Nitrosourea-induced pulmonary injury

INTRODUCTION — The nitrosourea drugs, a class of DNA alkylating agents related to nitrogen mustards, include carmustine (BiCNU or BCNU, the most widely used nitrosourea), lomustine (CCNU), and fotemustine (available in some countries outside the United States). Pulmonary toxicity has been described following treatment with all of these agents [1-5].

Nitrosoureas are used predominantly to treat lymphoma, brain tumors, and melanoma, in which treatment usually involves repeated cycles for periods up to two years. Carmustine and lomustine are also used in single dose or short course preparative regimens for hematopoietic cell transplantation (HCT) in combination with other chemotherapy agents [6].

The clinical characteristics and management of pulmonary toxicity associated with nitrosoureas will be reviewed here. Pulmonary complications of allogeneic and autologous hematopoietic cell transplantation and pulmonary toxicity due to other antineoplastic agents are discussed separately. (See "Pulmonary complications after autologous hematopoietic cell transplantation" and "Pulmonary complications after allogeneic hematopoietic cell transplantation: Causes" and "Pulmonary toxicity associated with systemic antineoplastic therapy: Clinical presentation, diagnosis, and treatment" and "Pulmonary toxicity associated with antineoplastic therapy: Cytotoxic agents" and "Pulmonary toxicity associated with antineoplastic therapy: Molecularly targeted agents".)

IDIOPATHIC PNEUMONIA SYNDROME AFTER HEMATOPOIETIC CELL TRANSPLANT — The idiopathic pneumonia syndrome (IPS) occurs in patients who have undergone either allogeneic or autologous hematopoietic cell transplantation (HCT) and represents a heterogeneous group of noninfectious disorders associated with lung injury that have common pathologic findings of interstitial pneumonitis and/or diffuse alveolar damage.

Clinical features — IPS is characterized by widespread alveolar injury, defined as multilobar opacities on chest radiograph or computed tomography (CT) PLUS signs and symptoms of pneumonia PLUS evidence of abnormal pulmonary physiology manifested by an increased alveolar-arterial oxygen gradient or the need for supplemental oxygen [7]. The clinical presentation is typically an acute onset of dyspnea and cough, with or without fever, and rapid progression to hypoxemia and respiratory failure [8].

Timing — The onset of IPS is typically in the first four months following HCT with the majority of cases developing in the pre-engraftment or early post-engraftment phase with median times of onset of 12 to 19 days (range 0 to 112 days) [8-11].

Incidence and risk factors — IPS develops in 2 to 10 percent of patients undergoing allogeneic HCT and a smaller portion of those undergoing autologous HCT [7-9,12,13]. Risk factors for IPS include age less than 54 years, prior HCT, prior chest irradiation, total body irradiation as part of the conditioning regimen, presence of acute graft-versus-host disease, and the specific preparative conditioning regimen, including the dose of carmustine [8,9,14-16]. As examples:

●Among 210 children who underwent allogeneic HCT, 14 (7 percent) developed IPS and risk factors included prior HCT and acute graft-versus-host-disease [9].

●The rate of IPS following allogeneic HCT appears lower in patients who receive reduced intensity conditioning (RIC) regimens compared with a conventional myeloablative conditioning regimen (2.2 versus 8.4 percent) [16].

●In a series of 4917 patients undergoing autologous HCT for lymphoma, the incidence of IPS was 3 to 6 percent [14]. The incidence was higher among those who received cyclophosphamide, carmustine (approximately 450 mg/m²), and etoposide (CVB; high-dose carmustine, hazard ratio 1.9) with total body irradiation (TBI; HR 2.0), compared with carmustine (low-dose, approximately 300 mg/m²), etoposide, cytarabine, and melphalan (BEAM).

