Note: This copy is for your personal non-commercial use only. To order presentation-ready copies for distribution to your colleagues or clients, contact us at www.rsna.org/rsnarights.
AFIP ARCHIVES
883
From the Archives of the AFIP Pulmonary Alveolar Proteinosis1 CME FEATURE See accompanying test at http:// www.rsna.org /education /rg_cme.html
LEARNING OBJECTIVES FOR TEST 6 After reading this article and taking the test, the reader will be able to: ■■Discuss
the clinical manifestations, diagnostic pathologic features, and theoretic pathogenesis of pulmonary alveolar proteinosis.
■■Describe
the spectrum of radiologic manifestations of pulmonary alveolar proteinosis and formulate a differential diagnosis of crazypaving at CT.
■■List
the original AFIP observations made at radiologicpathologic correlation in pulmonary alveolar proteinosis.
Aletta Ann Frazier, MD • Teri J. Franks, MD • Erinn O. Cooke, MPH Tan-Lucien H. Mohammed, MD, FCCP • Robert D. Pugatch, MD Jeffrey R. Galvin, MD Pulmonary alveolar proteinosis (PAP) may develop in a primary (idiopathic) form, chiefly during middle age, or less commonly in the setting of inhalational exposure, hematologic malignancy, or immunodeficiency. Current research supports the theory that PAP is the result of pathophysiologic mechanisms that impair pulmonary surfactant homeostasis and lung immune function. Clinical symptomatology is variable, ranging from mild progressive dyspnea to respiratory failure. There is a strong association with tobacco use. The predominant computed tomographic feature of PAP is a “crazy-paving” pattern (smoothly thickened septal lines on a background of widespread ground-glass opacity), often with lobular or geographic sparing. The radiologic differential diagnosis of crazy-paving includes pulmonary edema, pneumonia, alveolar hemorrhage, diffuse alveolar damage, and lymphangitic carcinomatosis. Definitive diagnosis is made with lung biopsy or bronchoalveolar lavage specimens that reveal intraalveolar deposits of proteinaceous material, dissolved cholesterol, and eosinophilic globules. Symptomatic treatment includes whole-lung lavage, and multiple procedures may be required. New therapies directed toward the identified defect in immune defense have met with moderate clinical success. radiographics.rsnajnls.org
TEACHING POINTS See last page
Abbreviations: AFIP = Armed Forces Institute of Pathology, BAL = bronchoalveolar lavage, GM-CSF = granulocyte-macrophage colony-stimulating factor, H-E = hematoxylin-eosin, ILS = interlobular septa, PAP = pulmonary alveolar proteinosis, WLL = whole-lung lavage RadioGraphics 2008; 28:883–899 • Published online 10.1148/rg.283075219 • Content Code: 1 From the Departments of Radiologic Pathology (A.A.F., J.R.G.) and Pulmonary and Mediastinal Pathology (T.J.F.), Armed Forces Institute of Pathology, 14th St and Alaska Ave NW, Washington, DC 20306; Department of Diagnostic Radiology, University of Maryland School of Medicine (A.A.F., R.D.P., J.R.G.), and University of Maryland School of Medicine (E.O.C.), Baltimore, Md; and Section of Thoracic Imaging, Division of Radiology, Cleveland Clinic, Cleveland, Ohio (T.-L.H.M.). Received December 3, 2007; revision requested December 19 and received February 5, 2008; accepted February 21. All authors have no financial relationships to disclose. Address correspondence to A.A.F. (e-mail:
[email protected]).
The opinions and assertions contained herein are the private views of the authors and are not to be construed as official nor as representing the views of the Departments of the Navy, Army, or Defense.
884 May-June 2008
Introduction
Teaching Point
Pulmonary alveolar proteinosis (PAP) is a rare disease characterized by abnormal intraalveolar accumulation of surfactant-like material (1). This year marks the 50th anniversary of its initial description by the eminent pathologists Rosen, Castleman, and Liebow (2). In 1958, Dr Rosen was Chief of Pulmonary and Mediastinal Pathology at the Armed Forces Institute of Pathology (AFIP), and the majority of patients in his article originated from the AFIP. In the intervening years, less than 500 cases have been reported in the literature (3). In this article, we discuss and illustrate PAP in terms of clinical manifestations and evaluation, radiologic findings and differential diagnosis, treatment and prognosis, and radiologic-pathologic correlation. In so doing, we make use of the largest collection of PAP cases to date, contained in the archives of the Departments of Radiologic Pathology and Pulmonary and Mediastinal Pathology at the AFIP. Despite the rarity of PAP, the conceptual understanding and the factors linked to the pathogenesis of this disease entity have advanced remarkably in recent decades. Three distinct subgroups of PAP are currently recognized: idiopathic, secondary, and congenital. Idiopathic PAP (also termed “acquired” or “adult-type” PAP) accounts for the great majority of cases (90%). This form occurs worldwide, with an incidence of 0.36 new cases per 1 million persons each year and a prevalence of 3.7 cases per 1 million persons (4). Secondary PAP (5%–10% of cases) is recognized in patients with industrial inhalational exposure to materials such as silica particles, cement dust, aluminum dust, titanium dioxide, nitrogen dioxide, and fiberglass; underlying hematologic malignancy; or immunodeficiency disorders (including cytotoxic or immunosuppressive therapy and human immunodeficiency virus infection) (1,3). Congenital PAP is quite rare (2% of cases) and manifests in the neonatal period with severe hypoxia (3,4). “Congenital PAP” is not universally regarded as a true form of PAP, but may instead represent the disease entity termed “chronic pneumonitis of infancy” (1); in any case, the prognosis is poor in this subtype (1). PAP appears to be the final common expression of distinct but related pathophysiologic mechanisms that affect pulmonary surfactant and lung immune function. Inherited or acquired
RG ■ Volume 28 • Number 3
genetic mutations appear to lead to a deficiency of surfactant proteins or impaired function of granulocyte-macrophage colony-stimulating factor (GM-CSF), an essential regulator of immune defense and surfactant homeostasis (3,5–7). Patients with idiopathic PAP (but not those with secondary or congenital PAP) also have high levels of autoantibodies against GM-CSF in blood and tissues, including the pulmonary alveoli (3,5,6). Research suggests that these antibodies neutralize the role of GM-CSF in the terminal differentiation of alveolar macrophages, thereby critically impairing the process of surfactant clearance in the lung (6,8). Furthermore, these antibodies appear to neutralize the antimicrobial activity of GM-CSF by inducing lung neutrophil dysfunction (6,9).
