- Email: [email protected]
Diffuse alveolar hemorrhage in patients with hematological malignancies: HRCT patterns of pulmonary involvement and disease course
Clinical Imaging 37 (2013) 680–686
Contents lists available at SciVerse ScienceDirect
Clinical Imaging journal homepage: http://www.clinicalimaging.org
Diffuse alveolar hemorrhage in patients with hematological malignancies: HRCT patterns of pulmonary involvement and disease course Daniel Spira a,⁎, Stefan Wirths b, Felix Skowronski a, Jan Pintoffl b, Sascha Kaufmann a, Harald Brodoefel c, Marius Horger a a b c
Department of Diagnostic and Interventional Radiology, Eberhard-Karls-University, 72076 Tübingen, Germany Department of Oncology and Hematology, Eberhard-Karls-University, 72076 Tübingen, Germany Department of Radiology, Harvard Medical School, Beth Israel Deaconess Medical Center, WCC 308A, Boston, MA 02215, USA
a r t i c l e
i n f o
Article history: Received 27 June 2012 Received in revised form 25 August 2012 Accepted 7 November 2012 Keywords: High-resolution CT Drugs Reactions Hematological malignancies Diffuse alveolar hemorrhage
a b s t r a c t Objective: To analyze high-resolution computed tomography (HRCT) patterns of lung involvement and disease course in patients with hematological malignancies experiencing diffuse alveolar hemorrhage (DAH) after chemotherapy ±allogeneic stem cell transplantation (allo-SCT). Materials and methods: Sixteen patients experiencing DAH after chemotherapy±allo-SCT were enrolled. A total of 74 computed tomography (CT) scans obtained before, during, and after onset of DAH were evaluated retrospectively. Results: CT features of DAH are each, by oneself, nonspecific. However, conjoint bilateral, diffuse, and dependent ground glass opacity±crazy paving, accompanied by airspace bronchograms, should suggest this complication. The HRCT course comprises a wide range of trends that are not predictive for patient's outcome, but progression of parenchymal consolidations at follow up was more often detrimental. © 2013 Elsevier Inc. All rights reserved.
1. Introduction An abundance of noninfectious pulmonary complications is induced by chemotherapy and hematopoietic (particularly allogeneic) allogeneic stem cell transplantation (allo-SCT), commonly including diffuse alveolar damage (DAD), pulmonary edema, idiopathic pneumonia syndrome (IPS), diffuse alveolar hemorrhage (DAH), and manifestations related to acute pulmonary graft-versus-host disease (GVHD) [1–4]. DAH in the allo-SCT setting holds a poor prognosis with a mortality rate ranging between 64% and 100% [5]. It is characterized by multilobar pulmonary infiltrates, hypoxemia, absence of infection compatible with the diagnosis, progressively bloodier return in bronchoalveolar lavage (BAL), and the presence of blood in ≥ 30% of alveolar surfaces at autopsy [5,6]. The exact pathogenesis still remains elusive although capillary endothelial damage by chemotherapy followed by intravascular microthrombosis and transendothelial neutrophil engraftment following alloSCT is known to cause leakage of red blood cells into the pulmonary alveoli [7]. Clinical symptoms and signs are those of pneumonia, that is, dyspnea, cough, fever, tachycardia, and rarely hemoptysis [8].
⁎ Corresponding author. Department of Diagnostic and Interventional Radiology, Eberhard-Karls-University, Hoppe-Seyler-Str. 3, 72076 Tübingen, Germany. Tel.: +497071-2987212; fax: +49-7071-295845. E-mail address: [email protected] (D. Spira). 0899-7071/$ – see front matter © 2013 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.clinimag.2012.11.005
Accurate diagnosis and close monitoring of DAH via imaging is imperative for patient care. The radiographic appearance of DAH is described as unspecific with an interstitial or alveolar pattern of lung opacification primarily involving the central portion of the lung with predilection for the middle and lower lung zones [9]. High-resolution computed tomography (HRCT) is the most accurate imaging method displaying parenchymal lung involvement and is ideally suited to recognize and follow the extent and macroscopic morphology of DAH. However, apart from several excellent review articles and book chapters overviewing the field of DAH [10–12], we did not find a comprehensive analysis of HRCT findings in patients with hematological malignancies developing DAH. Thus, we analyzed lung involvement and disease course by HRCT in patients with hematological malignancies who developed DAH after chemotherapy±allo-SCT. Our work addresses two principle points: to determine the leading computed tomography (CT) morphologic finding or combination of findings and their prevalence in this patient group and to trace the temporal course of lung parenchymal changes. 2. Methods 2.1. Patient demographics This retrospective study was approved by our institutional review board that waived informed consent. Between May 2008 and
D. Spira et al. / Clinical Imaging 37 (2013) 680–686
December 2011, 16 patients with hematological malignancies (11 men, 5 women; age range: 21–70 years, mean: 46.4 years) who developed DAH after chemotherapy±allo-SCT were enrolled (Table 1). Eight patients underwent allo-SCT before the development of DAH. All patients were examined for signs of acute GVHD in the skin, the liver, or the gut after allo-SCT. 2.2. Inclusion criteria DAH was defined (according to current guidelines of National Institutes of Health) as the sudden development of multilobar infiltrates, symptoms and signs of pneumonia with abnormal pulmonary physiology, the absence of active lower respiratory tract infection, and the presence of hemorrhage as determined by BAL or retrospectively by autopsy [5]. In our study, the following additional criteria were considered necessary for confirmation of DAH and subsequent patient enrollment: (a) BAL at ≤ 2 days after the beginning of symptoms showing progressively bloodier return (n= 10); (b) autopsy revealing blood in N30% of the alveolar surfaces of lung tissue (n= 1); or (c) the combination of thrombocytopenia (b50.000/μl), a sudden drop in hemoglobin of ≥ 1g/dl, multilobar infiltrates, and an acute drop in oxygen saturation accompanied by clinically evident hemoptysis (n= 3) or generalized bleeding diathesis (i.e., massive epistaxis, cutaneous, and retinal bleeding) (n= 2) were considered diagnostic for DAH. 2.3. Exclusion criteria Patients with evidence of viral, bacterial or fungal infection, vasculitis, connective tissue disorders, trauma, as well as those with evident focal bleeding to the lung from bronchial or pulmonary vessels were excluded from the analysis. 2.4. CT evaluation All CT examinations were performed on either a 64- or a 128-row multidetector CT (Sensation 64/Definition+; Siemens, Erlangen, Germany) using a 250–330-mm field of view, a 512×512 reconstruction matrix, 120 kV, 100–120 effective mAs, and a tube rotation time of 0.5/0.3 ms. No intravenous contrast material was applied. A single spiral acquisition was obtained from the apex to the base during one breath-hold at suspended end-inspiratory volume. Examinations were performed with patients in the supine position. Reconstruction of the data was done at 1-mm slices with a sharp reconstruction algorithm (filter, B70, equivalent to HRCT) and 1-mm
Table 1 Treatment regimens Patient Disease
Treatment regimen
1
Multiple myeloma
2 3 4 5 6 7
AML Mycosis fungoides AML ALL AML Mantle cell lymphoma AML AML AML T-NHL AML AML AML ALL AML
Total body irradiation, busulfan, cyclophosphamide Clofabrine, thiotepa, melphalan Fludarabine, melphalan, MabCampath Total body irradiation, fludarabine Clofabrine, thiotepa, melphalan Cyclophosphamide, anti-thymocyte globulin Fludarabine, melphalan
8 9 10 11 12 13 14 15 16
Cytarabin, idarubicin Clofabrine, thiotepa, melphalan Total body irradiation, cyclophosphamide Total body irradiation, etoposide Fludarabine, cytarabine, amsacrine Cytarabine Clofabrine, THIOTEPA, MELPHALAN Treosulfane, etoposide, cyclophosphamide Total body irradiation, cyclophosphamide
681
reconstruction increment for visual assessment. Systems were calibrated on air daily [13]. 2.5. Clinical and laboratory findings The time interval from radio-/chemotherapy as well as allo-SCT to the development of DAH was recorded. At onset of DAH, peripheral blood platelet count, C-reactive protein (CRP), clinical presentation, and the presence/absence of signs of acute GVHD were assessed in all patients. 2.6. Chest CT evaluation of lung parenchymal abnormalities Images were viewed at lung (window width 1200 HU, window level 600 HU) and mediastinal (window width 350 HU, window level 50 HU) window settings. CT images were reevaluated retrospectively by two radiologists with 17 and 4 years (MH, DS) experience in reading chest CT via consensus reading. Both investigators were blinded to the original interpretations and mortality data. Criteria for HRCT findings were those defined by the Fleischner Society's glossary of terms for chest CT [14]. Ground glass opacity (GGO) was defined as a hazy increase in lung attenuation with preservation of bronchial and vascular margins being distributed either diffuse or patchy (Fig. 1A). Airspace consolidation was defined as an area of dense increase in attenuation with obscuration of the underlying vessels and airway walls, showing different morphology either in form of segmental or subsegmental or patchy opacifications (Fig. 1B). Reticulation was defined as an interlacing line shadow suggesting a mesh or net (Fig. 1C). Crazy paving was defined as a superimposition of GGO and reticulation (Fig. 1D). GGO, airspace consolidation, reticulation, or crazy paving, distributed throughout the parenchyma without zonal predominance, was called diffuse, while focal parenchymal infiltrates with lobular, segmental, or lobar distribution involving one or both lungs were named focal. If more or less sharply demarcated regions of different density (to the point of unattained secondary lobules) were noted within infiltrated lung parenchyma, the term mosaic pattern is used. A dependent distribution was defined as an increase in attenuation of more dorsally located lung parenchymal abnormalities (i.e., GGO, crazy paving, reticulation, consolidation) or a clear predominance of such abnormalities in dependent lung areas with preservation of nondependent regions. Airspace bronchograms were defined as air-filled bronchi surrounded by GGO, crazy paving, or consolidation. Tree-in-bud sign was defined as nodular dilatation of centrilobular branching structures resembling a budding tree. The presence or the absence of associated centrilobular nodules, lymphadenopathy, or pleural effusion was likewise recorded. Furthermore, we analyzed all residual changes in lung texture such as traction bronchiectasis, persistent septal thickening, coarse reticulation, linear opacities, focal residual consolidation, micro- or macrocystic honeycombing, or distortion of lung parenchyma compatible with fibrosis. 2.7. Treatment of DAH Treatment of DAH included high-dose steroids in 9/16 patients (56%), tumor necrosis factor alpha blockage via etanercept in 5/16 patients (31%), and platelet transfusion in all patients with clinically evident hemoptysis or profound thrombocytopenia (10/16 patients; 63%). No specific treatment of DAH was applied in 3/16 patients (19%) with an inconspicuous clinical presentation. 2.8. Disease course, follow up, and outcome All 16 patients underwent routine nonenhanced chest CT before initiation of consolidation chemotherapy to exclude lower respiratory tract infection and preexisting fibrosis, plus as symptoms
682
D. Spira et al. / Clinical Imaging 37 (2013) 680–686
Fig. 1. CT morphologic patterns of DAH. Twenty-two-year-old male with common acute lymphocytic leukemia experiencing DAH 2 days after allo-SCT. HRCT demonstrates central GGO with peripheral sparing of lung parenchyma (A). Twenty-two-year-old female with acute myeloid leukemia (AML) developed DAH after conditioning chemotherapy. Sixteen days postchemotherapy HRCT shows parenchymal consolidation and mosaic GGO pattern in a dependent lobar distribution (B). Sixty-seven-year-old male with AML experienced DAH after conditioning chemotherapy. Ten days postchemotherapy HRCT reveals fine reticulation in a central distribution (C). Twenty-seven-year-old male with AML experiencing DAH 2 days after conditioning chemotherapy. HRCT demonstrates predominant crazy paving in a central and dependent distribution (D).