●A separate report of 94 patients with Hodgkin lymphoma who received high dose chemotherapy including carmustine with autologous HCT found that the incidence of IPS increased with increasing doses of carmustine [15]:

•15 percent incidence with doses less than 475 mg/m²

•32 percent with doses between 475 and 525 mg/m²

•47 percent with doses greater than 525 mg/m² (26 percent fatal)

Evaluation, diagnosis, and treatment — The evaluation, diagnosis, and treatment of IPS following allogeneic and autologous HCT are discussed separately. (See "Pulmonary complications after allogeneic hematopoietic cell transplantation: Causes", section on 'Idiopathic pneumonia syndrome' and "Pulmonary complications after autologous hematopoietic cell transplantation", section on 'Idiopathic pneumonia syndrome'.)

ACUTE AND SUBACUTE INTERSTITIAL PNEUMONITIS — Nitrosourea-induced interstitial pneumonitis in the absence of hematopoietic cell transplantation (HCT) is reported in approximately 10 to 30 percent of patients treated with carmustine (BCNU) for malignant glioma [17,18] and is also reported after repeated cycles of treatment with lomustine (CCNU) and fotemustine [5,19,20]. The onset is generally later than that described after HCT, but typically within 36 months of initiating treatment, and tends to be less fulminant than the IPS reported after HCT.

Incidence and risk factors in patients not undergoing HCT — The reported incidence of carmustine-induced pulmonary toxicity has ranged from 20 to 30 percent in some studies, to as low as 1 percent in others [1]. The low incidence of interstitial pneumonitis in some series may be due to a short observation period, low dose carmustine therapy, or the short median survival of patients who have received carmustine in the setting of a biologically aggressive, advanced malignancy [17].

Several factors have been identified as risk factors for lung toxicity following treatment with carmustine (BCNU) or lomustine (CCNU), including dose (at least for carmustine; risk is especially high for cumulative doses >1500mg/m² [21,22])concomitant radiation therapy, concomitant administration of cyclophosphamide, and pre-existing lung disease [3,18]. As examples:

●In the non-HCT setting, dose dependency of lung toxicity has been noted with carmustine (BCNU) [18,21,22]. In a study of 73 children with glioma, seven developed interstitial pneumonitis 3 to 11 months after adjuvant chemotherapy with carmustine, vincristine, and cisplatin [23]. The median cumulative dose of carmustine was 460 mg/m². Six of the seven patients died of progressive lung disease.

It is not known if exposure to a single dose of carmustine carries a greater or lesser risk of lung toxicity than exposure to the same total dose given in smaller aliquots over a period of one to two years.

●The risk of pneumonitis in patients treated with lomustine for brain cancer or fotemustine appears to be lower than with carmustine, but it is not clear that lower doses of either drug are associated with a lower risk:

•In two studies of anaplastic oligodendroglioma, a total of 280 patients received procarbazine, lomustine (110 to 130 mg/m² on day 1 every six weeks for four cycles), and vincristine (PCV); one patient developed grade 2 to 3 pneumonitis, but none developed pulmonary fibrosis; the median survival was over four years [24].

•A single case report describes one patient who developed progressive and fatal respiratory failure three months after receiving lomustine (three doses of 200 mg each at six week intervals) [19].

•In two treatment regimens for malignant glioma that included lomustine (90 to 130 mg/m² per cycle every six weeks), no lung toxicity was reported in the combined total of 204 patients followed up for 9 to 16 months [25,26], although conclusions about safety are limited by the relatively short duration of followup.

•Only a few case reports are available regarding pulmonary toxicity with fotemustine. A single case report describes acute interstitial lung disease that developed 20 days after the fourth course of fotemustine monotherapy for metastatic melanoma [5]. The condition regressed spontaneously following drug discontinuation. Another report describes rapidly fatal pulmonary toxicity in of one of 24 patients receiving dacarbazine plus fotemustine for advanced melanoma [27]; however, the attribution of the pulmonary toxicity (dacarbazine versus fotemustine) was not established.

