Objectives: Dynamic myocardial CT perfusion (CTP) enables the functional assessment of coronary artery disease. We aimed to assess the impact of iterative reconstruction using hybrid- (HIR) and model-based iterative reconstruction (IMR) - as compared to filtered back projection (FBP) on CTP-derived myocardial blood flow (MBF). Methods: In our observational study we analyzed 35 patients who underwent dynamic myocardial CTP for the assessment of myocardial ischemia. Images were reconstructed using FBP, HIR and IMR algorithms. Qualitative and quantitative analysis of myocardial perfusion was performed by visual assessment and measurement of MBF on all reconstructed datasets. Image quality and interobserver variability were also assessed. Results: Mean MBF was 105.61 +- 27.54mL/100 g min-1 for FBP, 105.07 +- 27.72mL/100 g min-1 for HIR and 103.69 +- 27.62mL/100 g min
1 for IMR, without significant differences between the algorithms (P 5 0.49). Image noise was significantly lower using novel iterative reconstructions compared to FBP (mean SDaorta: 48.3 +- 19.7 for FBP, 30.4 +- 13.3 for HIR and 19.5 +- 5.9 for IMR, P < 0.001 for all). Intraclass correlation coefficient for MBF was 0.96, while kappa value was 0.45 for visual assessment. Conclusion: Iterative reconstruction techniques reduce noise without altering MBF values of dynamic myocardial CTP. Therefore, further dose reduction may be achieved for dynamic myocardial CTP imaging.  Kulcsszavak: dynamic CT perfusion, myocardial blood flow, coronary computed tomography angiography, iterative reconstruction