required for assessment of disease course and follow up. Survey time ranged from 0 to 384 days with a total of 74 CT scans (2–7 CT scans/ patient) performed. Lung parenchymal abnormalities were monitored via HRCT follow up and the time intervals until complete resolution of opacifications, fibrotic residues, or death were recorded, respectively. Values are expressed as median and range. 3. Results 3.1. Clinical and laboratory findings at onset of DAH All 16 patients received radio-/chemotherapy regimens at a median of 13 days (range: 3–44 days) prior to the development of DAH. Eight over 16 patients (50%) underwent allo-SCT at a median of 12 days (range: 0–38 days) before the development of DAH. Nine over 16 patients (56%) had profound thrombocytopenia (b 20.000 platelets/μl) at DAH onset, 2/16 patients (13%) had a peripheral blood platelet count of N 50.000 platelets/μl [10 and 21 days after stem cell transplantation (SCT)], and 5/16 patients (31%) had a peripheral blood platelet count between 20.000 and 50.000 platelets/μl. Initial clinical presentation included dyspnea in 14/16 patients (88%), elevated CRP in 7/16 patients (44%), fever in 4/16 patients (25%), hemoptysis in 3/16 patients (19%), cough in 2/16 patients (13%), and tachycardia with hypotension in 1/16 patients (6%). No signs of acute GVHD in the skin, the liver, or the gut were noted in any of the patients. 3.2. Lung parenchymal abnormalities at onset of DAH (Table 2) Lung parenchymal abnormalities in our series mainly included GGO (16/16 patients; 100%) and crazy paving (8/16 patients; 50%) and only rarely consolidation (5/16 patients; 31%) or fine reticulation (1/16 patients; 6%). When classified by their leading HRCT pattern, two main groups were recognized—those with preponderant crazy paving (6/16 patients; 38%) and those with predominantly isolated GGO (6/16 patients; 38%). Lung parenchymal abnormalities in
patients without allo-SCT did not differ from those with allo-SCT. Airspace bronchograms were seen in most patients (14/16 patients; 88%), and a dependent distribution was frequently discernible (7/16 patients; 44%) (Fig. 1B, D). Pulmonary involvement was bilateral (15/16 patients; 94%) with either a preference for the middle and lower lobes (9/16 patients; 56%), an upper lobe dominance (2/16 patients; 13%), or a random pattern regarding an upper/lower lobe dominance (5/16 patients; 31%). A primarily perihilar or diffuse distribution (also involving subpleural areas) was present in 6/16 patients (38%) and 9/16 patients (56%), respectively. Only 1/16 patients Table 2 Chest CT findings in subgroups defined by predominant HRCT pattern at onset of DAH Dominant CT finding
Morphology Air bronchograms Assoc. cons. Assoc. GGO Assoc. crazy paving gradient Distribution Bilateral Symmetric Middle and lower Upper Randomz-axis Central Peripheral Mosaic, patchy Randomtransverse plane Associated findings Tree-in-bud Centrilob. nodules Septal thickening Lymphadenopathy Pleural effusion
Crazy paving
Isolated GGO
Fine retic.
Consolidation
(n=6)
(n=6)
(n=1)
(n=3)
5/6 1/6 6/6 – 2/6
5/6 1/6 – 1/6 3/6
– – 1/1 – –
3/3 – 3/3 1/3 2/3
5/6 4/6 3/6 1/6 2/6 3/6 – 4/6 3/6
6/6 4/6 4/6 – 2/6 1/3 – 3/6 5/6
1/1 1/1 – – 1/1 1/1 – – –
3/3 1/3 2/3 1/3 – 1/3 1/3 – 1/3
1/6 3/6 1/6 – 3/6
4/6 3/6 3/6 3/6 3/6
– 1/1 – – –
1/3 1/3 1/3 – 2/3
Fine retic.=fine reticulation; assoc. cons.=associated consolidation.
D. Spira et al. / Clinical Imaging 37 (2013) 680–686 Table 3 Early (≤12 days) changes of chest CT findings in subgroups defined by predominant HRCT pattern at onset of DAH Dominant CT finding at onset of DAH Crazy paving Isolated GGO Fine retic. Consolidation
Increasing infiltrates Constant morphology Constant distribution Dominant GGO Dominant cons. Reticulation Traction bronchiectasis Decreasing infiltrates Constant morphology Constant distribution Dominant GGO Dominant cons. Reticulation Traction bronchiectasis No follow-up CT Clinical improvement Clinical deterioration
(n=6)
(n=6)
(n=1)
(n=3)
0/6 – – – – – – 4/6 – 3/4 3/4 1/4 1/4 1/4 2/6 1/2 1/2
2/6 2/2 2/2 2/2 – – 1/2 4/6 2/4 2/4 2/4 1/4 1/4 1/4 – – –
– – – – – – – 1/1 1/1 1/1 – – 1/1 – – – –
2/3 1/2 1/2 1/2 1/2 1/2 1/2 1/3 – 1/1 1/1 – – – – – –
Bolded figures highlight the respective subgroups. Fine retic.= fine reticulation; dominant cons.=dominant consolidation.