Clinical features — The clinical presentation is heterogeneous. Most patients with nitrosourea-induced interstitial pneumonitis present within weeks to months of treatment with dry cough, breathlessness, and bilateral inspiratory crackles [18]. Others may present within days to weeks of carmustine treatment with a fulminant illness similar to acute respiratory distress syndrome (ARDS); however, the patient in this case report had received a high dose of carmustine (1255 mg/m² ) [28]. Still others may have subclinical disease detected by pulmonary function testing [2].

Evaluation — The evaluation of possible nitrosourea lung toxicity combines assessment of the severity of respiratory impairment with testing to exclude or include other processes. Broad spectrum microbiologic testing, radiographic evaluation, a trial of drug cessation, and bronchoalveolar lavage are the primary ways to narrow the differential diagnosis.

●Laboratory tests – The main role of laboratory testing is to determine whether other disease processes are contributing to the patient's respiratory compromise. Thus, complete blood cell counts, coagulation tests, B-type natriuretic peptide (BNP), blood cultures, sputum cultures, and viral serology can be helpful.

Arterial blood gas analysis reveals hypoxemia in advanced cases (with normal or low PaCO₂).

●Imaging studies – The chest radiograph and computed tomographic (CT) scan appearances are variable. Radiographic imaging most commonly demonstrates bibasilar reticular and micronodular opacities. In fulminant cases, the radiographic picture may suggest pulmonary edema or acute respiratory distress syndrome (ARDS) [18,28].

●Pulmonary function tests – The main role for pulmonary function testing is to assess the degree of respiratory impairment by measuring spirometry, diffusing capacity for carbon monoxide, and ambulatory oximetry.

Interstitial pneumonitis may be associated with reduced forced vital capacity (FVC) and total lung capacity (TLC). In addition, a reduced diffusing capacity for carbon monoxide (DLCO) is one of the most sensitive markers of early carmustine lung toxicity, and may occur when patients are asymptomatic [2,17,23]. Measurements of DLCO can be a very useful early marker of pulmonary toxicity, since diffusion abnormalities frequently precede the radiologic changes in affected patients [2]. (See "Diffusing capacity for carbon monoxide".)

●Bronchoscopy and lung biopsy – The main role of bronchoscopy and bronchoalveolar lavage (BAL) is to exclude other processes such as infection, diffuse alveolar hemorrhage, and lymphangitic spread of tumor. Lung biopsy is generally not necessary in suspected nitrosourea lung toxicity, unless symptoms and signs are rapidly progressive and refractory to drug cessation.

●Histopathology – Biopsy and autopsy reports have primarily described interstitial fibrosis with varying degrees of Type II cell hyperplasia, edema, and hyaline membrane formation [28,29]. Electron microscopic studies (performed as part of research) have revealed disappearance of type I pneumocytes, marked hyperplasia and hypertrophy of type II pneumocytes, and interstitial deposits of collagen. Increased expression of fibrogenic factors, including platelet-derived growth factor-B and transforming growth factor beta-1, has also been reported [30].

Diagnosis and differential diagnosis — The diagnosis of nitrosourea-induced interstitial pneumonitis is usually based on a history of drug exposure, a compatible clinical picture (cough, dyspnea, decrease in DLCO), and exclusion of alternative diagnoses. The differential diagnosis includes infection, radiation-induced lung injury, pulmonary edema, lung involvement by an underlying malignancy, and pulmonary thromboembolism. (See "Pulmonary toxicity associated with systemic antineoplastic therapy: Clinical presentation, diagnosis, and treatment", section on 'Diagnosis'.)

Treatment — The optimal treatment for suspected nitrosourea-induced lung toxicity is not known, although a key initial step is drug cessation for patients receiving ongoing cycles of nitrosourea-containing chemotherapy [5,17,18].

●Stopping therapy early based on pulmonary function tests – We monitor pulmonary function prior to and during treatment with either carmustine or lomustine, consistent with United States Food and Drug Administration (FDA)-approved prescribing information [31,32]. Specifically, we measure FVC and DLCO prior to each dose and withdraw the drug if there is a measurable fall (ie, >10 percent) in either [2,18].

