The ability to identify patients with severe coronary artery disease (CAD) by analysis of perfusion defects is limited. The lung/heart ratio (LHR) and transient ischaemic dilatation (TID) have been used for this purpose in thallium-201 scintigraphy. The value of these parameters in technetium-99m sestamibi single-photon emission tomography (SPET) imaging is controversial. In this study, therefore, we determined TID and LHR in a single-day rest/stress 99mTc-sestamibi SPET perfusion protocol and compared these measurements with perfusion defect size (PDS) and angiographic severity of CAD. Severe CAD was defined as >75% left main coronary stenosis and/or >90% proximal left anterior descending artery stenosis and/or >90% proximal stenosis in the left circumflex and right coronary arteries. LHR was determined from a stress anterior planar image recorded < or =6 min after exercise. TID ratio was derived from automatically calculated left ventricular rest/stress volumes, and PDS was measured based on semi-automated computer software (CEqual). Diagnostic accuracy and predictive values were compared between 22 patients with severe and 98 patients without severe CAD. LHRs showed a higher sensitivity (73%) for the assessment of severe CAD as compared to PDS and TID ratio (41% and 23% respectively, P<0.01), whereas specificity was highest for TID ratio [95%, P<0.01 when compared to PDS (84%) and LHR (82%)]. It is concluded that increased LHR in 99mTc-sestamibi myocardial perfusion imaging seems to yield good diagnostic accuracy in the detection of patients with severe CAD and may be derived from a single-day rest/stress study.