(6%) had peripherally located infiltrates with central sparing. No definite connection was observed between the occurrence of profound thrombocytopenia (b 20.000 platelets/μl; n= 9) and the leading initial HRCT pattern [3/9 (33%) GGO, 3/9 (33%) crazy paving, 1/9 (11%) fine reticulation, 2/9 (22%) consolidation]. Bilateral pleural effusions (8/16 patients; 50%), as well as centrilobular nodules (8/16 patients; 50%), constituted common associated findings. 3.3. Early (≤ 12 days) disease course and follow up (Table 3) In early follow up, a first subgroup of 10/16 patients (63%) experienced partial resolution of lung parenchymal opacifications: 4/ 6 with initial crazy paving, 4/6 with initial GGO, 1/1 with initial fine reticulation, and 1/3 with initial consolidation. In this group, initial crazy paving or consolidation showed a definite transformation towards GGO in 4/5 patients (80%), whereas in the fifth patient (20%), crazy paving turned into residual focal consolidation and coarse reticulation with traction bronchiectasis. Already initially dominant GGO either remained constant in morphology and distribution despite fading in intensity (2/4 patients; 50%) or changed into a coarse reticular and tractile appearance suggestive of secondary pulmonary fibrosis (2/4 patients; 50%) (Fig. 2). The second subgroup [4/16 patients (25%)] showed progressive lung parenchymal opacifications: 2/6 patients (33%) who presented with initial GGO and 2/3 patients (67%) with initial consolidation. Both patients with increas-
683
ing GGO showed no change in character of parenchymal opacities (Fig. 3), whereas 1 patient with central consolidation and reticulation at initial presentation demonstrated a conversion towards a diffuse, expansive GGO followed by increased coarse reticulation, traction bronchiectasis, and death (Fig. 4). Mechanical ventilation was necessary in 2/6 patients with initially dominant crazy paving (33%), in 3/6 patients with initially dominant isolated GGO (50%), and in 2/3 patients with initially dominant consolidation (67%). 3.4. Patient's outcome (Table 4) Four over 16 patients (25%) died from progressive DAH at a median of 21 days (range: 0–24 days) after the beginning of symptoms (Fig. 4). All 4 patients needed mechanical ventilation at final disease stage. DAH clinically resolved in 12/16 patients (75%). In the CT morphologic disease course, 6 of these 12 patients (50%) showed complete resolution of lung parenchymal opacifications at a median of 78 days (range: 12–231 days). In another, 3 of those 12 patients (75%) residues with a fibrotic appearance (i.e., parenchymal bands, traction bronchiectasis, and reticulation) were observed at a median of 56 days (range: 42–67 days) (Fig. 2)—with equal distribution of dominant CT findings at onset of DAH. Two over 12 patients (17%) exhibited only mild residual changes on early follow-up CT at 6 days and 29 days after DAH onset (Fig. 3). One over 12 patients (8%) did not receive CT follow up. 4. Discussion Bleeding to the lung may originate from bronchial vessels, pulmonary vessels, or the alveolar microcirculation [15]. Diffuse pulmonary hemorrhage is caused by injury to the alveolar microcirculation and can occur with a variety of diseases such as different vasculitides, connective tissue disorders, infections, barotraumas, coagulation disorders, drug toxicities, or after hematopoietic SCT [12,15]. Depending on the cause and the clinical setting, DAH may exhibit a wide range of morbidity and mortality. Reported death rates between 45% and 100% for DAH are encountered mainly in patients who developed this complication after allo-SCT and cytotoxic chemotherapy [5,8]. These particular settings certainly are responsible for different pathomechanisms compared to simple pulmonary hemorrhage due to temporary thrombocytopenia, barotrauma, infection, or autoimmune diseases. Radiographic studies have long been and still are an integral part of the diagnostic work-up in patients with DAH [9]. However, due to its superiority for illustrating even subtle macroscopic patterns of lung parenchymal involvement, HRCT represents the mainstay of morphologic lung imaging in patients with hematological malignancies at risk for both, infectious and noninfectious pulmonary complications. Hence, HRCT is often obtained before definite confirmation of DAH by BAL. Yet, the field of CT morphologic
Fig. 2. CT morphologic disease course of DAH. Twenty-two-year-old male with common ALL experiencing DAH after allo-SCT. Initial findings at Day 17 posttransplant mainly consist of ill-defined centrilobular GGO with a bilateral and partly confluent, patchy distribution (A). Twenty-eight days later, following mechanical ventilation, high-dose steroids, and etanercept, opacifications evolve into centrally located coarse peribronchovascular consolidations accompanied by incipient lung parenchymal retraction and widening of bronchial lumina in the right upper lobe (B). Note concomitant pneumomediastinum, presumably due to mechanical ventilation. Another 28 days later, fibrotic changes partially resolve, leaving residual GGO and parenchymal distortion (C). The pneumomediastinum completely subsides.
684
D. Spira et al. / Clinical Imaging 37 (2013) 680–686
Fig. 3. CT morphologic disease course of DAH. Fifty-three-year-old male with mycosis fungoides experiencing DAH after allo-SCT. At Day 10 posttransplant, HRCT shows lobulardependent GGO with characteristic sparing of the ventral lung regions (A). Ten days later, the pattern is extended in distribution with increasing parenchymal attenuation but unchanged in character (B). The dependent distribution of lung parenchymal changes is still preserved. Another 19 days later, opacifications resolve, and only minimal residual abnormalities (i.e., few parenchymal bands) are noted (C). The patient survives, and no mechanical ventilation during the disease course was needed.
DAH imaging in patients with hematological malignancies is still neglected in the literature, and aside from rather scarce reports, we did not find a comprehensive analysis of HRCT findings in patients with hematological malignancies developing DAH [10–12]. The onset of DAH is usually within the first 30 days (median: 11– 24 days) following SCT [16], which was comparable to our patient group (median: 12 days). Symptoms typically include dyspnea, fever, and cough, whereas hemoptysis is rare [5]. In addition, we observed a significant proportion of patients with an initially elevated CRP (44%), which is explained by the inflammatory processes accompanying neutrophile influx through the endothelial wall into the lung parenchyma, chemotherapy-induced mucositis, and others. Even though severe, acute GVHD is recognized as a risk factor for DAH after SCT, no signs of acute GVHD in the skin, the liver, or the gut were noted in any of our patients [5,17]. Other defined risk factors are advanced age, total body irradiation, myeloablative conditioning regimen, and allogeneic donor source [5,17]. Some review articles and textbooks describe bilateral areas of GGO or consolidation as the most common CT findings in DAH [11,18,19], whereas others noted GGO with a superimposed reticular pattern as characteristical [12,20]. This is well represented in our series where two main groups were recognized at onset of DAH—those with predominant crazy paving and those with predominantly isolated GGO. However, both patterns reflect the same pathologic background of an acute alveolar/interstitial disease and should rather be seen as a continuum than separate entities, therefore constituting the main dominant finding in 76% of our patient group. At initial presentation, consolidation was seen in 31% of our patients, constituting the domi-
nant pattern in only 19%. Patterns of distribution and associated findings were similar in both subgroups and comparable to the radiographic analysis conducted by Witte et al. [9]. The preference of DAH for the middle/lower lung areas (probably due to gravitational forces on intraalveolar hemorrhage) and its bilateral distribution need to be regarded as nonspecific, even though infectious complications such as Pneumocystis jirovecii and cytomegalovirus show a preference for the upper lung areas or are randomly distributed. In our study, airspace bronchograms were seen in most patients, and a characteristic dependent distribution was frequently discernible (44%). It is tempting to explain both features with DAH pathophysiology and pulmonary microanatomy: bleeding in DAH originates from the alveolar capillaries [15], therefore primarily involves the alveolar spaces and alveolar ducts with sparing of the bronchial tree [21,22] and the consecutive appearance of airspace bronchograms. Furthermore, communications of alveoli through pores of Kohn and alveolar fenestrae as well as direct communications between alveoli and peripheral bronchioles (canals of Lambert) enable collateral ventilation and/or expansion of infiltrates, thus explaining the observed dependent distribution [23–25]. The appearance of a tree-in-bud sign is strongly suggestive for an infectious process. At times, however (in our study 6/16 DAH patients), it is also seen in patients with pulmonary hemorrhage, probably due to hemoptysis and consecutive microaspiration of blood into bronchioli [26]. The radiologist should be aware of the possibility of tree-in-bud occurring in DAH. Nevertheless, a high level of suspicion for infection needs to be maintained when the tree-in-bud sign is spotted.
Fig. 4. CT morphologic disease course of DAH. Forty-one-year-old male with T-cell non-Hodgkin lymphoma developed DAH after allo-SCT. Initial DAH opacifications at Day 14 posttransplant demonstrate central consolidation and reticulation with airspace bronchograms in the upper lobes and sparing of the lung periphery (A). Fourteen days later, the pattern is changed into a mixture of GGO, coarse reticulation (ventrally located), and fine reticulation (dorsally located). Note traction bronchiectasis in the middle lobe (B). Another 17 days later, the lung is strongly increased in attenuation due to a mixture of consolidation, diffuse GGO and reticulation, as well as persistent bronchiectasis (C). Also note extensive pneumomediastinum and thoracic wall emphysema, presumably due to mechanical ventilation. Despite mechanical ventilation, treatment with high-dose steroids, and etanercept, the patient dies the same day.
D. Spira et al. / Clinical Imaging 37 (2013) 680–686 Table 4 Outcome of subgroups defined by predominant HRCT pattern at onset of DAH Dominant CT finding at onset of DAH Crazy paving Isolated GGO Fine retic. Consolidation
Clinically resolved CT residua (at N40 d) Dominant GGO Retic./Parench. bands Traction bronchiectasis Death due to DAH Constant morphology Constant distribution Dominant GGO Dominant cons. Retic./Parench. bands Traction bronchiectasis
(n=6)
(n=6)
(n=1)
(n=3)
5/6 – – – – 1/6 1/1 1/1 – – – –
5/6 1/5 1/5 1/5 1/5 1/6 – 1/1 1/1 – 1/1 1/1
1/1 1/1 – 1/1 – – – – – – – –
1/3 1/1 1/1 – – 2/3 1/2 1/2 – 2/2 1/2 1/2
Bolded figures highlight the respective subgroups. Fine retic.=fine reticulation; retic./parench. bands=reticulation and parenchymal bands; dominant cons.=dominant consolidation.
