Introduction The term ‘lymphoma’ refers to a primary malignant tu- mour of the immune system. It encompasses a spectrum of diseases among which Hodgkin’s disease (HD) repre- sents a specific clinical-pathological entity and is distinct from the heterogeneous group of lymphoproliferative disorders which are collectively termed non-Hodgkin lymphomas (NHLs). Malignant lymphomas should be differentiated from other non-malignant lymphomatous disorders that can occur in the chest, such as Castle- man’s disease, lymphomatoid granulomatosis and lym- phoid interstitial pneumonia [1]; these will not be dis- cussed in this article. HD accounts for approximately 0.5–1 % of all newly diagnosed cancers in adults and shows a characteristic, and unexplained, bimodal peak in the third and sixth decades of life; it is more prevalent in males. NHLs are approximately 4 times more common than HD. Males are affected more commonly than females at all ages; the median age at the time of diagnosis is 55 years [2, 3]. The aetiology of malignant lymphomas, like that of other malignancies, remains unknown, although a possi- ble role of Epstein-Barr virus and some retroviruses in determining the onset of the disease has been proposed. However, no definite cause-effect relationship has been established. It is well known that patients whose im- mune system is impaired and patients with acquired im- mune deficiency syndrome (AIDS) are at a higher risk for developing malignant disorders (40–100 times more than the normal population) and that NHLs are particu- larly common in this group of patients [4–7]. Pathology The pathological classification of HD was developed by Lukes and Butler in 1966 [8], with subsequent slight modification at the RYE conference in 1971, and it is still accepted all over the world (Table 1). The nodular sclerosis subtype is the most common, occurring in up to three-quarters of all cases in different series. The pathological classification of NHLs is more com- plex as the spectrum of histological types is broader than that of HD. The Working Formulation, a recent pathological classification of NHLs [9], has increasingly supplanted the Rappaport classification introduced in 1966 and is widely accepted, especially in the United States. The Working Formulation groups the histologi- cal subtypes into low-, intermediate- and high-grade lymphomas, thus providing improvements in determin- ing clinical behaviour, therapeutic approaches and prog- nosis. The Kiel classification, which is focused on the or- igin of the malignant cell, is used almost exclusively in Europe [10, 11]. Table 2 shows the correlation among the three most important classifications of NHLs. Eur. Radiol. 7, 1179–1189 (1997)  Springer-Verlag 1997 Chest radiology European Radiology R eview article Staging of thoracic lymphoma by radiological imaging L.Bonomo, C.Ciccotosto, A.Guidotti, B.Merlino, M.L.Storto Istituto di Scienze Radiologiche, Universita ` ‘G.D’Annunzio’, Ospedale ‘SS. Annunziata’ Via PA Valignani 1, I-66100 Chieti, Italy Received 4 April 1996; Revision received 7 November 1996; Accepted 7 November 1996 Abstract. Thoracic lymphomas, which are very com- mon especially in Hodgkin’s disease patients, are characterised by enlargement of mediastinal lymph nodes, parenchymal abnormalities, and pleural, peri- cardial and chest wall involvement. The use of several imaging techniques has been proposed in order to as- sess the extent of the disease correctly and to plan therapy. The most relevant results in this field, espe- cially those using computed tomography (CT), mag- netic resonance imaging (MRI) and gallium scanning, are summarised in this review. Presently CT is widely and successfully used in staging patients, whereas MRI seems to be preferable, as a second-step tech- nique, if pericardial, pleural and chest wall involve- ment are suspected. The role of gallium scanning is limited in the staging, although it could be relevant in the follow-up of treated patients. Key words: Chest, lymphoma – Lymphoma, CT – Lymphoma, MRI – Lymphoma, radionuclide studies Correspondence to: L. Bonomo