Aims: The aim of the study was to identify the characteristics of the culprit lesions compared to non-culprit lesions in patients with non-ST-elevation-myocardial infarction using dual energy computed tomography (DECT). Methods and results: In 29 patients, we identified 29 culprit lesions and 227 non-culprit lesions. Quantitative values such as the effective atomic number (effective-Z) and Hounsfield Units (HU) values were measured. Furthermore, all the lesions were characterised using characteristics such as composition (non-calcified, predominantly-non-calcified, predominantly-calcified, or calcified), presence of spotty calcification, remodelling index, and napkin ring sign. The mean effective-Z and HU values were significantly lower in culprit lesions than in non-culprit lesions (8.99 ± 1.21 vs 9.79 ± 1.52; p = 0.0066 and 87.41 ± 84.97 vs. 154.45 ± 176.13; p = 0.0447). The culprit lesions had a higher frequency of non-calcified plaques and predominantly non-calcified plaques, and were with a greater presence of napkin ring signs in comparison with non-culprit lesions. There were no differences in the presence of spotty calcification or remodelling index. By adding effective-Z to plaque characteristics such as non-calcified, positive remodelling, spotty calcification, and napkin rings we observed a significant increased sensitivity of detecting culprit lesions (65.5% vs.44.8%), but no significant changes in area under curve (AUC). Conclusion: The use of DECT adds new information of the plaque composition expressed by the effective-Z, which differs significantly in culprit lesions in comparison with non-culprit lesions. The use of the effective-Z improves the diagnostic sensitivity in detection of culprit lesions.