulmonary hypertension is a marker of poor outcome in COPD.• MRI can predict invasively measured mean pulmonary artery pressure.• Cardiac MRI allows for estimation of survival in COPD.• Cardiac MRI may be useful for follow up or future trials.• MRI is potentially useful to assess pulmonary hypertension in patients with COPD.