Clinical Manifestations and Evaluation
Patients with idiopathic or secondary PAP experience nonspecific, moderate respiratory symptoms including progressive dyspnea (mean duration, 7 months; however, onset may last for years) and dry or minimally productive cough (3,4). Less common signs and symptoms include fatigue, weight loss, low-grade fever, chest pain, and hemoptysis (3,4). Physical examination may reveal crackles, clubbing, or cyanosis (3). The mean patient age at diagnosis is 40 years (SD ± 13 years). There is a strong association with tobacco use: About three-quarters of PAP patients are smokers, and in this subgroup, men are three times more frequently affected than women. In nonsmokers, there is no gender predilection (3,4,9). The most common elevated serologic marker for PAP is an elevated lactate dehydrogenase level (82% of cases), but this finding is nonspecific. Blood gases show moderate hypoxemia with increased arterial oxygen tension (PaO2) and increased alveolar-arterial oxygen tension difference (AaPO2) (3,5). Pulmonary spirometry in PAP reveals gas exchange impairment (decreased carbon monoxide–diffusing capacity) and mild to moderate restrictive ventilatory defect (3,4). As noted earlier, patients with idiopathic PAP also have antibodies to GM-CSF in the blood and tissues, as well as in bronchoalveolar lavage (BAL) fluid (5,10,11). Although these antibodies are considered highly sensitive and specific markers for idiopathic PAP, the titers do not correlate with other markers of disease severity such as serum lactate dehydrogenase, PaO2, or AaPO2 (5,9–11).
RG ■ Volume 28 • Number 3
Frazier et al 885
Figure 1. (a) PAP in a 61-year-old man with chronic myelogenous leukemia and recent onset of fatigue and cough. Posteroanterior chest radiograph shows symmetric, perihilar ground-glass and reticulonodular opacities. Note the relative sparing of the costophrenic angles. (b) PAP in a 17-year-old boy with mild cough and dyspnea that had persisted for several years. Chest radiograph shows dense bilateral consolidation with relative sparing of the apices and right costophrenic angle. (c) PAP in a 36-year-old man with a history of inhalational exposure to beryllium. Chest radiograph shows symmetric perihilar consolidation with sparing of the costophrenic angles and apices.
tibodies (9). It is considered surprising that PAP patients develop microbial infection relatively uncommonly, given the conditions of macrophage and neutrophil dysfunction (8,12).
Radiologic Findings and Differential Diagnosis
Superimposed infectious pneumonia affects approximately 13% of all PAP patients (9). The increased risk of developing pneumonia in PAP may be due to macrophage dysfunction or the microbial growth medium provided by intraalveolar proteinaceous material (12). Complicating infectious pneumonias in PAP are often opportunistic, and agents include Nocardia, Candida, Cryptococcus neoformans, Aspergillus, cytomegalovirus, tuberculous and nontuberculous mycobacteria, Histoplasma capsulatum, Pneumocystis jirovecii, and Streptococcus pneumoniae (9,12–15). These infections may be disseminated, suggesting a systemic disorder such as the proposed “neutralization” of GM-CSF activity by circulating autoan-
Chest radiography is a helpful first step in diagnostic imaging but remains nonspecific for PAP. The typical radiograph reveals bilateral central and symmetric lung opacities, with relative sparing of the apices and costophrenic angles (Fig 1) (2,3). Less commonly, radiographs show multifocal asymmetric opacities or extensive diffuse consolidation without any clear zonal predominance (Fig 2) (3,9,16). Opacities range from a groundglass appearance with indistinct margins, to reticular or reticulonodular, to consolidation with air bronchograms (Fig 2b) (2,17–19). Although
886 May-June 2008
RG ■ Volume 28 • Number 3
Figure 2. (a) PAP in a 45-year-old female smoker with a 3-month history of dyspnea and productive cough (whitish sputum). Chest radiograph shows asymmetric reticulonodular opacities and multifocal consolidation. (b) PAP in a 31-year-old man with a 6-month history of mild dyspnea and digital clubbing. Chest radiograph demonstrates bilateral asymmetric opacities ranging from consolidated to reticulonodular to ground-glass opacities.
Figure 3. PAP in a 46-year-old man with mild progressive dyspnea. (a) Chest radiograph shows bilateral reticulonodular and patchy consolidated opacities limited to the midlung zones. (b) Follow-up chest radiograph obtained 5 months later shows increased opacification in the right lung with partial interval resolution of left-sided opacities, findings that reflect the evanescent nature of PAP in some cases.
the radiographic appearance of PAP suggests pulmonary interstitial edema, pleural effusion and cardiomegaly are absent (12,13,16,17,20). The Kerley B (septal) lines depicted at computed
tomography (CT) are rarely detectable on radiographs (2,16,21). There is often notable disparity between the moderate clinical symptomatology of PAP and the more impressive radiographic abnormalities (“clinicoradiologic discrepancy”) (3,17,19).
