Somatostatin receptor imaging (SRI) was carried out as part of the initial staging of 26 patients with histologically proven Hodgkin's (3) and non-Hodgkin's (23) lymphoma, and in the assessment of the first treatment's efficacy in seven of these patients. Static acquisitions over the whole body were performed 4 and 24 h after intravenous administration of 150 MBq of indium-111 pentetreotide. SRI data were compared with the results of conventional methods (clinical data, abdominal and thoracic computed tomography, bone marrow biopsy). Only 50 of the 86 (58%) confirmed extra-medullary tumour sites were detected by SRI. Twelve previously unknown localizations were visualized in seven patients. The Ann Arbor clinical stage was modified in only one of them. When tumoral tracer uptake was present, a tumour uptake index (TUI) was calculated using two regions of interest (one over the tumoral hot spot and one over the shoulder) on 24-h planar images. The patients were classified into three groups: high tumour uptake (TUI > 2.5 in all tumour sites, group A, six patients), low tumour uptake (1.5 < TUI < 2.5 in all tumour sites, group B, 18 patients), and no tumour uptake (group C, two patients). The sensitivity of SRI detection was higher in group A (90%) than in group B (52%) (P < 0.001). Six weeks after the fourth chemotherapy cycle, conventional methods and SRI were concordant in five of seven investigated cases (four complete remissions and one residual active thoracic mass showing tracer uptake), and discordant in two. SRI demonstrated residual tumoral tracer uptake in these two patients, who had previously been considered to be in complete remission. In conclusion, SRI does not seem to be reliable for the initial staging of lymphomas because of the highly variable and usually low tumoral tracer uptake. It may be more useful in the diagnosis of residual masses after treatment. However, further studies are needed to assess its specificity.