Aims Troponin-positive chest pain patients with unobstructed coronaries represent a clinical dilemma. Cardiovascular magnetic resonance (CMR) imaging has an increasingly prominent role in the assessment of these patients; however, its utility in addition to expert clinical judgement is unclear. We sought to determine the incremental diagnostic value of CMR and the heterogeneity in diagnoses by experienced cardiologists when presented with blinded clinical and investigative data in this population. Methods and results A total of 125 consecutive patients presenting to a tertiary centre between 2010 and 2014 with cardiac chest pain, elevated troponin (.29 ng/L), and unobstructed coronaries were enrolled and underwent CMR. A panel of three experienced cardiologists unaware of the CMR diagnosis and blinded to each other's assessment provided a diagnosis based on clinical and investigative findings. A consensus panel diagnosis was defined as two or more cardiologists sharing the same clinical diagnosis. Findings were classified into acute myocarditis, Takotsubo cardiomyopathy, acute myocardial infarction (AMI), or indeterminate. CMR provided a diagnosis in 87% of patients. Consensus panel diagnosis and CMR were concordant in 65/125 (52%) patients. There was an only moderate level of agreement between the three cardiologists (k= 0.47, P , 0.05) and a poor level of agreement between the consensus panel and CMR (k = 0.38, P , 0.05) with the most disagreement seen in patients with AMI diagnosed on CMR. Conclusion The clinical diagnosis of patients with non-obstructive coronaries and positive troponin remains a challenge. The concordance between CMR and clinical diagnosis is poor. CMR provides a diagnosis in majority of these patients.