Dose-intensive chemotherapy combined with localized or total body irradiation (TBI) followed by autologous or allogeneic hematologic stem-cell transplantation is increasingly being used to treat a number of solid tissue and hematologic malignancies. In the case of some hematologic cancers (e.g., AML, ALL, CML, and CLL), which are refractory to conventional dose chemotherapy, several studies now report that allogeneic transplantation offers increased cure rates and increased long-term survival (1–3). Even patients with nonmalignant chronic conditions, such as aplastic anemia, paroxysmal nocturnal hemoglobinuria, sickle cell, and thalassemias (4–6), as well as those with autoimmune disorders (7, 8), are being considered for hematogenous stem-cell transplantation. At some centers, autologous stem-cell transplantation after high-dose chemotherapy is frequently performed for advanced stage and high-risk breast cancer (9). Despite the acclaim and promise of this emerging combined pharmacologic and immunologic technology, side effects remain one of the major limitations to its success. In some cases, those side effects can limit an individual’s future therapeutic options. In a majority of regimens, noninfectious forms of pulmonary toxicity ( see Table 1) (infectious pneumonias are not addressed in this perspective) all too often inflict unwanted complications, resulting in increased morbidity and mortality. The incidence of noninfectious pulmonary toxicity after bone marrow transplantation has been estimated to be 12 to 70%, and much depends on the specific patient population being analyzed, the details of their respective therapeutic regimen, and the criteria used to define lung toxicity. In 1991, the National Heart, Lung, and Blood Institute held a workshop to review clinical-pathologic features of idiopathic pneumonia syndrome (IPS), a severe form of lung toxicity seen in about 12% of allogeneic marrow transplant recipients (10). The diagnosis of IPS in this setting requires evidence of widespread alveolar injury, as well as an absence of an active lower respiratory tract infection. However, less severe forms of lung toxicity, not meeting this criteria, often occur after autologous transplant. For example, delayed pulmonary toxicity syndrome (DPTS) develops in up to 64% of patients who undergo high-dose chemotherapy and autologous bone marrow transplantation for advanced stage breast cancer, and DPTS is best diagnosed on the basis of reductions in serial D L CO measurements (11). What challenges a clinical investigator’s ability to dissect out important factors in the development of lung toxicity is the myriad of variables specifically assigned to an individual’s treatment regimen. Table 2 outlines a partial list of variables that are likely to play important roles in the pathogenesis of post–BMT lung toxicity (12, 13). Further limiting our understanding of the pathophysiology of lung toxicity has been the lack of animal models. Recently, however, several groups have begun characterizing mouse models of IPS that appear to mimic the human disease process. Shankar and colleagues (14) at the University of Kentucky have developed a mouse model of IPS that utilizes a semi-allogeneic parental → F1 transplant strategy after a conditioning regimen of TBI to induce a mild form of graft versus host disease (GVHD). Several features of this model appear particularly relevant to the human condition. First, these animals develop progressive lung injury over a 3to 12-wk period that is pathologically characterized by prominent perivascular and peribronchiolar inflammation with diffuse alveolar interstitial mononuclear cell inflammation. Second, the acute phase of this GVHD appears mediated by CD8 1 cells, whereas CD4 1 cells are associated with the chronic form of GVHD. Third, these animals develop mild, localized interstitial fibrosis at the later time points analyzed. What features are necessary for the development of lung disease in this model? Shankar and associates (15) begin to answer this question in this current issue, but in doing so, generate additional questions. First, the kinetics of pulmonary injury seen in this model are unique when compared with that of other tissues (including liver, ear, skin, colon, and tongue) in that the temporal progression develops relatively slowly over a 9to 25-wk time frame, whereas the other extrapulmonary target organs analyzed showed maximal GVHD within 3 wk posttransplant followed by resolution over the remaining time frame. Second, preconditioning irradiation was found to be necessary for the development of IPS on the basis of several observations. The degree of lung injury showed irradiation dose-dependency. Animals that were not pretreated with conditioning irradiation showed no lung injury, whereas ( Received in original form April 13, 1999 )