Currently available methods for the estimation of disease activity in the orbits of patients with Graves' ophthalmopathy (GO) have the disadvantages of being either expensive or time consuming. The aim of this study was to develop a quick, reliable method using Tc labelled diethylenetriaminepentaacetic acid (DTPA) and four-headed single photon emission computed tomography (SPECT) for this purpose. The theoretical basis of the method is that the high capillarization and oedema in the orbit may be reflected on Tc-DTPA images in GO. SPECT data of 28 orbits of 14 patients with GO were compared with magnetic resonance imaging (MRI) T2 relaxation time scores. Based on the number of 'active' muscles with T2 relaxation times of more than 70 ms in a given orbit, an MRI score of 0 (no active muscle) to 4 (all rectus muscles active) was assigned to the orbit. With MRI, 18 orbits were inactive, and 10 were active. Thirty minutes after the intravenous administration of 7 MBq.kg Tc-DTPA, 128 projections were acquired by a four-headed SPECT. On the sum of six transaxial slices containing the entire bulbar region of the skull, a triangle-like region of interest (ROI) was drawn (OR ROI). This ROI was 'slipped' to the right temporal region of the brain as reference site (B ROI). The count ratios of OR/B were calculated and compared to MRI score values. In the group of 18 inactive orbits (an MRI score of zero on both sides), in the transaxial plane, the mean OR/B value was 6.4+/-1.17, and in the group of 10 active orbits (an MRI score of 1-3) 8.30+/-2.08, the difference being significant (P <0.05). Tc-DTPA orbital SPECT is a promising method for the estimation of disease activity in the orbits of patients with Graves' ophthalmopathy.