Radiographically, the course of DAH is described as rapidly progressive from an initially mild interstitial or alveolar pattern in the central and lower lung towards a diffuse, severe alveolar pattern [9]. However, only scarce data exist on the CT morphologic course of DAH infiltrates. In our series, the disease course was unpredictable and independent of primary DAH CT morphology spanning a wide range of CT morphologic conversion varying from complete resolution (38%, median of 78 days) over fibrotic residues (37%, median of 56 days) to a diffusely fibrotic appearance, DAH progression, and death (25%, median: 21 days). This is at least in part explained by the various underlying mechanisms and response characteristics of the host, different combinatory treatment regimens applied, and variable extent of thrombocytopenia. Hence, an either intermittent or definite fibrotic conversion is suggested by their CT morphologic course in several of our patients. Progression of parenchymal consolidations at follow up was more often followed by worsening of respiratory function and death. We did not find any data concerning the later histopathology in patients surviving DAH after allo-SCT. However, DAH is intimately associated with the histologic picture of DAD [11,27,28], which is characterized by a uniform interstitial fibrosis in the organizing phase followed by either gradual resolution or continued interstitial fibrosis with architectural remodeling and progressive respiratory compromise [22]. When compared to the reported death rate of 45%–100% [5,8], the initial mortality due to progressive DAH was very low in our series (25%). The reduced mortality rate may primarily be related to the heterogeneity of pathomechanisms in the chemotherapy and SCT setting and variable risk profiles. However, a concise evaluation in this respect was beyond the scope of this study. Specialized patient care at the intensive care unit using modern treatment regimens such as early high-dose corticosteroids±cytokine inhibitors, including tumor necrosis factor alpha blockers, besides supportive measures (noninvasive ventilation, etc.) may also have contributed to improved patient outcome. Due to often similar CT morphology and distribution of opacifications, P. jirovecii pneumonia and viral lung infections (such as cytomegalovirus, herpes simplex virus, respiratory syncytial virus, and influenza virus) need to be excluded clinically and via BAL. Yet, a basal dominance and a dependent distribution are rarely seen in these infections and should rather suggest DAH in the appropriate clinical setting. The main noninfectious differential diagnoses to be considered are pulmonary edema, IPS, engraftment syndrome, granulocyte colony-stimulating factor (G-CSF)-induced lung changes, transplantassociated thrombotic microangiopathy (TA-TMA) and, eventually, transfusion-associated lung injury (TRALI). Pulmonary edema due to
685
fluid overload, renal or hepatic insufficiency, or cardiotoxicity rapidly responds to diuretic treatment and was excluded in this analysis. IPS is often similar in appearance but shows a variable yet nondependent distribution, holding a tendency towards a more rapid fibrotic conversion at follow up, frequently necessitating early mechanical ventilation. TA-TMA is diagnosed by clinical and laboratory data such as red blood cell fragmentation (≥ 2 schistocytes per high-power field), concurrent increased serum lactate dehydrogenase, concurrent renal and neurologic dysfunction, as well as negative direct and indirect Coombs test. TRALI is characterized by its intimate timely relation to stem cell transfusion, whereas DAH is a known complication of G-CSF application. Our study holds a number of limitations. Firstly, the dictate of the clinical course on the number and time windows between CT follow up complicates adequate comparison. Secondly, a large number of other variables related to the underlying disorders, such as type and intensity of therapies, age, and associated comorbidities, may have a significant impact on the evolution and outcome of DAH. Third, therapeutic management of DAH varied in our cohort, being determined by the degree of respiratory distress and other clinical parameters. The retrospective character and the rather small number of patients request larger prospective follow-up studies in order to validate our results. Nonetheless, to our knowledge, this is the first HRCT study to analyze the development, patterns of distribution, and course of this serious complication in patients with hematological malignancies. The intention of this work was to exhibit the radiologist's role in identifying characteristic changes and monitoring disease course of DAH, thereby endowing the treating hematologist with additional information and aiding in treatment decisions. 5. Conclusion Chest CT features of DAH are each by oneself nonspecific; however, the conjoint occurrence of bilateral, diffuse, and dependent GGO± crazy paving, accompanied by airspace bronchograms in the appropriate clinical setting, is consistent with this complication. Nevertheless, a broad differential diagnosis needs to be considered, including infectious agents (e.g., viruses, P. jirovecii) and noninfectious, treatment-related causes (e.g., pulmonary edema, IPS, TA-TMA, GCSF, and TRALI). The disease course of DAH is unforeseeable, spanning a wide range of CT morphologic conversion varying from complete resolution to increasing fibrosis, DAH progression and death, and can be well monitored via HRCT. References [1] Camus P, Fanton A, Bonniaud P, Camus C, Foucher P. Interstitial lung disease induced by drugs and radiation. Respiration 2004;71(4):301-26. [2] Limper AH. Chemotherapy-induced lung disease. Clin Chest Med 2004;25(1): 53-64. [3] Abratt RP, Morgan GW, Silvestri G, Willcox P. Pulmonary complications of radiation therapy. Clin Chest Med 2004;25(1):167-77. [4] Goldman AL, Enquist R. Hyperacute radiation pneumonitis. Chest 1975;67(5): 613-5. [5] Afessa B, Tefferi A, Litzow MR, Krowka MJ, Wylam ME, Peters SG. Diffuse alveolar hemorrhage in hematopoietic stem cell transplant recipients. Am J Respir Crit Care Med 2002;166(5):641-5. [6] Weisdorf DJ. Diffuse alveolar hemorrhage: an evolving problem? Leukemia 2003;17(6):1049-50. [7] Carreras E, Diaz-Ricart M. The role of the endothelium in the short-term complications of hematopoietic SCT. Bone Marrow Transplant 2011;46(12): 1495-502. [8] Gupta S, Jain A, Warneke CL, Gupta A, Shannon VR, Morice RC, Onn A, Jimenez CA, Bashoura L, Giralt SA, Dickey BF, Eapen GA. Outcome of alveolar hemorrhage in hematopoietic stem cell transplant recipients. Bone Marrow Transplant 2007;40(1):71-8. [9] Witte RJ, Gurney JW, Robbins RA, Linder J, Rennard SI, Arneson M, Vaughan WP, Reed EC, Dicke KA. Diffuse pulmonary alveolar hemorrhage after bone marrow transplantation: radiographic findings in 39 patients. AJR Am J Roentgenol 1991; 157(3):461-4.
686
D. Spira et al. / Clinical Imaging 37 (2013) 680–686
[10] Primack SL, Miller RR, Müller NL. Diffuse pulmonary hemorrhage: clinical, pathologic, and imaging features. AJR Am J Roentgenol 1995;164(2):295-300. [11] Frazer RS, Müller NL, Colman N, Paré PD, editors. Diagnosis of diseases of the chest. 4th ed. Philadelphia: W.B. Saunders; 1999. p. 1726-7. [12] Webb WR, Müller NL, Naidich DP, editors. High-resolution CT of the lung. 4th ed. Philadelphia, MD: Lippincott Williams & Wilkins; 2009. p. 462-7. [13] Parr DG, Stoel BC, Stolk J, Nightingale PG, Stockley RA. Influence of calibration on densitometric studies of emphysema progression using computed tomography. Am J Respir Crit Care Med 2004;170(8):883-90. [14] Austin JH, Müller NL, Friedman PJ, Hansell DM, Naidich DP, Remy-Jardin M, Webb WR, Zerhouni EA. Glossary of terms for CT of the lungs: recommendations of the Nomenclature Committee of the Fleischner Society. Radiology 1996;200(2): 327-31. [15] Lara AR, Schwarz MI. Diffuse alveolar hemorrhage. Chest 2010;137(5):1164-71. [16] Afessa B, Peters SG. Major complications following hematopoietic stem cell transplantation. Semin Respir Crit Care Med 2006;27(3):297-309. [17] Majhail NS, Parks K, Defor TE, Weisdorf DJ. Diffuse alveolar hemorrhage and infection-associated alveolar hemorrhage following hematopoietic stem cell transplantation: related and high-risk clinical syndromes. Biol Blood Marrow Transplant 2006;12(10):1038-46. [18] Afessa B, Peters SG. Noninfectious pneumonitis after blood and marrow transplant. Curr Opin Oncol 2008;20(2):227-33. [19] Worthy SA, Flint JD, Müller NL. Pulmonary complications after bone marrow transplantation: high-resolution CT and pathologic findings. Radiographics 1997; 17(6):1359-71.
[20] Coy DL, Ormazabal A, Godwin JD, Lalani T. Imaging evaluation of pulmonary and abdominal complications following hematopoietic stem cell transplantation. Radiographics 2005;25(2):305-17. [21] Roychowdhury M, Pambuccian SE, Aslan DL, Jessurun J, Rose AG, Manivel JC, Gulbahce HE. Pulmonary complications after bone marrow transplantation: an autopsy study from a large transplantation center. Arch Pathol Lab Med 2005;129(3):366-71. [22] Beasley MB. The pathologist's approach to acute lung injury. Arch Pathol Lab Med 2010;134(5):719-27. [23] Takaro T, Price HP, Parra SC. Ultrastructural studies of apertures in the interalveolar septum of the adult human lung. Am Rev Respir Dis 1979;119(3):425-34. [24] Pump KK. Fenestrae in the alveolar membrane of the human lung. Chest 1974;65(4):431-6. [25] Lambert MW. Accessory bronchiolealveolar communications. J Pathol Bacteriol 1955;70(2):311-4. [26] Okada F, Ando Y, Yoshitake S, Ono A, Tanoue S, Matsumoto S, Wakisaka M, Maeda T, Mori H. Clinical/pathologic correlations in 553 patients with primary centrilobular findings on high-resolution CT scan of the thorax. Chest 2007;132(6):1939-48. [27] Robbins RA, Linder J, Stahl MG, Thompson AB, Haire W, Kessinger A, Armitage JO, Arneson M, Woods G, Vaughan WP, et al. Diffuse alveolar hemorrhage in autologous bone marrow transplant recipients. Am J Med 1989;87(5):511-8. [28] Agustí C, Ramirez J, Picado C, Xaubet A, Carreras E, Ballester E, Torres A, Battochia C, Rodriguez-Roisin R. Diffuse alveolar hemorrhage in allogeneic bone marrow transplantation. A postmortem study. Am J Respir Crit Care Med 1995;151(4):1006-10.
