Evaluation of the response to therapy is important for optimal selection of treatment strategy in patients with Hodgkin's disease (HD). Refractory disease requires intensive high-dose chemotherapy, whereas unnecessary treatment should be avoided in patients in complete remission. The purpose of this study was to evaluate the contribution of gallium-67 scintigraphy in predicting the clinical outcome in patients with HD and mediastinal involvement on the basis of scan results at the end of chemotherapy. Seventy-four patients with HD and mediastinal involvement were retrospectively investigated with 67Ga scintigraphy 72 h after injection of 220 MBq 67Ga citrate (planar and single-photon emission tomographic studies) following the completion of chemotherapy. At the same time, they all underwent computed tomography (CT). Patients were followed up for an average of 63 months (range 28-124 months). The disease status was newly diagnosed disease in 64 of the patients and relapse in 10. Systemic symptoms were absent (A) in 34 cases and present (B) in 40 cases. Forty-one patients had stage I or II disease and 33 patients had stage III or IV disease. Twenty-two patients had bulky disease on initial diagnosis. At the end of chemotherapy, all 74 patients showed regression of the mass by more than 50% (50%-100%) on CT. Patients were divided into two groups according to the positivity or negativity of the gallium scan after chemotherapy: 61 patients had negative and 13 patients had positive gallium scans. In the gallium-negative group, 19.7% of the patients relapsed and 91.8% were alive at the end of the follow-up. Relapse occurred in 20% of the patients with residual mass and in 19.6% of the patients without residual mass. In the gallium-positive group, 84.6% of the patients had recurrent disease and 61.5% were alive after intensive chemotherapy. There was a statistically significant difference in overall survival between patients with positive and patients with negative gallium results (P=0.0034). Disease-free survival differed significantly between patients with positive and patients with negative gallium scans at the end of chemotherapy (P<0.0001). The relative risk of death was 5.2 and the relative risk of relapse was 11.3 for patients with positive gallium scans, in comparison to those with negative gallium scans. The positive and negative predictive values for predicting relapse were 85% and 87%, respectively. It is concluded that even if gallium scan is performed at the end of chemotherapy, it can predict outcome. Alternative therapy may be required on the basis of gallium scan results obtained after treatment.