RG ■ Volume 28 • Number 3
Frazier et al 887
Figure 4. (a) Known PAP in a 43-year-old woman with a 6-month history of cough and acute onset of complicating pneumonia. Chest radiograph shows rounded consolidation with central cavitation (arrowheads) in the retrocardiac portion of the left lower lobe. (b) Known PAP in a 46-year-old woman who presented with acute fever and chills. Chest radiograph reveals the recent development of dense left perihilar consolidation and pleural effusion, findings that are suggestive of superimposed pneumonia in the appropriate clinical setting. Nocardial pneumonia was confirmed at autopsy. Figure 5. PAP in a 50-year-old man with severe dyspnea. CT scan (lung window) acquired through the upper lobes demonstrates widespread ground-glass opacity with focal areas of sparing and strikingly prominent septal lines (crazy-paving pattern).
Prior to the routine use of therapeutic lung lavage, abnormalities at chest radiography might persist or evanesce over months or years (Fig 3); only in exceptional cases did opacities resolve significantly. In that era, findings of pleural effusion, lymphadenopathy, or focal cavitary consolidation were considered highly suggestive of complicating pulmonary infection, a principle that is still operative today (Fig 4) (2,13,14,17,20,22). Before the broad clinical and radiographic recognition of PAP, this disease was occasionally misdiagnosed
as a primary lung infection, usually active tuberculosis or “burnt-out” pneumocystis pneumonia (2,14,23). CT provides greater anatomic detail and information concerning disease extent. The CT appearance of “crazy-paving,” defined as a network of smoothly thickened reticular (septal) lines superimposed on areas of ground-glass opacity, was first described in a group of six patients with PAP (Fig 5) (16,24,25). Areas of crazy-paving in PAP are typically widespread and bilateral, often with sharply marginated areas of geographic or lobular sparing (Fig 6) (13,16,22). There are widely variable patterns of regional or zonal predominance, including symmetric or asymmetric apical, basilar, central, peripheral, lobar, or diffuse lung involvement (13,19,22). Extent and degree of CT ground-glass opacity or consolidation appear
Teaching Point
888 May-June 2008
RG ■ Volume 28 • Number 3
Figure 6. Crazy-paving in PAP. (a) Coronal reformatted image (lung window) obtained in a 35-year-old man shows geographic areas of ground-glass opacity and septal thickening in an asymmetric distribution. (b) Coronal reformatted image (lung window) obtained in a 45-year-old man demonstrates an extensive crazy-paving pattern with sparing of the costophrenic angles, basilar subpleural zones, and lung apices.
Figure 7. PAP in a 50-year-old man. (a) CT scan (lung window) obtained prior to therapeutic BAL shows patchy areas of ground-glass opacity, thickened septal lines, and consolidation. (b) Post-BAL CT scan (lung window) obtained at the same level reveals residual septal lines but near-complete resolution of the ground-glass opacity.
Teaching Point
to correlate directly with severity of compromised pulmonary functional parameters—namely, restrictive ventilation, decreased diffusing capacity, and hypoxemia (19). Posttherapeutic BAL CT may reveal persistent septal lines despite interval resolution of ground-glass opacity (Fig 7) (17,19,22). Although the CT finding of crazy-paving is highly characteristic of PAP, it is also seen in several infectious, hemorrhagic, neoplastic, inha-
lational, and idiopathic conditions as well as in straightforward hydrostatic pulmonary edema. Therefore, the radiologic differential diagnosis of crazy-paving is broad and includes left heart failure, pneumonia (especially pneumocystis pneumonia), alveolar hemorrhage, bronchoalveolar carcinoma, lymphangitic carcinomatosis, diffuse alveolar damage (adult respiratory distress syndrome), radiation- or drug-induced pneumonitis, hypersensitivity pneumonitis, and pulmonary veno-occlusive disease (Fig 8) (26–32).
Figure 8. CT scans show a spectrum of lung diseases that are part of the differential diagnosis of PAP and manifest with varying degrees of ground-glass opacity, septal lines, and consolidation: cryptococcal pneumonia (a), diffuse alveolar damage (b), Erdheim-Chester disease (c), pulmonary veno-occlusive disease (d), cardiogenic pulmonary edema (e), pneumocystis pneumonia (f), pulmonary hemorrhage (g), and lymphangitic carcinomatosis (h).
RG ■ Volume 28 • Number 3
Frazier et al 889
890 May-June 2008
Treatment and Prognosis In unusual cases, clinical remission or quiescence of PAP may occur, but intervention is required in the majority of patients, and treatment depends on the particular form of disease (4,9). Congenital PAP requires supportive measures or lung transplantation. Secondary PAP demands removal of the inciting agent from the patient’s environment. Patients with idiopathic PAP are treated with sequential therapeutic whole-lung lavage (WLL), a procedure for removing lipoproteinaceous material from pulmonary alveoli with use of saline solution and chest percussion (3). Approximately 63% of patients with idiopathic PAP require WLL within 5 years of diagnosis (4). Therapeutic lung lavage originated in 1960, when Dr Ramirez-Rivera performed “segmental flooding” of the lung in PAP patients, inciting copious “white viscid” sputum production that led to significant respiratory functional improvement; Ramirez-Rivera initiated WLL for PAP in 1964 (9). The mortality rate in PAP approached 30% prior to the broad application of WLL, but over the past 3 decades, the 5-year survival rate of patients receiving WLL has been 95% (3). The Cleveland Clinic reports equal success with unilateral WLL or bilateral WLL performed in a single session. Potential postprocedural complications include pneumonia, sepsis, adult respiratory distress syndrome, and pneumothorax (3). The mean symptom-free interval after WLL is 15 months, and repeat treatments are often necessary (3). After undergoing two sequential WLLs, more than 60% of patients regain normal “exertional capacity” (3). Recent investigation suggests that BAL fluid levels of anti–GM-CSF antibodies may prove helpful in predicting the need for repeated WLL, but the utility of assessing serum or BAL levels of anti–GM-CSF antibodies for monitoring disease activity or treatment response remains uncertain (5). Efforts since 1994 to address the underlying pathophysiology of idiopathic PAP have met with moderate success (6,33). To combat GM-CSF autoantibodies, exogenous GM-CSF has been administered as an aerosol or subcutaneously to several PAP patients, with an overall response rate of approximately 50% (3,4,9). Unfortunately, because of differences in underlying pathogenesis and genetic defects, patients with congenital PAP do not respond to GM-CSF therapy (3). Double lung transplantation may be performed in congenital PAP, when a patient fails to improve with WLL, or when PAP (rarely) progresses
RG ■ Volume 28 • Number 3
to pulmonary fibrosis (3,34). PAP has been reported to recur following double transplantation, which argues in favor of a systemic disorder probably related to the circulating antibody to GM-CSF (9,34). It is interesting to note that a PAP-like condition may also develop in lung allograft recipients without prior PAP, possibly secondary to immunosuppressive therapy (3,35). Overall survival rates for PAP now approach 100%, but if death does result directly from PAP, it is due to either respiratory failure (80% of cases) or pulmonary infection (20%) (3,4).