Contents lists available at SciVerse ScienceDirect
Clinical Imaging journal homepage: http://www.clinicalimaging.org
Diffuse alveolar hemorrhage in patients with hematological malignancies: HRCT patterns of pulmonary involvement and disease course Daniel Spira a,⁎, Stefan Wirths b, Felix Skowronski a, Jan Pintoffl b, Sascha Kaufmann a, Harald Brodoefel c, Marius Horger a a b c
Department of Diagnostic and Interventional Radiology, Eberhard-Karls-University, 72076 Tübingen, Germany Department of Oncology and Hematology, Eberhard-Karls-University, 72076 Tübingen, Germany Department of Radiology, Harvard Medical School, Beth Israel Deaconess Medical Center, WCC 308A, Boston, MA 02215, USA
a r t i c l e
i n f o
Article history: Received 27 June 2012 Received in revised form 25 August 2012 Accepted 7 November 2012 Keywords: High-resolution CT Drugs Reactions Hematological malignancies Diffuse alveolar hemorrhage
a b s t r a c t Objective: To analyze high-resolution computed tomography (HRCT) patterns of lung involvement and disease course in patients with hematological malignancies experiencing diffuse alveolar hemorrhage (DAH) after chemotherapy ±allogeneic stem cell transplantation (allo-SCT). Materials and methods: Sixteen patients experiencing DAH after chemotherapy±allo-SCT were enrolled. A total of 74 computed tomography (CT) scans obtained before, during, and after onset of DAH were evaluated retrospectively. Results: CT features of DAH are each, by oneself, nonspecific. However, conjoint bilateral, diffuse, and dependent ground glass opacity±crazy paving, accompanied by airspace bronchograms, should suggest this complication. The HRCT course comprises a wide range of trends that are not predictive for patient's outcome, but progression of parenchymal consolidations at follow up was more often detrimental. © 2013 Elsevier Inc. All rights reserved.
1. Introduction An abundance of noninfectious pulmonary complications is induced by chemotherapy and hematopoietic (particularly allogeneic) allogeneic stem cell transplantation (allo-SCT), commonly including diffuse alveolar damage (DAD), pulmonary edema, idiopathic pneumonia syndrome (IPS), diffuse alveolar hemorrhage (DAH), and manifestations related to acute pulmonary graft-versus-host disease (GVHD) [1–4]. DAH in the allo-SCT setting holds a poor prognosis with a mortality rate ranging between 64% and 100% [5]. It is characterized by multilobar pulmonary infiltrates, hypoxemia, absence of infection compatible with the diagnosis, progressively bloodier return in bronchoalveolar lavage (BAL), and the presence of blood in ≥ 30% of alveolar surfaces at autopsy [5,6]. The exact pathogenesis still remains elusive although capillary endothelial damage by chemotherapy followed by intravascular microthrombosis and transendothelial neutrophil engraftment following alloSCT is known to cause leakage of red blood cells into the pulmonary alveoli [7]. Clinical symptoms and signs are those of pneumonia, that is, dyspnea, cough, fever, tachycardia, and rarely hemoptysis [8].
⁎ Corresponding author. Department of Diagnostic and Interventional Radiology, Eberhard-Karls-University, Hoppe-Seyler-Str. 3, 72076 Tübingen, Germany. Tel.: +497071-2987212; fax: +49-7071-295845. E-mail address: [email protected] (D. Spira). 0899-7071/$ – see front matter © 2013 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.clinimag.2012.11.005
Accurate diagnosis and close monitoring of DAH via imaging is imperative for patient care. The radiographic appearance of DAH is described as unspecific with an interstitial or alveolar pattern of lung opacification primarily involving the central portion of the lung with predilection for the middle and lower lung zones [9]. High-resolution computed tomography (HRCT) is the most accurate imaging method displaying parenchymal lung involvement and is ideally suited to recognize and follow the extent and macroscopic morphology of DAH. However, apart from several excellent review articles and book chapters overviewing the field of DAH [10–12], we did not find a comprehensive analysis of HRCT findings in patients with hematological malignancies developing DAH. Thus, we analyzed lung involvement and disease course by HRCT in patients with hematological malignancies who developed DAH after chemotherapy±allo-SCT. Our work addresses two principle points: to determine the leading computed tomography (CT) morphologic finding or combination of findings and their prevalence in this patient group and to trace the temporal course of lung parenchymal changes. 2. Methods 2.1. Patient demographics This retrospective study was approved by our institutional review board that waived informed consent. Between May 2008 and
D. Spira et al. / Clinical Imaging 37 (2013) 680–686
December 2011, 16 patients with hematological malignancies (11 men, 5 women; age range: 21–70 years, mean: 46.4 years) who developed DAH after chemotherapy±allo-SCT were enrolled (Table 1). Eight patients underwent allo-SCT before the development of DAH. All patients were examined for signs of acute GVHD in the skin, the liver, or the gut after allo-SCT. 2.2. Inclusion criteria DAH was defined (according to current guidelines of National Institutes of Health) as the sudden development of multilobar infiltrates, symptoms and signs of pneumonia with abnormal pulmonary physiology, the absence of active lower respiratory tract infection, and the presence of hemorrhage as determined by BAL or retrospectively by autopsy [5]. In our study, the following additional criteria were considered necessary for confirmation of DAH and subsequent patient enrollment: (a) BAL at ≤ 2 days after the beginning of symptoms showing progressively bloodier return (n= 10); (b) autopsy revealing blood in N30% of the alveolar surfaces of lung tissue (n= 1); or (c) the combination of thrombocytopenia (b50.000/μl), a sudden drop in hemoglobin of ≥ 1g/dl, multilobar infiltrates, and an acute drop in oxygen saturation accompanied by clinically evident hemoptysis (n= 3) or generalized bleeding diathesis (i.e., massive epistaxis, cutaneous, and retinal bleeding) (n= 2) were considered diagnostic for DAH. 2.3. Exclusion criteria Patients with evidence of viral, bacterial or fungal infection, vasculitis, connective tissue disorders, trauma, as well as those with evident focal bleeding to the lung from bronchial or pulmonary vessels were excluded from the analysis. 2.4. CT evaluation All CT examinations were performed on either a 64- or a 128-row multidetector CT (Sensation 64/Definition+; Siemens, Erlangen, Germany) using a 250–330-mm field of view, a 512×512 reconstruction matrix, 120 kV, 100–120 effective mAs, and a tube rotation time of 0.5/0.3 ms. No intravenous contrast material was applied. A single spiral acquisition was obtained from the apex to the base during one breath-hold at suspended end-inspiratory volume. Examinations were performed with patients in the supine position. Reconstruction of the data was done at 1-mm slices with a sharp reconstruction algorithm (filter, B70, equivalent to HRCT) and 1-mm
Table 1 Treatment regimens Patient Disease
Treatment regimen
1
Multiple myeloma
2 3 4 5 6 7
AML Mycosis fungoides AML ALL AML Mantle cell lymphoma AML AML AML T-NHL AML AML AML ALL AML
Total body irradiation, busulfan, cyclophosphamide Clofabrine, thiotepa, melphalan Fludarabine, melphalan, MabCampath Total body irradiation, fludarabine Clofabrine, thiotepa, melphalan Cyclophosphamide, anti-thymocyte globulin Fludarabine, melphalan
8 9 10 11 12 13 14 15 16
Cytarabin, idarubicin Clofabrine, thiotepa, melphalan Total body irradiation, cyclophosphamide Total body irradiation, etoposide Fludarabine, cytarabine, amsacrine Cytarabine Clofabrine, THIOTEPA, MELPHALAN Treosulfane, etoposide, cyclophosphamide Total body irradiation, cyclophosphamide
681
reconstruction increment for visual assessment. Systems were calibrated on air daily [13]. 2.5. Clinical and laboratory findings The time interval from radio-/chemotherapy as well as allo-SCT to the development of DAH was recorded. At onset of DAH, peripheral blood platelet count, C-reactive protein (CRP), clinical presentation, and the presence/absence of signs of acute GVHD were assessed in all patients. 2.6. Chest CT evaluation of lung parenchymal abnormalities Images were viewed at lung (window width 1200 HU, window level 600 HU) and mediastinal (window width 350 HU, window level 50 HU) window settings. CT images were reevaluated retrospectively by two radiologists with 17 and 4 years (MH, DS) experience in reading chest CT via consensus reading. Both investigators were blinded to the original interpretations and mortality data. Criteria for HRCT findings were those defined by the Fleischner Society's glossary of terms for chest CT [14]. Ground glass opacity (GGO) was defined as a hazy increase in lung attenuation with preservation of bronchial and vascular margins being distributed either diffuse or patchy (Fig. 1A). Airspace consolidation was defined as an area of dense increase in attenuation with obscuration of the underlying vessels and airway walls, showing different morphology either in form of segmental or subsegmental or patchy opacifications (Fig. 1B). Reticulation was defined as an interlacing line shadow suggesting a mesh or net (Fig. 1C). Crazy paving was defined as a superimposition of GGO and reticulation (Fig. 1D). GGO, airspace consolidation, reticulation, or crazy paving, distributed throughout the parenchyma without zonal predominance, was called diffuse, while focal parenchymal infiltrates with lobular, segmental, or lobar distribution involving one or both lungs were named focal. If more or less sharply demarcated regions of different density (to the point of unattained secondary lobules) were noted within infiltrated lung parenchyma, the term mosaic pattern is used. A dependent distribution was defined as an increase in attenuation of more dorsally located lung parenchymal abnormalities (i.e., GGO, crazy paving, reticulation, consolidation) or a clear predominance of such abnormalities in dependent lung areas with preservation of nondependent regions. Airspace bronchograms were defined as air-filled bronchi surrounded by GGO, crazy paving, or consolidation. Tree-in-bud sign was defined as nodular dilatation of centrilobular branching structures resembling a budding tree. The presence or the absence of associated centrilobular nodules, lymphadenopathy, or pleural effusion was likewise recorded. Furthermore, we analyzed all residual changes in lung texture such as traction bronchiectasis, persistent septal thickening, coarse reticulation, linear opacities, focal residual consolidation, micro- or macrocystic honeycombing, or distortion of lung parenchyma compatible with fibrosis. 2.7. Treatment of DAH Treatment of DAH included high-dose steroids in 9/16 patients (56%), tumor necrosis factor alpha blockage via etanercept in 5/16 patients (31%), and platelet transfusion in all patients with clinically evident hemoptysis or profound thrombocytopenia (10/16 patients; 63%). No specific treatment of DAH was applied in 3/16 patients (19%) with an inconspicuous clinical presentation. 2.8. Disease course, follow up, and outcome All 16 patients underwent routine nonenhanced chest CT before initiation of consolidation chemotherapy to exclude lower respiratory tract infection and preexisting fibrosis, plus as symptoms
682
D. Spira et al. / Clinical Imaging 37 (2013) 680–686
Fig. 1. CT morphologic patterns of DAH. Twenty-two-year-old male with common acute lymphocytic leukemia experiencing DAH 2 days after allo-SCT. HRCT demonstrates central GGO with peripheral sparing of lung parenchyma (A). Twenty-two-year-old female with acute myeloid leukemia (AML) developed DAH after conditioning chemotherapy. Sixteen days postchemotherapy HRCT shows parenchymal consolidation and mosaic GGO pattern in a dependent lobar distribution (B). Sixty-seven-year-old male with AML experienced DAH after conditioning chemotherapy. Ten days postchemotherapy HRCT reveals fine reticulation in a central distribution (C). Twenty-seven-year-old male with AML experiencing DAH 2 days after conditioning chemotherapy. HRCT demonstrates predominant crazy paving in a central and dependent distribution (D).