Evidence in favor of a benefit to stopping carmustine at the first sign of lung toxicity comes from a case series in which the reported mortality associated with carmustine-induced interstitial pneumonitis was lower when the drug was stopped at the onset of respiratory symptoms or reduced pulmonary function [2,17]. However, mortality varies substantially among studies, ranging from 16 to 70 percent [18].

●Supportive care – In addition to stopping the nitrosourea, supportive care includes supplemental oxygen as needed. (See "Long-term supplemental oxygen therapy", section on 'Indications'.)

●Glucocorticoid treatment - Based on clinical experience, for patients with moderate to severe, early-onset nitrosourea-induced pulmonary toxicity (eg, dyspnea at rest, a decrease in oxygen saturation below 90 percent or more than 4 percent decrease from baseline, or worsening clinical status) and no evidence of infection, we suggest initiation of oral glucocorticoid therapy. We typically use a dose equivalent to prednisolone 40 to 60 mg per day, or 1 mg/kg per day, until symptoms and pulmonary function abnormalities improve or resolve [6,33-36].

Studies of glucocorticoid treatment have yielded mixed results, but occasional patients respond. In one series of 73 children receiving a carmustine-based chemotherapy regimen for glioma, seven developed interstitial pneumonitis; only one was successfully treated with glucocorticoids, whereas the others all died [23]. On the other hand, no benefit for glucocorticoids was suggested by case reports of patients who received lomustine for brain cancer and developed a rapid onset of respiratory symptoms and fatal respiratory failure despite dexamethasone 6 mg daily [19] or hydrocortisone 10 mg/kg every six hours [37]. High-dose glucocorticoids are often used in patients who develop pneumonitis following carmustine in preparation for HCT, although evidence in favor of benefit is minimal. (See "Pulmonary complications after allogeneic hematopoietic cell transplantation: Causes", section on 'Idiopathic pneumonia syndrome'.)

LATE ONSET PLEUROPARENCHYMAL FIBROELASTOSIS — Some patients develop a late onset of apical lung disease 2 to 17 years after treatment that is consistent with a rare sub-type of fibrotic interstitial lung disease known as pleuroparenchymal fibroelastosis (PPFE) (picture 1) [4,38-43]. PPFE can be idiopathic, or may be associated with a chemotherapeutic agent (including nitrosoureas and cyclophosphamide) [40,41], rheumatic disease (eg, Sjögren's disease, systemic sclerosis) [44], aluminosilicate exposure [45], or an idiopathic interstitial pneumonia [46,47]. It has also been described as a late complication of hematopoietic cell and lung transplantation [41].

The causality of nitrosoureas in this clinical picture of PPFE has been questioned, largely because of the idiopathic occurrence of the syndrome, lack of a clear dose response relationship, and progression to fibrosis only after a long treatment-free interval [41]. However, factors in favor of an association include the normalcy of pretherapy chest imaging [40], consistent reports of the same distinctive nitrosourea-associated disease from several countries [48], the notion that two separate alkylating agents have been associated with this condition [40,41], and the occurrence of PPFE in young individuals [4,42], in whom fibrotic conditions are relatively rare. The strongest evidence in favor of this association is that histological changes consistent with PPFE were found in all seven subjects who consented to lung biopsy 12 to 17 years after carmustine therapy [39]. (See "Interpretation of lung biopsy results in interstitial lung disease", section on 'Rare histopathologic interstitial pneumonia patterns'.)

Risk factors — The mechanism of PPFE due to nitrosoureas is not known and the severity is not always related to the dose of carmustine or lomustine [4,38-40,42,43]. However, in one series, all of the patients who developed lung toxicity consistent with PPFE had received at least 700 mg/ m² of carmustine, and 12 of 17 had received >1000 mg/ m², including 7 of 9 who died of pulmonary fibrosis [4,38,42,43]. Young age at treatment might also be a risk factor [42]. (See 'Clinical features' below.)