AFIP Archives Case Review
Ninety-eight cases of PAP were identified in the archives of the Departments of Radiologic Pathology and Pulmonary and Mediastinal Pathology at the AFIP (Table). Some cases may have been published previously, given the secondary consultation nature of practice at the AFIP. Clinical information was not always available, but many cases included records of clinical presentation, surgery reports, pathology reports, or hospital discharge summaries. Chest radiographs were available in 89 patients and CT scans in 28 patients. Imaging studies were reviewed retrospectively by three thoracic radiologists (A.A.F., J.R.G., R.D.P.). Selected gross specimen photographs and hematoxylin-eosin (H-E)–stained tissue sections were reviewed by a pulmonary pathologist (T.J.F.) with two thoracic radiologists (A.A.F., J.R.G.). Patients ranged in age from 8 months to 64 years (mean age, 38 years). The ratio of male to female patients was greater than 2:1. Data on clinical presentation were available in 72 of 98 patients. The most common presenting symptoms were dyspnea (59% of cases) and cough (54%). The mean period of clinical onset was 4 months (range, 3 days–10 years). Less common signs and symptoms included fever (13% of cases), chest pain (11%), fatigue (10%), hemoptysis (6%), cyanosis and clubbing (4%), and respiratory distress requiring intubation (1%). Twenty-one patients had a documented history of tobacco use, but we believe this is a significant underestimation given the fact that in many of the earlier AFIP cases, the patients were male soldiers. Fourteen patients had some form of documented inhalational exposure to materials including beryllium, cement dust, wood dust, drywall, masonry and bricks, embalming fluid, agricultural materials, and metal dust. There were 11 cases of complicating pulmonary infection, two of which were specifically identified as nocardial pneumonia and two as cryptococcal pneumonia. Hematologic malignancy (myeloid leukemia) was documented in two patients.
RG ■ Volume 28 • Number 3
Frazier et al 891
AFIP Case Review of 98 Cases of PAP No. of Cases*
Features Patients (n = 98) Male Female Symptoms (n = 72) Dyspnea Cough Fever Chest pain Fatigue Hemoptysis Cyanosis and clubbing Respiratory distress requiring intubation Radiographic findings (n = 89) Disease distribution Perihilar Peripheral Diffuse Patterns of sparing Costophrenic angle Apical Peripheral CT findings (n = 28) Disease distribution Diffuse Central Peripheral Multifocal Patterns of sparing Costophrenic angle Subpleural Geographic-lobular Disease extent† 0%–25% 25%–50% 50%–75% 75%–100% Airway irregularity Fissural distortion
67 (68) 31 (32) 43 (59) 39 (54) 9 (13) 8 (11) 9 (10) 8 (6) 7 (4) 1 (1) 57 (64) 12 (14) 20 (22) 65 (73) 65 (73) 42 (47) 20 (71) 4 (14) 1 (4) 3 (11) 14 (50) 9 (32) 15 (54) 3 (11) 3 (11) 6 (21) 15 (57) 17 (39) 20 (29)
*Numbers in parentheses indicate percentages. †Percentage
of lungs.
Several patterns of radiographic opacity were appreciated at posteroanterior chest radiography. The most common pattern was reticulonodular opacity (26% of cases), followed by combined reticulonodular opacity and consolidation (25%), consolidation only (25%), ground-glass opacity only (17%), and combined ground-glass and reticulonodular opacity (8%). Disease extent was bilateral in 99% of cases and unilateral in 1%. The majority of cases (56%) demonstrated bilateral symmetric opacities, whereas 44% appeared asymmetric. Radiographic opacities were most
often perihilar in distribution (64% of cases) and much less commonly diffuse (22%) or peripheral (14%). The midlung zones were most severely affected in 44% of cases, the lower lung zones in 27%, the entire lung in 25%, and the upper lung zones in 5%. Chest radiographs showed regional sparing in 84% of cases. With some overlap in findings, there were patterns of costophrenic angle sparing in 73% of cases, apical sparing in 73%, and peripheral sparing in 47%. Although the preponderance (75%) of CT cases demonstrated the crazy-paving pattern, 25% alternatively showed ground-glass opacity without evidence of prominent septal lines; additional focal areas of consolidation were present in 46% of cases. CT opacity was always bilateral and demonstrated the following spectrum of distribution patterns: diffuse (71% of cases), chiefly central (14%), multifocal-patchy (11%), and chiefly peripheral (4%). Disease was distributed most often throughout all lung zones (71% of cases), followed by an equal number of cases in which the disease predominantly affected the upper (14%) or lower (14%) lung zones. Overall extent of parenchymal opacification was divided into four categories: 75%–100% of the lungs (57% of cases), 50%–75% (21%), 25%–50% (11%), and 0%–25% (11%). Three patterns of sparing were noted, with some overlap in findings: geographic-lobular (54% of cases); costophrenic angle, including the supradiaphragmatic area (50%); and subpleural (32%). The airways demonstrated irregularity in 39% of cases. Fissural irregularity or distortion was evident in 29% of cases. Two CT scans demonstrated lobar atelectasis. Pleural effusion and lymphadenopathy were absent at all CT studies.