required for assessment of disease course and follow up. Survey time ranged from 0 to 384 days with a total of 74 CT scans (2–7 CT scans/ patient) performed. Lung parenchymal abnormalities were monitored via HRCT follow up and the time intervals until complete resolution of opacifications, fibrotic residues, or death were recorded, respectively. Values are expressed as median and range. 3. Results 3.1. Clinical and laboratory findings at onset of DAH All 16 patients received radio-/chemotherapy regimens at a median of 13 days (range: 3–44 days) prior to the development of DAH. Eight over 16 patients (50%) underwent allo-SCT at a median of 12 days (range: 0–38 days) before the development of DAH. Nine over 16 patients (56%) had profound thrombocytopenia (b 20.000 platelets/μl) at DAH onset, 2/16 patients (13%) had a peripheral blood platelet count of N 50.000 platelets/μl [10 and 21 days after stem cell transplantation (SCT)], and 5/16 patients (31%) had a peripheral blood platelet count between 20.000 and 50.000 platelets/μl. Initial clinical presentation included dyspnea in 14/16 patients (88%), elevated CRP in 7/16 patients (44%), fever in 4/16 patients (25%), hemoptysis in 3/16 patients (19%), cough in 2/16 patients (13%), and tachycardia with hypotension in 1/16 patients (6%). No signs of acute GVHD in the skin, the liver, or the gut were noted in any of the patients. 3.2. Lung parenchymal abnormalities at onset of DAH (Table 2) Lung parenchymal abnormalities in our series mainly included GGO (16/16 patients; 100%) and crazy paving (8/16 patients; 50%) and only rarely consolidation (5/16 patients; 31%) or fine reticulation (1/16 patients; 6%). When classified by their leading HRCT pattern, two main groups were recognized—those with preponderant crazy paving (6/16 patients; 38%) and those with predominantly isolated GGO (6/16 patients; 38%). Lung parenchymal abnormalities in
patients without allo-SCT did not differ from those with allo-SCT. Airspace bronchograms were seen in most patients (14/16 patients; 88%), and a dependent distribution was frequently discernible (7/16 patients; 44%) (Fig. 1B, D). Pulmonary involvement was bilateral (15/16 patients; 94%) with either a preference for the middle and lower lobes (9/16 patients; 56%), an upper lobe dominance (2/16 patients; 13%), or a random pattern regarding an upper/lower lobe dominance (5/16 patients; 31%). A primarily perihilar or diffuse distribution (also involving subpleural areas) was present in 6/16 patients (38%) and 9/16 patients (56%), respectively. Only 1/16 patients Table 2 Chest CT findings in subgroups defined by predominant HRCT pattern at onset of DAH Dominant CT finding
Morphology Air bronchograms Assoc. cons. Assoc. GGO Assoc. crazy paving gradient Distribution Bilateral Symmetric Middle and lower Upper Randomz-axis Central Peripheral Mosaic, patchy Randomtransverse plane Associated findings Tree-in-bud Centrilob. nodules Septal thickening Lymphadenopathy Pleural effusion
Crazy paving
Isolated GGO
Fine retic.
Consolidation
(n=6)
(n=6)
(n=1)
(n=3)
5/6 1/6 6/6 – 2/6
5/6 1/6 – 1/6 3/6
– – 1/1 – –
3/3 – 3/3 1/3 2/3
5/6 4/6 3/6 1/6 2/6 3/6 – 4/6 3/6
6/6 4/6 4/6 – 2/6 1/3 – 3/6 5/6
1/1 1/1 – – 1/1 1/1 – – –
3/3 1/3 2/3 1/3 – 1/3 1/3 – 1/3
1/6 3/6 1/6 – 3/6
4/6 3/6 3/6 3/6 3/6
– 1/1 – – –
1/3 1/3 1/3 – 2/3
Fine retic.=fine reticulation; assoc. cons.=associated consolidation.
D. Spira et al. / Clinical Imaging 37 (2013) 680–686 Table 3 Early (≤12 days) changes of chest CT findings in subgroups defined by predominant HRCT pattern at onset of DAH Dominant CT finding at onset of DAH Crazy paving Isolated GGO Fine retic. Consolidation
Increasing infiltrates Constant morphology Constant distribution Dominant GGO Dominant cons. Reticulation Traction bronchiectasis Decreasing infiltrates Constant morphology Constant distribution Dominant GGO Dominant cons. Reticulation Traction bronchiectasis No follow-up CT Clinical improvement Clinical deterioration
(n=6)
(n=6)
(n=1)
(n=3)
0/6 – – – – – – 4/6 – 3/4 3/4 1/4 1/4 1/4 2/6 1/2 1/2
2/6 2/2 2/2 2/2 – – 1/2 4/6 2/4 2/4 2/4 1/4 1/4 1/4 – – –
– – – – – – – 1/1 1/1 1/1 – – 1/1 – – – –
2/3 1/2 1/2 1/2 1/2 1/2 1/2 1/3 – 1/1 1/1 – – – – – –
Bolded figures highlight the respective subgroups. Fine retic.= fine reticulation; dominant cons.=dominant consolidation.
