An important problem of myocardial SPECT/CT imaging with attenuation correction (AC) is a reduction of radiotracer activity at the apex in reconstructed images. This problem is known as the false apical defect. False apical defects mask real defects and lead to interpretation uncertainties. The aim of the present work is to study this phenomenon to understand the cause of apical defects on AC images. Mathematical modeling and clinical studies are carried out. Special mathematical models were designed to study the effect of the myocardial left ventricle (LV) form and the effect of patient constitution on false apical defects severity. Computer simulations of the SPECT/CT myocardial perfusion imaging procedure with and without attenuation correction were performed. The dependence of apical false defects on the LV form and on the data acquisition scheme was studied. The results of mathematical modeling have shown that the cause of false apical defects is not associated with attenuation correction. Rather, it is related to the LV form, especially to the presence of anatomical apical thinning. The LV form with apical thinning in diastole phase can not be exactly reconstructed by using the standard data acquisition schemes with limited 32 and 60 views. The reconstruction error appears in the rounded apical region in reconstructions with and without attenuation correction. But, in non-AC images, this error is masked by the second error related to the lack of attenuation correction. Interpretation based on the results of two compensating errors cannot be reliable. Taking into account the results obtained, we suggest the following hypothesis: gated systolic AC images can be useful in the interpretations of findings at the apex. This hypothesis was confirmed in preliminary clinical study.