Clinical features — PPFE after carmustine or lomustine generally presents with cough, dyspnea, and sometimes chest pain, which may be pleuritic [4,40-43]. The onset of symptoms can be many years after completion of therapy. However, many subjects remain free of symptoms despite restrictive lung function and evidence of PPFE on CT scans and lung biopsies emphasizing the need for careful review of such patients even in the absence of symptoms [4].

The following clinical characteristics of late onset fibrosis are drawn from observation of a cohort of 17 pediatric survivors who had been treated with carmustine for brain tumors, six of whom died of lung fibrosis (four after a symptom-free interval of 7 to 12 years) [4]. Among the four cases with late onset fibrosis, the mean interval from treatment to first symptoms was nine years, but the interval from first symptoms to death was only one year; all died within two years of their first respiratory symptoms. All had been given 30 to 40 mg prednisone daily with no documented benefit.

Of the 11 long-term survivors, eight had detailed respiratory follow-up 13 to 17 years post treatment [4]. Most patients had no symptoms for several years, although the chest radiographs and lung function studies showed evidence of lung fibrosis in six of eight cases. All patients had restricted lung function but only one had crackles on chest auscultation. Four patients developed breathlessness or cough. Of the six who had prior spirometry, three had reductions in vital capacity from their initial measurement, but one symptomatic patient had unchanged spirometry over a five-year period.

In a later report of this cohort with three additional years of follow-up, two more patients died of pulmonary fibrosis [42]. Notably, there was a significant interaction between age at carmustine exposure and the risk of pulmonary fibrosis. In the entire cohort, of the eight patients who died of pulmonary fibrosis, the median age at treatment was 2.5; in contrast the nine-long-term survivors had a median age of treatment of 10 years.

Evaluation — The most important component of the evaluation of possible PPFE after nitrosourea therapy is high resolution computed tomography (HRCT).

●Imaging studies – The typical HRCT findings of PPFE include a dense, irregular apical pleural rim, wedge-shaped areas of consolidation that extend into the parenchyma along the septa, and sometimes cephalad retraction of the hila and adjacent traction bronchiectasis [40,41].

Changes consistent with PPFE were described in patients who were long-term survivors after childhood treatment with carmustine (BCNU) for gliomas: apical pleural thickening, spontaneous pneumothoraces (both usually many months after treatment), and fibrotic "stranding" have been described (image 1) [4,38,49]. HRCT scans in six patients showed coarse linear strands lying mainly in the upper lobes with some confluent sub-pleural bands that suggested pleural thickening (image 2) [4,38].

Gallium lung scans, which are not typically performed in this setting, were negative in all cases, consistent with noninflammatory fibrosis.

●Pulmonary function tests – Patients with late-presenting carmustine lung fibrosis have typically have restricted lung volumes (forced vital capacity [FVC] and total lung capacity [TLC] 50 to 60 percent of predicted values). However, unlike early onset carmustine fibrosis, the DLCO may be normal [4,40]. (See 'Evaluation' above.)

●Bronchoalveolar lavage (BAL) - BAL is generally not needed in the evaluation of PPFE, unless infection is suspected. In the few patients who have undergone BAL, no overall excess of polymorphonuclear cells or lymphocytes (or lymphocyte subsets) was observed on in the lavage fluid [4]. However, mononuclear cells yielded a lower than normal population of mature macrophages and a higher than normal population of immature monocytes.

●Lung pathology – Lung biopsy is generally not needed in patients with a clear history of nitrosourea exposure, typical imaging findings, and gradually progressive disease. When obtained, light microscopy reveals interstitial pulmonary fibrosis and elastosis but no inflammatory infiltrate [4,39,41]. The boundary between the pleuropulmonary fibrotic area and normal lung is usually sharp [41]. At autopsy, the lungs show subpleural apical fibrosis (picture 1). In one series, the type I (membranous) pneumocytes had been shed in some areas, leaving a bare basement membrane [4]. Type II cells had an increased number of lamellar bodies, and some contained multiple large vacuoles.