(6%) had peripherally located infiltrates with central sparing. No definite connection was observed between the occurrence of profound thrombocytopenia (b 20.000 platelets/μl; n= 9) and the leading initial HRCT pattern [3/9 (33%) GGO, 3/9 (33%) crazy paving, 1/9 (11%) fine reticulation, 2/9 (22%) consolidation]. Bilateral pleural effusions (8/16 patients; 50%), as well as centrilobular nodules (8/16 patients; 50%), constituted common associated findings. 3.3. Early (≤ 12 days) disease course and follow up (Table 3) In early follow up, a first subgroup of 10/16 patients (63%) experienced partial resolution of lung parenchymal opacifications: 4/ 6 with initial crazy paving, 4/6 with initial GGO, 1/1 with initial fine reticulation, and 1/3 with initial consolidation. In this group, initial crazy paving or consolidation showed a definite transformation towards GGO in 4/5 patients (80%), whereas in the fifth patient (20%), crazy paving turned into residual focal consolidation and coarse reticulation with traction bronchiectasis. Already initially dominant GGO either remained constant in morphology and distribution despite fading in intensity (2/4 patients; 50%) or changed into a coarse reticular and tractile appearance suggestive of secondary pulmonary fibrosis (2/4 patients; 50%) (Fig. 2). The second subgroup [4/16 patients (25%)] showed progressive lung parenchymal opacifications: 2/6 patients (33%) who presented with initial GGO and 2/3 patients (67%) with initial consolidation. Both patients with increas-
683
ing GGO showed no change in character of parenchymal opacities (Fig. 3), whereas 1 patient with central consolidation and reticulation at initial presentation demonstrated a conversion towards a diffuse, expansive GGO followed by increased coarse reticulation, traction bronchiectasis, and death (Fig. 4). Mechanical ventilation was necessary in 2/6 patients with initially dominant crazy paving (33%), in 3/6 patients with initially dominant isolated GGO (50%), and in 2/3 patients with initially dominant consolidation (67%). 3.4. Patient's outcome (Table 4) Four over 16 patients (25%) died from progressive DAH at a median of 21 days (range: 0–24 days) after the beginning of symptoms (Fig. 4). All 4 patients needed mechanical ventilation at final disease stage. DAH clinically resolved in 12/16 patients (75%). In the CT morphologic disease course, 6 of these 12 patients (50%) showed complete resolution of lung parenchymal opacifications at a median of 78 days (range: 12–231 days). In another, 3 of those 12 patients (75%) residues with a fibrotic appearance (i.e., parenchymal bands, traction bronchiectasis, and reticulation) were observed at a median of 56 days (range: 42–67 days) (Fig. 2)—with equal distribution of dominant CT findings at onset of DAH. Two over 12 patients (17%) exhibited only mild residual changes on early follow-up CT at 6 days and 29 days after DAH onset (Fig. 3). One over 12 patients (8%) did not receive CT follow up. 4. Discussion Bleeding to the lung may originate from bronchial vessels, pulmonary vessels, or the alveolar microcirculation [15]. Diffuse pulmonary hemorrhage is caused by injury to the alveolar microcirculation and can occur with a variety of diseases such as different vasculitides, connective tissue disorders, infections, barotraumas, coagulation disorders, drug toxicities, or after hematopoietic SCT [12,15]. Depending on the cause and the clinical setting, DAH may exhibit a wide range of morbidity and mortality. Reported death rates between 45% and 100% for DAH are encountered mainly in patients who developed this complication after allo-SCT and cytotoxic chemotherapy [5,8]. These particular settings certainly are responsible for different pathomechanisms compared to simple pulmonary hemorrhage due to temporary thrombocytopenia, barotrauma, infection, or autoimmune diseases. Radiographic studies have long been and still are an integral part of the diagnostic work-up in patients with DAH [9]. However, due to its superiority for illustrating even subtle macroscopic patterns of lung parenchymal involvement, HRCT represents the mainstay of morphologic lung imaging in patients with hematological malignancies at risk for both, infectious and noninfectious pulmonary complications. Hence, HRCT is often obtained before definite confirmation of DAH by BAL. Yet, the field of CT morphologic
Fig. 2. CT morphologic disease course of DAH. Twenty-two-year-old male with common ALL experiencing DAH after allo-SCT. Initial findings at Day 17 posttransplant mainly consist of ill-defined centrilobular GGO with a bilateral and partly confluent, patchy distribution (A). Twenty-eight days later, following mechanical ventilation, high-dose steroids, and etanercept, opacifications evolve into centrally located coarse peribronchovascular consolidations accompanied by incipient lung parenchymal retraction and widening of bronchial lumina in the right upper lobe (B). Note concomitant pneumomediastinum, presumably due to mechanical ventilation. Another 28 days later, fibrotic changes partially resolve, leaving residual GGO and parenchymal distortion (C). The pneumomediastinum completely subsides.
684
D. Spira et al. / Clinical Imaging 37 (2013) 680–686
Fig. 3. CT morphologic disease course of DAH. Fifty-three-year-old male with mycosis fungoides experiencing DAH after allo-SCT. At Day 10 posttransplant, HRCT shows lobulardependent GGO with characteristic sparing of the ventral lung regions (A). Ten days later, the pattern is extended in distribution with increasing parenchymal attenuation but unchanged in character (B). The dependent distribution of lung parenchymal changes is still preserved. Another 19 days later, opacifications resolve, and only minimal residual abnormalities (i.e., few parenchymal bands) are noted (C). The patient survives, and no mechanical ventilation during the disease course was needed.
DAH imaging in patients with hematological malignancies is still neglected in the literature, and aside from rather scarce reports, we did not find a comprehensive analysis of HRCT findings in patients with hematological malignancies developing DAH [10–12]. The onset of DAH is usually within the first 30 days (median: 11– 24 days) following SCT [16], which was comparable to our patient group (median: 12 days). Symptoms typically include dyspnea, fever, and cough, whereas hemoptysis is rare [5]. In addition, we observed a significant proportion of patients with an initially elevated CRP (44%), which is explained by the inflammatory processes accompanying neutrophile influx through the endothelial wall into the lung parenchyma, chemotherapy-induced mucositis, and others. Even though severe, acute GVHD is recognized as a risk factor for DAH after SCT, no signs of acute GVHD in the skin, the liver, or the gut were noted in any of our patients [5,17]. Other defined risk factors are advanced age, total body irradiation, myeloablative conditioning regimen, and allogeneic donor source [5,17]. Some review articles and textbooks describe bilateral areas of GGO or consolidation as the most common CT findings in DAH [11,18,19], whereas others noted GGO with a superimposed reticular pattern as characteristical [12,20]. This is well represented in our series where two main groups were recognized at onset of DAH—those with predominant crazy paving and those with predominantly isolated GGO. However, both patterns reflect the same pathologic background of an acute alveolar/interstitial disease and should rather be seen as a continuum than separate entities, therefore constituting the main dominant finding in 76% of our patient group. At initial presentation, consolidation was seen in 31% of our patients, constituting the domi-
nant pattern in only 19%. Patterns of distribution and associated findings were similar in both subgroups and comparable to the radiographic analysis conducted by Witte et al. [9]. The preference of DAH for the middle/lower lung areas (probably due to gravitational forces on intraalveolar hemorrhage) and its bilateral distribution need to be regarded as nonspecific, even though infectious complications such as Pneumocystis jirovecii and cytomegalovirus show a preference for the upper lung areas or are randomly distributed. In our study, airspace bronchograms were seen in most patients, and a characteristic dependent distribution was frequently discernible (44%). It is tempting to explain both features with DAH pathophysiology and pulmonary microanatomy: bleeding in DAH originates from the alveolar capillaries [15], therefore primarily involves the alveolar spaces and alveolar ducts with sparing of the bronchial tree [21,22] and the consecutive appearance of airspace bronchograms. Furthermore, communications of alveoli through pores of Kohn and alveolar fenestrae as well as direct communications between alveoli and peripheral bronchioles (canals of Lambert) enable collateral ventilation and/or expansion of infiltrates, thus explaining the observed dependent distribution [23–25]. The appearance of a tree-in-bud sign is strongly suggestive for an infectious process. At times, however (in our study 6/16 DAH patients), it is also seen in patients with pulmonary hemorrhage, probably due to hemoptysis and consecutive microaspiration of blood into bronchioli [26]. The radiologist should be aware of the possibility of tree-in-bud occurring in DAH. Nevertheless, a high level of suspicion for infection needs to be maintained when the tree-in-bud sign is spotted.
Fig. 4. CT morphologic disease course of DAH. Forty-one-year-old male with T-cell non-Hodgkin lymphoma developed DAH after allo-SCT. Initial DAH opacifications at Day 14 posttransplant demonstrate central consolidation and reticulation with airspace bronchograms in the upper lobes and sparing of the lung periphery (A). Fourteen days later, the pattern is changed into a mixture of GGO, coarse reticulation (ventrally located), and fine reticulation (dorsally located). Note traction bronchiectasis in the middle lobe (B). Another 17 days later, the lung is strongly increased in attenuation due to a mixture of consolidation, diffuse GGO and reticulation, as well as persistent bronchiectasis (C). Also note extensive pneumomediastinum and thoracic wall emphysema, presumably due to mechanical ventilation. Despite mechanical ventilation, treatment with high-dose steroids, and etanercept, the patient dies the same day.
D. Spira et al. / Clinical Imaging 37 (2013) 680–686 Table 4 Outcome of subgroups defined by predominant HRCT pattern at onset of DAH Dominant CT finding at onset of DAH Crazy paving Isolated GGO Fine retic. Consolidation
Clinically resolved CT residua (at N40 d) Dominant GGO Retic./Parench. bands Traction bronchiectasis Death due to DAH Constant morphology Constant distribution Dominant GGO Dominant cons. Retic./Parench. bands Traction bronchiectasis
(n=6)
(n=6)
(n=1)
(n=3)
5/6 – – – – 1/6 1/1 1/1 – – – –
5/6 1/5 1/5 1/5 1/5 1/6 – 1/1 1/1 – 1/1 1/1
1/1 1/1 – 1/1 – – – – – – – –
1/3 1/1 1/1 – – 2/3 1/2 1/2 – 2/2 1/2 1/2
Bolded figures highlight the respective subgroups. Fine retic.=fine reticulation; retic./parench. bands=reticulation and parenchymal bands; dominant cons.=dominant consolidation.
