The best method to measure right (RV) and left (LV) ventricle volumes of patients with corrected tetralogy of Fallot is considered cardiac magnetic resonance (CMR). However, to date, no standard protocol to measure RV volumes by CMR exists. RV volumes can be measured from a stack of short-axis slices or a stack of axial slices through the patient's chest. Therefore, the aim of this study was to determine whether short-axis or axial slices are more reliable for routine measurement of RV and LV volumes in patients with corrected tetralogy of Fallot. We studied consecutive patients with corrected tetralogy of Fallot (n = 46) undergoing routine CMR. The end-diastolic and end-systolic RV and LV volumes were measured by 2 investigators unaware of the results of the other measurements using short-axis and axial slices, and the inter- and intraobserver variances were compared. The design of the study was based on the Standards for Reporting of Diagnostic Accuracy. Interobserver variance was significantly smaller using axial slices than using short-axis slices for the RV end-systolic volumes (127.9%(2) vs 315.1%(2); p = 0.003), LV end-diastolic volumes (11.4%(2) vs 36.1%(2); p <0.001), and LV end-systolic volumes (31.9%(2) vs 176.1%(2); p <0.001). Intraobserver variance was significantly smaller using axial slices than using short-axis slices for the RV end-diastolic volumes (26.7%(2) vs 51.1%(2); p = 0.032), LV end-diastolic volumes (11.0%(2) vs 23.5%(2); p = 0.012), and LV end-systolic volumes (34.3%(2) vs 86.1%(2); p = 0.003). In conclusion, axial slices are more reproducible than short-axis slices for measuring ventricular volumes of patients with corrected tetralogy of Fallot by CMR.