OBJECTIVE: To determine the cost-effectiveness [CE] of non-invasive myocardial perfusion imaging [MPI] in asymptomatic individuals with diabetes mellitus [DM] for detection of CAD. METHODS: A decision tree analysis using TreeAge™ software was completed to determine the probabilistic CE of MPI in comparison to stress echocardiography [ECHO] and do-nothing management [DNM]. Probabilities for the analysis included cardiac event incidence such as non-fatal MI and cardiac death and percent of test results reported as abnormal/positive or normal/negative for each diagnostic technique. Utilities consisted of PPV and NPV values obtained from the published literature; preference scores [zero to one scale] assigned to disease states were similarly obtained. Medicare charges for 2004 were used as surrogates of cost; mean charges for initial clinical assessment [labs, etc.], exercise treadmill testing, ECHO, and MPI were all included in the CEA. RESULTS: MPI demonstrated PPV of 92.9% in detecting CAD in asymptomatic patients with DM compared to a stress ECHO PPV of 80%. However, ECHO is the dominant strategy with a mean cost of $701 and effectiveness ratio of 0.51. MPI demonstrated a mean cost of $1084 with an effectiveness ratio of 0.60. The resulting ICER is $4324/QALY. Importantly, DNM is dominated by MPI [p<.006] CONCLUSIONS: MPI is more effective and more costly than stress ECHO but meets generally-regarded WTP thresholds in asymptomatic individuals with DM. Use of diagnostic data from MPI may impact clinical outcomes such as cardiac events by influencing early management decisions.