Radiographically, the course of DAH is described as rapidly progressive from an initially mild interstitial or alveolar pattern in the central and lower lung towards a diffuse, severe alveolar pattern [9]. However, only scarce data exist on the CT morphologic course of DAH infiltrates. In our series, the disease course was unpredictable and independent of primary DAH CT morphology spanning a wide range of CT morphologic conversion varying from complete resolution (38%, median of 78 days) over fibrotic residues (37%, median of 56 days) to a diffusely fibrotic appearance, DAH progression, and death (25%, median: 21 days). This is at least in part explained by the various underlying mechanisms and response characteristics of the host, different combinatory treatment regimens applied, and variable extent of thrombocytopenia. Hence, an either intermittent or definite fibrotic conversion is suggested by their CT morphologic course in several of our patients. Progression of parenchymal consolidations at follow up was more often followed by worsening of respiratory function and death. We did not find any data concerning the later histopathology in patients surviving DAH after allo-SCT. However, DAH is intimately associated with the histologic picture of DAD [11,27,28], which is characterized by a uniform interstitial fibrosis in the organizing phase followed by either gradual resolution or continued interstitial fibrosis with architectural remodeling and progressive respiratory compromise [22]. When compared to the reported death rate of 45%–100% [5,8], the initial mortality due to progressive DAH was very low in our series (25%). The reduced mortality rate may primarily be related to the heterogeneity of pathomechanisms in the chemotherapy and SCT setting and variable risk profiles. However, a concise evaluation in this respect was beyond the scope of this study. Specialized patient care at the intensive care unit using modern treatment regimens such as early high-dose corticosteroids±cytokine inhibitors, including tumor necrosis factor alpha blockers, besides supportive measures (noninvasive ventilation, etc.) may also have contributed to improved patient outcome. Due to often similar CT morphology and distribution of opacifications, P. jirovecii pneumonia and viral lung infections (such as cytomegalovirus, herpes simplex virus, respiratory syncytial virus, and influenza virus) need to be excluded clinically and via BAL. Yet, a basal dominance and a dependent distribution are rarely seen in these infections and should rather suggest DAH in the appropriate clinical setting. The main noninfectious differential diagnoses to be considered are pulmonary edema, IPS, engraftment syndrome, granulocyte colony-stimulating factor (G-CSF)-induced lung changes, transplantassociated thrombotic microangiopathy (TA-TMA) and, eventually, transfusion-associated lung injury (TRALI). Pulmonary edema due to
685
fluid overload, renal or hepatic insufficiency, or cardiotoxicity rapidly responds to diuretic treatment and was excluded in this analysis. IPS is often similar in appearance but shows a variable yet nondependent distribution, holding a tendency towards a more rapid fibrotic conversion at follow up, frequently necessitating early mechanical ventilation. TA-TMA is diagnosed by clinical and laboratory data such as red blood cell fragmentation (≥ 2 schistocytes per high-power field), concurrent increased serum lactate dehydrogenase, concurrent renal and neurologic dysfunction, as well as negative direct and indirect Coombs test. TRALI is characterized by its intimate timely relation to stem cell transfusion, whereas DAH is a known complication of G-CSF application. Our study holds a number of limitations. Firstly, the dictate of the clinical course on the number and time windows between CT follow up complicates adequate comparison. Secondly, a large number of other variables related to the underlying disorders, such as type and intensity of therapies, age, and associated comorbidities, may have a significant impact on the evolution and outcome of DAH. Third, therapeutic management of DAH varied in our cohort, being determined by the degree of respiratory distress and other clinical parameters. The retrospective character and the rather small number of patients request larger prospective follow-up studies in order to validate our results. Nonetheless, to our knowledge, this is the first HRCT study to analyze the development, patterns of distribution, and course of this serious complication in patients with hematological malignancies. The intention of this work was to exhibit the radiologist's role in identifying characteristic changes and monitoring disease course of DAH, thereby endowing the treating hematologist with additional information and aiding in treatment decisions. 5. Conclusion Chest CT features of DAH are each by oneself nonspecific; however, the conjoint occurrence of bilateral, diffuse, and dependent GGO± crazy paving, accompanied by airspace bronchograms in the appropriate clinical setting, is consistent with this complication. Nevertheless, a broad differential diagnosis needs to be considered, including infectious agents (e.g., viruses, P. jirovecii) and noninfectious, treatment-related causes (e.g., pulmonary edema, IPS, TA-TMA, GCSF, and TRALI). The disease course of DAH is unforeseeable, spanning a wide range of CT morphologic conversion varying from complete resolution to increasing fibrosis, DAH progression and death, and can be well monitored via HRCT. References [1] Camus P, Fanton A, Bonniaud P, Camus C, Foucher P. Interstitial lung disease induced by drugs and radiation. Respiration 2004;71(4):301-26. [2] Limper AH. Chemotherapy-induced lung disease. Clin Chest Med 2004;25(1): 53-64. [3] Abratt RP, Morgan GW, Silvestri G, Willcox P. Pulmonary complications of radiation therapy. Clin Chest Med 2004;25(1):167-77. [4] Goldman AL, Enquist R. Hyperacute radiation pneumonitis. Chest 1975;67(5): 613-5. [5] Afessa B, Tefferi A, Litzow MR, Krowka MJ, Wylam ME, Peters SG. Diffuse alveolar hemorrhage in hematopoietic stem cell transplant recipients. Am J Respir Crit Care Med 2002;166(5):641-5. [6] Weisdorf DJ. Diffuse alveolar hemorrhage: an evolving problem? Leukemia 2003;17(6):1049-50. [7] Carreras E, Diaz-Ricart M. The role of the endothelium in the short-term complications of hematopoietic SCT. Bone Marrow Transplant 2011;46(12): 1495-502. [8] Gupta S, Jain A, Warneke CL, Gupta A, Shannon VR, Morice RC, Onn A, Jimenez CA, Bashoura L, Giralt SA, Dickey BF, Eapen GA. Outcome of alveolar hemorrhage in hematopoietic stem cell transplant recipients. Bone Marrow Transplant 2007;40(1):71-8. [9] Witte RJ, Gurney JW, Robbins RA, Linder J, Rennard SI, Arneson M, Vaughan WP, Reed EC, Dicke KA. Diffuse pulmonary alveolar hemorrhage after bone marrow transplantation: radiographic findings in 39 patients. AJR Am J Roentgenol 1991; 157(3):461-4.
686
D. Spira et al. / Clinical Imaging 37 (2013) 680–686
[10] Primack SL, Miller RR, Müller NL. Diffuse pulmonary hemorrhage: clinical, pathologic, and imaging features. AJR Am J Roentgenol 1995;164(2):295-300. [11] Frazer RS, Müller NL, Colman N, Paré PD, editors. Diagnosis of diseases of the chest. 4th ed. Philadelphia: W.B. Saunders; 1999. p. 1726-7. [12] Webb WR, Müller NL, Naidich DP, editors. High-resolution CT of the lung. 4th ed. Philadelphia, MD: Lippincott Williams & Wilkins; 2009. p. 462-7. [13] Parr DG, Stoel BC, Stolk J, Nightingale PG, Stockley RA. Influence of calibration on densitometric studies of emphysema progression using computed tomography. Am J Respir Crit Care Med 2004;170(8):883-90. [14] Austin JH, Müller NL, Friedman PJ, Hansell DM, Naidich DP, Remy-Jardin M, Webb WR, Zerhouni EA. Glossary of terms for CT of the lungs: recommendations of the Nomenclature Committee of the Fleischner Society. Radiology 1996;200(2): 327-31. [15] Lara AR, Schwarz MI. Diffuse alveolar hemorrhage. Chest 2010;137(5):1164-71. [16] Afessa B, Peters SG. Major complications following hematopoietic stem cell transplantation. Semin Respir Crit Care Med 2006;27(3):297-309. [17] Majhail NS, Parks K, Defor TE, Weisdorf DJ. Diffuse alveolar hemorrhage and infection-associated alveolar hemorrhage following hematopoietic stem cell transplantation: related and high-risk clinical syndromes. Biol Blood Marrow Transplant 2006;12(10):1038-46. [18] Afessa B, Peters SG. Noninfectious pneumonitis after blood and marrow transplant. Curr Opin Oncol 2008;20(2):227-33. [19] Worthy SA, Flint JD, Müller NL. Pulmonary complications after bone marrow transplantation: high-resolution CT and pathologic findings. Radiographics 1997; 17(6):1359-71.
[20] Coy DL, Ormazabal A, Godwin JD, Lalani T. Imaging evaluation of pulmonary and abdominal complications following hematopoietic stem cell transplantation. Radiographics 2005;25(2):305-17. [21] Roychowdhury M, Pambuccian SE, Aslan DL, Jessurun J, Rose AG, Manivel JC, Gulbahce HE. Pulmonary complications after bone marrow transplantation: an autopsy study from a large transplantation center. Arch Pathol Lab Med 2005;129(3):366-71. [22] Beasley MB. The pathologist's approach to acute lung injury. Arch Pathol Lab Med 2010;134(5):719-27. [23] Takaro T, Price HP, Parra SC. Ultrastructural studies of apertures in the interalveolar septum of the adult human lung. Am Rev Respir Dis 1979;119(3):425-34. [24] Pump KK. Fenestrae in the alveolar membrane of the human lung. Chest 1974;65(4):431-6. [25] Lambert MW. Accessory bronchiolealveolar communications. J Pathol Bacteriol 1955;70(2):311-4. [26] Okada F, Ando Y, Yoshitake S, Ono A, Tanoue S, Matsumoto S, Wakisaka M, Maeda T, Mori H. Clinical/pathologic correlations in 553 patients with primary centrilobular findings on high-resolution CT scan of the thorax. Chest 2007;132(6):1939-48. [27] Robbins RA, Linder J, Stahl MG, Thompson AB, Haire W, Kessinger A, Armitage JO, Arneson M, Woods G, Vaughan WP, et al. Diffuse alveolar hemorrhage in autologous bone marrow transplant recipients. Am J Med 1989;87(5):511-8. [28] Agustí C, Ramirez J, Picado C, Xaubet A, Carreras E, Ballester E, Torres A, Battochia C, Rodriguez-Roisin R. Diffuse alveolar hemorrhage in allogeneic bone marrow transplantation. A postmortem study. Am J Respir Crit Care Med 1995;151(4):1006-10.