Radiologic-Pathologic Correlation at the AFIP Although the disease was first named and is generally still known as “pulmonary alveolar proteinosis,” the term lipoproteinosis is often used and better describes the chemical composition of the material filling alveoli (2). Evidence of lipid content is present grossly as well as microscopically. At gross examination, portions of lung parenchyma involved by PAP are firm with yellow cut surfaces that protrude above the level of the adjacent uninvolved lung. Tissue firmness results from filling of alveolar spaces, whereas the yellow color reflects the presence of lipid (Fig 9).
892 May-June 2008
RG ■ Volume 28 • Number 3
Figure 9. Photograph of a gross specimen shows areas involved by PAP with yellow cut surfaces that protrude above the adjacent uninvolved lung.
Figure 10. Histopathologic findings of PAP. (a) Medium-power photomicrograph (original magnification, ×200; H-E stain) of the lung shows intraalveolar deposits of lipoproteinaceous material circumscribed by normal alveolar walls (arrowhead). (b) On a different medium-power photomicrograph (original magnification, ×200; H-E stain), the alveolar walls (*) appear thickened due to infiltrates of chronic inflammatory cells and type 2 pneumocyte hyperplasia lining the walls. The thick walls enclose alveolar spaces filled with finely granular eosinophilic deposits in which there are acicular (needle-shaped) clefts of dissolved cholesterol; eosinophilic globules; and bluish, degenerating cell remnants. (c) Medium-power photomicrograph (original magnification, ×200; periodic acid–Schiff stain) demonstrates the weak positivity and granularity of the intraalveolar deposits with this stain. (d) Low-power photomicrograph (original magnification, ×40; H-E stain) shows how, in PAP, the interlobular septa (ILS) are remarkably abnormal, expanded by edema and massively dilated lymphatic channels (*).
RG ■ Volume 28 • Number 3
Frazier et al 893
Figure 11. (a) CT scan (lung window) obtained in a 47-year-old woman with PAP shows a crazy-paving pattern with notable lobular and geographic sparing. (b) Low-power photomicrograph (original magnification, ×100; H-E stain) shows a mildly edematous ILS (arrowhead) forming a barrier between secondary lobules, thereby limiting the extension of intraalveolar material. (c) Low-power photomicrograph (original magnification, ×10; H-E stain) shows the lung parenchyma, which has a vaguely nodular appearance owing to the limitation imposed by the ILS. (d) CT scan (lung window) obtained in a 30-year-old man with PAP demonstrates bilateral, multifocal punctate nodules (arrowheads) within patchy areas of ground-glass opacity.
Teaching Point
At microscopic examination, PAP is characterized by intraalveolar accumulations of lipoproteinaceous material circumscribed by normal or thickened alveolar walls (Fig 10a). The accumulations are finely granular eosinophilic deposits containing acicular clefts of dissolved cholesterol, eosinophilic globules, foamy macrophages, and ghostlike cell remnants (Fig 10b). Periodic acid–Schiff stain highlights the granularity of the deposits, which are weakly positive with this stain (Fig 10c). Although the interstitium of the lung is most often normal, alveolar walls can be quite abnormal, thickened by type 2 pneumocyte hyperplasia and variable combinations of chronic inflammatory cells and fibrosis (Fig 10a, 10b). The most common CT feature of PAP is widespread groundglass opacity and smoothly thickened ILS, the so-called crazy-paving pattern. The ground-glass opacity appears to represent the combined density
of alveolar walls and intraalveolar substance. Prominent septal lines at CT correspond to remarkably abnormal ILS that are expanded by edema and massively dilated lymphatic channels (Fig 10d). CT often reveals geographic or lobular sparing of the lung, with affected areas that are sharply demarcated from and interposed with normalappearing lung (Fig 11a). Involved lung is sharply demarcated from uninvolved lung by the physical barrier of the ILS, which limits the extension of intraalveolar material between secondary lobules (Fig 11b, 11c). Discrete nodules are also occasionally evident at CT (Fig 11d), and at microscopic examination, PAP may indeed appear as well-circumscribed nodular areas (Fig 11c).
894 May-June 2008
RG ■ Volume 28 • Number 3
Figure 12. (a, b) PAP in a 34-year-old man with chronic progressive dyspnea. (a) CT scan (lung window) obtained in the early phase of PAP shows extensive ground-glass opacity, scattered septal lines, and patchy consolidation in the upper lobes. (b) CT scan (lung window) obtained 1 year later reveals substantial clearing of the crazy-paving pattern and dependent consolidation. The upper lobe airways now appear distorted and ectatic (arrows), findings that are compatible with intervening fibrosis. (c) Coronal reformatted image (lung window) obtained in a 42-year-old woman with PAP shows bilateral lower lobe ground-glass opacity with associated mild bronchiectatic changes. Note the relative paucity of thickened septal lines. (d, e) Medium-power (original magnification, ×200; H-E stain) (d) and low-power (original magnification, ×100; H-E stain) (e) photomicrographs obtained in two other patients with PAP show fibrosis of the alveolar walls (* in d) and ILS (* in e), findings that correlate with the presence of traction bronchiectasis at CT and that are seen in some cases of PAP.