Diagnosis and differential diagnosis — The diagnosis of PPFE due to nitrosourea therapy is typically based on the history of exposure and the radiographic appearance on high-resolution CT [41]. (See 'Evaluation' above.).

The differential diagnosis focuses on disease processes that affect the upper lung zones in a subpleural pattern and on other causes of PPFE.

●It is important to exclude infection such as tuberculosis and aspergillosis, which have a predilection for the upper lung zones [41,50]. Ankylosing spondylitis can also cause apical fibrosis [41].

●PPFE can be idiopathic [51] or associated with the rheumatic diseases.

●PPFE can be caused by other chemotherapeutic agents (eg, cyclophosphamide [52,53]).

●Pneumonitis due to radiation therapy can cause a similar pattern, if the ports are compatible.

●Asbestosis can cause pleural plaques and interstitial fibrosis, but the changes are more commonly found in the lower lung zones than in the apex [41]. (See "Asbestos-related pleuropulmonary disease".)  

●Ergot alkaloid-related pleuropulmonary fibrosis also typically involves the lower lung zones.

Treatment and prognosis — No treatment has been shown to influence the course of PPFE due to nitrosoureas, and supportive care is the primary treatment. The absence of inflammation on CT scans and biopsies suggests that glucocorticoids and immunosuppressive therapy do not have a role in this PPFE. Lung transplantation offers the best hope of long-term survival for severely affected patients.

The rate of progression is variable with some patients progressing rapidly to respiratory failure [4] and others remaining stable for many years. The prognosis seems to be much worse for those exposed to nitrosoureas before the age of seven years. In one series, patients who were treated beyond the age of puberty had the best lung function and the slowest rate of decline [42].

In a series of nine long-term brain cancer survivors who had serial pulmonary function testing for up to 25 years, a gradual fall in vital capacity and total lung capacity was noted in 6, while 3 patients had stable lung function (figure 1) [43]. Nine patients eventually died of lung fibrosis, seven of late onset disease. The gradual decline in lung function over many years suggests that patients who have received carmustine are potentially at risk of late-presenting lung fibrosis, although as noted above, it appears to be most common when drug exposure occurs before the age of seven. (See 'Risk factors' above.)

RECALL PNEUMONITIS AFTER RADIOTHERAPY — Radiation recall pneumonitis refers to the development of interstitial pneumonitis following administration of certain antineoplastic agents to patients who have received prior radiation therapy to the lung. In a case report, a patient with lymphoma, who had received irradiation to the chest wall and right lung four and six months previously, developed pneumonitis about seven days after administration of carmustine [54]. Based on the chest radiograph, the area of lung injury was limited to the prior radiation therapy ports. (See "Radiation-induced lung injury", section on 'Prior radiation (recall)'.)

PREVENTION — The only way to completely avoid the risk of nitrosourea-induced pulmonary toxicity is to avoid the use of nitrosoureas. However, certain general measures may reduce the incidence and severity of lung toxicity:

●The lowest possible dose should be used because there is good evidence that higher doses are more likely to cause pulmonary toxicity [15,21,55]. As an example, most hematopoietic cell transplant centers avoid single doses of carmustine ≥475 mg/m² in the conditioning regimen [15]. (See 'Incidence and risk factors' above.)

●For patients receiving repeated cycles of nitrosourea chemotherapy, monitoring with serial measures of diffusing capacity for carbon monoxide (DLCO) can help identify early disease. If the DLCO decreases by ≥10 percent, the nitrosourea should be withheld. Additional therapy with glucocorticoids may be indicated. (See 'Treatment' above.)

