Pulmonary toxicity has historically been a frequent complication of high-dose therapy with autologous stem cell transplant (ASCT), particularly with conditioning regimens containing 1,3-bis[2-chloroethyl]-1-nitrosurea (BCNU, or carmustine). Numerous studies in lymphomas, malignant gliomas and breast cancer have established a strong correlation between BCNU dose and pulmonary toxicity. Th e incidence of pulmonary toxicity is 10% when cumulative BCNU doses are 800 mg/m 2 as a single agent [1], but in combination with other cytotoxic drugs, particularly cyclophosphamide, the tolerated dose is lower. Several studies have demonstrated that doses of 600 mg/m 2 or higher are associated with an increased risk of pneumonitis [2 – 4], and for many years 450 mg/m 2 has been considered the upper dose limit for BCNU and is the standard dose used in the CBV regimen (cyclophosphamide, BCNU, etoposide). Despite decades of experience with high-dose BCNU in the ASCT setting, many aspects of BCNU-associated lung injury remain unclear, including the precise safe dose limit, the predisposing risk factors and the pathophysiology. Th us, studies attempting to elucidate these questions, such as the one by Lane and colleagues in this issue of Leukemia and Lymphoma [5], are welcomed. Before initiating any retrospective analysis on this condition, however, it is important to defi ne its characteristics, in view of the diverse spectrum of lung complications encountered after hematopoietic transplant. A recent statement [6] from the American Th oracic Society is helpful in this regard. Th e most common non-infectious lung injury in the ASCT setting shares several common features with the idiopathic pneumonia syndrome (IPS) that occurs after allogeneic hematopoietic transplant: pulmonary symptoms (dyspnea or cough), scattered or diffuse pulmonary infi ltrates on imaging, evidence of abnormal pulmonary physiology and, most commonly, diff use alveolar damage on histology, all in the absence of infectious, cardiac, renal or iatrogenic causes [7,8]. Th is ASCT-associated condition has been described as “ delayed pulmonary toxicity syndrome ” (DPTS) [9], and is distinct from IPS since it occurs later (median of day 45 vs. 19), has a much better prognosis ( 10% mortality rate, compared with 60 – 80%), is associated with specifi c drugs (BCNU, cyclophosphamide and cisplatin) and is highly responsive to corticosteroid therapy. Th e study by Lane and colleagues retrospectively examines the incidence and risk factors of pneumonitis in a population of patients with lymphoma treated at two centers with high-dose CBV and ASCT [5]. Th e authors appropriately designate pulmonary symptoms and diff use radiologic infi ltrates in the absence of infection as the cornerstones of their defi nition of pneumonitis. A more rigorous defi nition might also have included a decrease in D L CO (diff using capacity of the lung for carbon monoxide), although the more minimalist classifi cation may be adequate, since the selected patients did exhibit a 31% median reduction in D L CO. Th e strengths of the study include well-defi ned inclusion criteria and an analysis limited to a relatively homogeneous group of patients treated with a uniform conditioning regimen, thus minimizing potential confounding factors.