Background: To evaluate the effectiveness of published nuclear cardiology training guidelines, the diagnostic accuracy of image interpretation by nuclear cardiology trainees was compared with that of experienced nuclear cardiologists.
Methods and results: The accuracy of three experienced nuclear cardiologists and three trainees with level II experience following Society of Nuclear Medicine/American College of Cardiology/American Society of Nuclear Cardiology guidelines in the interpretation of 114 exercise 99mTc-labeled sestamibi single-photon emission computed tomographic imaging studies was evaluated. Studies were selected randomly and included patients with less than 5% likelihood of coronary artery disease, as well as patients with angiographically demonstrated single and multivessel disease. Studies were interpreted by each reader without knowledge of clinical or exercise data. Each reader classified perfusion as normal or abnormal. Accuracy was assessed according to sensitivity, normalcy rate, and predictive accuracy. In addition, the ability of experienced readers and trainees to identify abnormal perfusion in patients with multivessel disease was compared. Trainees had high accuracy, comparable to experienced readers for sensitivity, normalcy rate, and predictive accuracy, as well as the ability to identify abnormal perfusion in patients with multivessel disease. In all categories, experienced interpretors demonstrated a trend toward greater accuracy with less observer variability than did trainees.
Conclusion: Structured training in nuclear cardiology following Society of Nuclear Medicine/American College of Cardiology/American Society of Nuclear Cardiology guidelines during clinical cardiology fellowship is effective, and trainees possess the skills to interpret myocardial perfusion images accurately. Interpretive skills can be expected to improve further with experience.