RG ■ Volume 28 • Number 3
Frazier et al 895
Figure 13. PAP in a 50-year-old man. (a, b) Coronal (a) and sagittal (b) reformatted images (lung window) show irregular thickening and distortion of the lobar fissures, accompanied by a patchy distribution of crazy-paving. (c) Low-power photomicrograph (original magnification, ×10; H-E stain) shows how fibrotic ILS (*) may even appear to tug on the pleural surface, producing a dimpled or puckered surface contour.
Unusual cases of PAP demonstrate CT findings of mild traction bronchiectasis and focal fissural distortion (Fig 12a, 12b). In this setting, the crazypaving pattern is often absent or less distinct, and there is an increase in the overall apparent density of the lung (Fig 12c). Light microscopy reveals evidence of fibrosis in the alveolar walls (Fig 12d) and ILS (Fig 12e). Fibrotic ILS may even appear to “tug on” or focally contract the pleural surface
(Fig 13). These histopathologic findings raise the possibility that underlying interstitial fibrosis may help explain the CT findings of airway traction, pleural distortion, and increased lung opacity that are occasionally evident in PAP. Although the pathologic differential diagnosis of PAP includes any disorder with eosinophilic intraalveolar deposits, pulmonary edema and pneumocystis pneumonia are the most common considerations. In contrast to PAP, edema fluid lacks granularity, acicular clefts, eosinophilic
896 May-June 2008
RG ■ Volume 28 • Number 3
Figure 14. Medium-power photomicrographs (original magnification, ×200; H-E stain) demonstrate the homogeneous quality of edema fluid (a) and the foamy or honeycomb appearance of pneumocystis pneumonia exudates (b). Figure 15. (a) Photograph shows a glass cylinder containing opalescent therapeutic BAL effluent. Scale is in milliliters. (b) High-power photomicrograph (original magnification, ×400; H-E stain) demonstrates the granular quality of BAL effluent. Granularity is characteristic of but not specific for PAP. Chest CT scans with features characteristic of PAP provide support for the cytologic diagnosis; thus, correlation with imaging studies in this setting is prudent.
globules, and foamy macrophages (Fig 14a). Pneumocystis gives rise to foamy or “honeycomb” exudates, which are readily distinguished from PAP by the presence of characteristic organisms noted with a Gomori methenamine silver stain (Fig 14b). Exclusion of micro-organisms is essential when exudates are present, and other stains such as Ziehl-Neelsen stain for mycobacteria may be required in addition to Gomori methenamine silver stain.
Definitive diagnosis of PAP is typically made at transbronchial or surgical lung biopsy, but cytologic specimens from sputum or BAL fluid can also be diagnostic (12). Cytologic preparations contain the characteristic finely granular proteinaceous material with rare background macrophages or inflammatory cells; however, these findings are not specific (Fig 15). As with biopsy material, exclusion of morphologically similar entities is essential.
Discussion At chest radiography, PAP typically manifests with bilateral, symmetric central opacities. In our re-
RG ■ Volume 28 • Number 3
Frazier et al 897
Figure 16. PAP in a 46-year-old man with a 3-month history of dyspnea. CT scan (lung window) obtained at the level of the right upper lobe bronchus shows a diffuse distribution of ground-glass-opacity nodules, with focal geographic and subfissural sparing.
Figure 17. (a) PAP in a 43-year-old woman. CT scan (lung window) shows a crazy-paving pattern with notable subpleural sparing, particularly in the lateral lung zones. (b) PAP in a 50-year-old man. CT scan (lung window) shows geographic and well-circumscribed nonuniform areas of subpleural sparing located predominantly in the posterior lungs and costophrenic angles.
view, the opacities are most often reticulonodular, either solely or combined with consolidation. Our review further demonstrates the broad spectrum of radiographic appearances of PAP (Fig 1), including a significant proportion (44%) of cases with asymmetry (Fig 2). Early AFIP cases with evanescent opacities evolving over months or years may reflect the natural course of disease prior to the era of therapeutic lung lavage (Fig 3). Dense focal consolidation, particularly with cavitation, suggests a complicating pneumonia (Fig 4). At CT, there is often remarkably widespread lung involvement in PAP. During our review, we noted a surprising discrepancy between the degree of opacification at chest radiography and the more extensive disease at CT (although our observations are not quantified). The predominant but non-
specific CT feature of PAP is diffuse or multifocal crazy-paving; interestingly, however, 25% of AFIP cases demonstrate only ground-glass opacity (Fig 16). Furthermore, our review illustrates many patterns of parenchymal sparing, including geographic, costophrenic angle, supradiaphragmatic, subpleural, and perifissural sparing (Figs 6, 16, 17). Our small sample of postlavage CT scans helps confirm prior descriptions of residual septal lines with diminished ground-glass opacity (Fig 7). Finally, it is important to note that a large proportion of PAP patients are smokers, and indeed, we found that PAP may be superimposed on emphysema; an underlying lung carcinoma may even be present (Fig 18).
