We retrospectively evaluated the use of 18F-FDG PET for assessment of residual disease in 27 patients after therapy for malignant lymphoma. The images were evaluated qualitatively and quantitatively using standardized uptake values (SUV). All findings were validated either by biopsy or by clinical follow-up and compared with corresponding CT findings. The impact of blood glucose concentration, body weight, body surface area, lesion diameter and the time between injection and imaging on the SUVs were analysed. All 15 patients with biopsy-proven residual disease or relapse during follow-up and 11 of 12 patients who remained relapse-free were correctly identified by qualitative interpretation of the PET images. A case of pneumonitis after radiotherapy/chemotherapy accounted for the only false-positive finding. Compared with CT imaging, PET had a significantly higher specificity (P < 0.01), accuracy (P < 0.05) and positive predictive value (P < 0.05). The mean and maximum SUV of the tumour lesions were positively correlated to lesion diameter (P < 0.01) and imaging time post-injection (P < 0.01). Standardized uptake values corrected for the partial volume effect and normalized to a standardized imaging time (SUVBPT) were significantly higher (P < 0.05) in high-grade than in low-grade non-Hodgkin's lymphoma. In conclusion, 18F-FDG PET may help in the identification of patients who need additional treatment after the completion of conventional therapy. Qualitative image interpretation appears sufficient for this purpose.