Introduction Carcinoid heart disease (NET-CHD) is a frequent and adverse complication of carcinoid syndrome due to right ventricular (RV) failure. Medical therapy alone has a 2-year survival of approximately 20% and while surgical valve replacement is effective in improving symptoms and may increase survival, peri-operative risk remains approximately 15–20%. Transthoracic echocardiography (TTE) is considered the gold standard for assessment of NET-CHD and data on the role of cardiac magnetic resonance imaging (CMR) are limited despite recognised advantages in assessment of the right heart. The aim of this study was to assess the role of CMR in assessment of NET-CHD.

Methods This is a retrospective cohort study of 50 consecutive patients with proven NET referred with elevated NT pro-BNP to the European Centre of Excellence for Neuroendocrine Tumours in Birmingham between 2005–2015. At referral, all subjects underwent comprehensive left ventricular (LV) and RV assessment with CMR (1.5T Siemens Avanto), including deformation (Tissue Tracking, cvi42® Circle Cardiovascular Imaging), and late gadolinium enhancement (LGE).

Results In total, 36 patients were diagnosed with NET-CHD and 14 without (CHD-neg). Right sided valve disease was universal in NET-CHD: severe tricupid regurgitation (97%), severe pulmonary regurgitation (86%). On CMR, RV end-diastolic volume (EDV) and end-systolic volume (ESV) were increased (120 ± 30ml/m² vs. 67 ± 14 ml/m², p < 0.01; 49 ± 20ml/m² vs. 11 ± 3ml/m², p < 0.01) but with no difference in RVEF (60 ± 14% vs. 60 ± 9% p = 0.92). There was early evidence of ventricular-ventricular interaction, with reduction in both LVEDV (53 ± 16ml/m² vs. 72 ± 16 ml/m², p < 0.01) and LVESV (19 ± 10 ml/m² vs. 28 ± 16ml/m², p < 0.05) in NET-CHD, but no difference in LV ejection fraction (67 ± 8% vs. 63 ± 14%, p = 0.3). There was no difference in LV global longitudinal strain (GLS) or circumferential strain (GCS) between groups. RV LGE indicative of endocardial plaques was present in 6/36 (17%) but not observed in CHD-neg. Diffuse LV LGE was present 5 NET-CHD patients. Over follow up (median 1.3 years [0.6–3.1]), 20 patients with NET-CHD died. These patients had a lower GCS (14.8 ± 4.6% vs. 18.2 ± 4.4%, p < 0.05) and lower GLS (14.7 ± 4.7% vs. 18.3 ± 4.4%, p < 0.05) on CMR but no difference in LVEF, RVEF, LV volumes, RV volumes or NT-proBNP. In a logistic regression model, LV GLS remained an independent predictor of death.

Conclusion Significant increase in RV size and the presence of RV plaques are measurable on CMR early following referral with NET-CHD. This is sufficient to adversely affect LV filling and global deformation, which may contribute to effort intolerance and adverse cardiovascular outcomes.