SUMMARY AND RECOMMENDATIONS

●Spectrum of adverse pulmonary effects from nitrosourea drugs – The nitrosourea drugs are DNA alkylating agents that are associated with idiopathic pneumonia syndrome, interstitial pneumonitis (usually within three years of treatment) and late-onset fibrosis (pleuroparenchymal fibroelastosis [PPFE], often occurring many years after treatment). Carmustine (BCNU) is the most commonly used nitrosourea. (See 'Introduction' above.)

●Idiopathic pneumonia syndrome – The idiopathic pneumonia syndrome (IPS) develops in approximately 10 percent of patients up to four months after undergoing hematopoietic cell transplantation (HCT). While the etiology of IPS is likely multifactorial, carmustine and lomustine have been used in the preparative regimen for HCT in patients with Hodgkin lymphoma and other malignancies and may be contributing factors. (See 'Idiopathic pneumonia syndrome after hematopoietic cell transplant' above and "Pulmonary complications after allogeneic hematopoietic cell transplantation: Causes", section on 'Idiopathic pneumonia syndrome'.)

●Interstitial pneumonitis

•Risk factors – Risk factors for interstitial pneumonitis include higher doses (especially doses above 1500 m^g/m²), the concurrent administration of other agents that cause pulmonary toxicity (eg, cyclophosphamide), and possibly underlying lung disease, localized chest wall radiotherapy, and atopy. (See 'Acute and subacute interstitial pneumonitis' above.)

•Differential diagnosis – The diagnosis of early-onset nitrosourea-indued lung toxicity is usually established based upon a history of drug exposure and a compatible clinical picture, but it is a diagnosis of exclusion. The differential diagnosis includes radiation-induced lung injury, pulmonary edema, lung involvement by an underlying malignancy, and pulmonary thromboembolism. (See 'Diagnosis and differential diagnosis' above.)

•Monitoring and threshold for discontinuation – For patients receiving ongoing treatment (eg, repeated cycles for a solid tumor) with carmustine or lomustine, most experts, including the authors and editors of this topic review, measure forced vital capacity (FVC) and diffusing capacity for carbon monoxide (DLCO) prior to each dose and withdraw the drug if there is a measurable fall (ie, >10 percent) in either. (See 'Treatment' above.)

•Systemic glucocorticoid therapy – For those patients with moderate to severe, early-onset nitrosourea-induced pulmonary toxicity (eg, dyspnea at rest, a decrease in oxygen saturation below 90 percent or more than 4 percent decrease from baseline, or worsening clinical status) and no evidence of infection, we suggest administering systemic glucocorticoid therapy, rather than observation and supportive care alone (Grade 2C). The initial dose is the equivalent of oral prednisone 40 to 60 mg/day or 1 mg/kg daily. (See 'Treatment' above.)

●Late onset pleuroparenchymal fibroelastosis – High doses of carmustine are also a risk factor for late-onset PPFE, which is fatal in approximately 50 percent and associated with radiographic or physiologic evidence of pulmonary fibrosis in long-term survivors. (See 'Late onset pleuroparenchymal fibroelastosis' above.)

•Clinical and radiographic features – Late onset PPFE typically presents with gradual onset of dyspnea and cough several years following cessation of carmustine. The main radiographic findings are a dense, irregular apical pleural rim, wedge-shaped areas of consolidation that extend into the parenchyma along the septa, and sometimes cephalad retraction of the hila and adjacent traction bronchiectasis. Pulmonary function tests show a restrictive pattern. (See 'Clinical features' above and 'Evaluation' above.)

•Diagnosis – The diagnosis of PPFE is usually based on the clinical features, history of nitrosourea exposure, and the HRCT pattern. (See 'Diagnosis and differential diagnosis' above.)

•Treatment – No treatment has been shown to influence the course of late nitrosourea-induced PPFE. Based on the histopathologic appearance of fibrosis without inflammation, glucocorticoids do not have a role. Lung transplantation may be an option. (See 'Treatment and prognosis' above.)

ACKNOWLEDGMENT — The UpToDate editorial staff acknowledges Ronan O’Driscoll, BSc, MD, FRCP, who contributed to earlier versions of this topic review.