RG ■ Volume 28 • Number 3
898 May-June 2008
Teaching Point
The sharply marginated zones of disease adjacent to normal lung correspond with zones of affected alveoli, filled with proteinaceous material and limited by adjacent ILS. Our radiologicpathologic observations further reveal that prominent septal lines in PAP correlate with remarkable ILS expansion due to edema and dilatation of the pulmonary lymphatic system, a strong sign of disturbed pulmonary fluid homeostasis. The presence of such lymphatic engorgement raises several questions. Is there an underlying abnormality in PAP, either primary or secondary, that leads to capillary injury and consequent leakage of fluid into the interstitium? Does the intraalveolar substance itself present a significant challenge to the lymphatic system as it strives to maintain pulmonary homeostasis by removing fluid and macromolecules? Does the pathophysiology of PAP affect the functional capacity of the lymphatic vessels by a mechanism of direct injury? We found a surprising number of AFIP cases with airway irregularity (almost 40%) and fissural distortion (almost 30%). When CT depicts these abnormalities, often accompanied by increased lung opacity, underlying interstitial fibrosis is evident at microscopic examination. It is interesting to ponder the possible sequence of events in PAP disease progression: Do active inflammation and edema in the early phase of PAP act as mediators, driving (at least in some patients) the disease process toward fibrosis? At the 50th anniversary of the original description of PAP, we stand on the threshold of active research and discovery concerning this disease entity. The early years led us to appreciate three distinct classes of PAP and to develop therapeutic measures such as WLL. More recently, exciting advances in animal model and human research have provided new clues to the mechanisms of pathogenesis in PAP. As Dr Ioachimescu at the Cleveland Clinic asserts, “PAP may be the first human disease in which an autoantibody against a growth factor (GM-CSF) is linked to disease pathogenesis” (3). Scientific discovery is driving forward new treatment strategies such as exogenous GM-CSF therapy, and ongoing research in PAP may enlighten us about other human conditions of impaired immune defense and pulmo-
Figure 18. PAP in a 62-year-old male smoker. CT scan (lung window) shows bilateral widespread areas of ground-glass opacity, scattered septal lines, and parenchymal emphysema. Note the spiculated density in the left upper lobe (arrowhead), a finding that represents a lung cancer.
nary homeostasis. As a complement to these advances, CT has shown objective correlation with serologic, BAL, and functional markers of disease severity; thus, pulmonary imaging holds promise for future disease assessment and surveillance. Acknowledgments: The authors express their deep appreciation to the radiology residents—past, present, and future—who provide case contributions to the Department of Radiologic Pathology at the AFIP. We extend sincere thanks to Tracy V. Faulkner, PharmD, for her skilled persistence in acquiring detailed patient information; and to E. James Britt, MD, at the University of Maryland School of Medicine for his keen clinical insights. We also thank Kim Jones, administrative assistant in the Department of Pulmonary and Mediastinal Pathology, as well as Janice Danqing Liu, Anika Torruella, and Robin E. Whitt in the Department of Radiologic Pathology for their gracious research assistance.
References 1. Katzenstein ALA. Katzenstein and Askin’s surgical pathology of non-neoplastic lung disease. 4th ed. Philadelphia, Pa: Saunders Elsevier, 2006. 2. Rosen SH, Castleman B, Liebow AA. Pulmonary alveolar proteinosis. N Engl J Med 1958;258: 1123–1142. 3. Ioachimescu OC, Kavuru MS. Pulmonary alveolar proteinosis. Chron Respir Dis 2006;3:149–159. 4. Presneill JJ, Nakata K, Inoue Y, Seymour JF. Pulmonary alveolar proteinosis. Clin Chest Med 2004; 25:593–613.
RG ■ Volume 28 • Number 3 5. Lin FC, Chang GD, Chern MS, Chen YC, Chang SC. Clinical significance of anti-GM-CSF antibodies in idiopathic pulmonary alveolar proteinosis. Thorax 2006;61:528–534. 6. Uchida K, Beck DC, Yamamoto T, et al. GM-CSF autoantibodies and neutrophil dysfunction in pulmonary alveolar proteinosis. N Engl J Med 2007; 356:567–579. 7. Yusen RD, Cohen AH, Hamvas A. Normal lung function in subjects heterozygous for surfactant protein-B deficiency. Am J Respir Crit Care Med 1999;159:411–414. 8. Doerschuk CM. Pulmonary alveolar proteinosis: is host defense awry? N Engl J Med 2007;356: 547–549. 9. Seymour JF, Presneill JJ. Pulmonary alveolar proteinosis: progress in the first 44 years. Am J Respir Crit Care Med 2002;166:215–235. 10. Bonfield TL, Russell D, Burgess S, Malur A, Kavuru MS, Thomassen MJ. Autoantibodies against granulocyte macrophage colony-stimulating factor are diagnostic for pulmonary alveolar proteinosis. Am J Respir Cell Mol Biol 2002;27:481–486. 11. Bonfield TL, Raychaudhuri B, Malur A, et al. PU.1 regulation of human alveolar macrophage differentiation requires granulocyte-macrophage colonystimulating factor. Am J Physiol Lung Cell Mol Physiol 2003;285:L1132–L1136. 12. Burkhalter A, Silverman JF, Hopkins MB 3rd, Geisinger KR. Bronchoalveolar lavage cytology in pulmonary alveolar proteinosis. Am J Clin Pathol 1996;106:504–510. 13. Godwin JD, Muller NL, Takasugi JE. Pulmonary alveolar proteinosis: CT findings. Radiology 1988; 169:609–613. 14. Burbank B, Morrione TG, Cutler SS. Pulmonary alveolar proteinosis and nocardiosis. Am J Med 1960;28:1002–1007. 15. Andriole VT, Ballas M, Wilson GL. The association of nocardiosis and pulmonary alveolar proteinosis: a case study. Ann Intern Med 1964;60:266–275. 16. Murch CR, Carr DH. Computed tomography appearances of pulmonary alveolar proteinosis. Clin Radiol 1989;40:240–243. 17. McCook TA, Kirks DR, Merten DF, Osborne DR, Spock A, Pratt PC. Pulmonary alveolar proteinosis in children. AJR Am J Roentgenol 1981;137: 1023–1027. 18. Miller PA, Ravin CE, Walker Smith GJ, Osborne DR. Pulmonary alveolar proteinosis with interstitial involvement. AJR Am J Roentgenol 1981;137: 1069–1071. 19. Lee KN, Levin DL, Webb WR, Chen D, Storto ML, Golden JA. Pulmonary alveolar proteinosis: high-resolution CT, chest radiographic, and functional correlations. Chest 1997;111:989–995. 20. Newell JD, Underwood GH Jr, Russo DJ, Bruno PP, Wilkerson GR, Black ML. Computed tomo-
