The aim of this study was to determine the yield of attenuation correction in myocardial perfusion imaging (MPI), before and after a 1-year experience period. In 48 consecutive patients referred for MPI, both non-corrected (NC) and attenuation-corrected (AC) images were analysed by three independent readers shortly after implementation of attenuation correction. The same images were re-analysed 1 year later, after having obtained experience in attenuation correction on a routine basis in >500 patients with clinical feedback. Results were compared with gold standards for ischaemia and infarction based on coronary angiography, follow-up, ultrasound and gated blood pool imaging. Sensitivity and specificity for the detection of coronary artery disease were 95% and 45% respectively for NC images and 77% and 70% respectively for AC images at first readings. After 1 year of AC experience, NC sensitivity/specificity was 100%/61%, and AC results were 89%/92%. It is concluded that attenuation correction improves the performance of MPI interpretation. With attenuation correction, specificity is increased and sensitivity is similar as compared with NC images. However, attenuation correction requires experience and is associated with a learning curve.