Frazier et al 899 graphic appearance of pulmonary alveolar proteinosis in adults. J Comput Tomogr 1984;8:21–29. 21. Prakash UB, Barham SS, Carpenter HA, Dines DE, Marsh HM. Pulmonary alveolar phospholipoproteinosis: experience with 34 cases and a review. Mayo Clin Proc 1987;62:499–518. 22. Holbert JM, Costello P, Li W, Hoffman RM, Rogers RM. CT features of pulmonary alveolar proteinosis. AJR Am J Roentgenol 2001;176:1287–1294. 23. Jones CC. Pulmonary alveolar proteinosis with unusual complicating infections: a report of two cases. Am J Med 1960;29:713–722. 24. Kang EY, Grenier P, Laurent F, Muller NL. Interlobular septal thickening: patterns at high-resolution computed tomography. J Thorac Imaging 1996;11:260–264. 25. Johkoh T, Itoh H, Muller NL, et al. Crazy-paving appearance at thin-section CT: spectrum of disease and pathologic findings. Radiology 1999;211: 155–160. 26. Lee CH. The crazy-paving sign. Radiology 2007; 243:905–906. 27. Murayama S, Murakami J, Yabuuchi H, Soeda H, Masuda K. “Crazy paving appearance” on high resolution CT in various diseases. J Comput Assist Tomogr 1999;23:749–752. 28. Webb WR. Thin-section CT of the secondary pulmonary lobule: anatomy and the image—the 2004 Fleischner lecture. Radiology 2006;239:322–338. 29. Rossi SE, Erasmus JJ, Volpacchio M, Franquet T, Castiglioni T, McAdams HP. “Crazy-paving” pattern at thin-section CT of the lungs: radiologic-pathologic overview. RadioGraphics 2003;23:1509–1519. 30. Franquet T, Giménez A, Rosón N, Torrubia S, Sabaté JM, Pérez C. Aspiration diseases: findings, pitfalls, and differential diagnosis. RadioGraphics 2000;20:673–685. 31. Miller WT, Shah RM. Isolated diffuse ground-glass opacity in thoracic CT: causes and clinical presentations. AJR Am J Roentgenol 2005;184:613–622. 32. Shah RM, Miller W Jr. Widespread ground-glass opacity of the lung in consecutive patients undergoing CT: does lobular distribution assist diagnosis? AJR Am J Roentgenol 2003;180:965–968. 33. Uchida K, Nakata K, Trapnell BC, et al. High-affinity autoantibodies specifically eliminate granulocyte-macrophage colony-stimulating factor activity in the lungs of patients with idiopathic pulmonary alveolar proteinosis. Blood 2004;103:1089–1098. 34. Parker LA, Novotny DB. Recurrent alveolar proteinosis following double lung transplantation. Chest 1997;111:1457–1458. 35. Gal AA, Bryan JA, Kanter KR, Lawrence EC. Cytopathology of pulmonary alveolar proteinosis complicating lung transplantation. J Heart Lung Transplant 2004;23:135–138.
This article meets the criteria for 1.0 credit hour in category 1 of the AMA Physician’s Recognition Award. To obtain credit, see accompanying test at http://www.rsna.org/education/rg_cme.html.
RG ■ Volume 28 • Number 3 • May-June 2008
Frazier et al
Pulmonary Alveolar Proteinosis Aletta Ann Frazier, MD, et al RadioGraphics 2008; 28:883–899 • Published online 10.1148/rg.283075219 • Content Code:
Page 884 Three distinct subgroups of PAP are currently recognized: idiopathic, secondary, and congenital.... Secondary PAP (5%–10% of cases) is recognized in patients with industrial inhalational exposure to materials such as silica particles, cement dust, aluminum dust, titanium dioxide, nitrogen dioxide, and fiberglass; underlying hematologic malignancy; or immunodeficiency disorders (including cytotoxic or immunosuppressive therapy and human immunodeficiency virus infection) (1,3).
Page 887 The CT appearance of “crazy-paving,” defined as a network of smoothly thickened reticular (septal) lines superimposed on areas of ground-glass opacity, was first described in a group of six patients with PAP (Fig 5) (16,24,25). Areas of crazy-paving in PAP are typically widespread and bilateral, often with sharply marginated areas of geographic or lobular sparing (Fig 6) (13,16,22).
Page 888 Although the CT finding of crazy-paving is highly characteristic of PAP, this pattern is also seen in several infectious, hemorrhagic, neoplastic, inhalational, and idiopathic conditions as well as in straightforward hydrostatic pulmonary edema. Therefore, the radiologic differential diagnosis of crazy-paving is broad and includes left heart failure, pneumonia (especially pneumocystis pneumonia), alveolar hemorrhage, bronchoalveolar carcinoma, lymphangitic carcinomatosis, diffuse alveolar damage (adult respiratory distress syndrome), radiation- or drug-induced pneumonitis, hypersensitivity pneumonitis, and pulmonary venoocclusive disease (Fig 8) (26–32).
Page 893 At microscopic examination, PAP is characterized by intraalveolar accumulations of lipoproteinaceous material circumscribed by normal or thickened alveolar walls (Fig 10a). The accumulations are finely granular eosinophilic deposits containing acicular clefts of dissolved cholesterol, eosinophilic globules, foamy macrophages, and ghostlike cell remnants (Fig 10b).
Page 898 Prominent septal lines in PAP correlate with remarkable ILS expansion due to edema and dilatation of the pulmonary lymphatic system, a strong sign of disturbed pulmonary fluid homeostasis.
