|
 |
|
|||||||
|
|||||||||
Copyright ©ERS Journals Ltd 2001
Classifying interstitial lung diseases in a fractal lung: a morphologist's view "anno Domini 2000"
Dept of Pathology, Katholieke Universiteit Leuven, Leuven, Belgium
CORRESPONDENCE: E.K. Verbeken, Dept of Pathology, U.Z. St. Rafaël, Minderbroedersstraat 12, 3000, Leuven, Belgium. Fax: 32 16336548
Keywords: classification, fractals, interstitial lung disease, morphology, pathology
 |
Abstract |
|---|
 TOP Abstract The pathological basis for...The pathological basis for...Bronchiolitis in interstitial...References |
|---|
Interstitial lung diseases (ILDs) remain a challenging problem for the pathologist. New insights in aetiology and pathogenesis, new diagnostic tools and successful research have led to a renewed interest in ILDs during the last few years, and highlighted the need for a novel classification, particularly of the chronic and/or idiopathic categories of interstitial pneumonias.
The present paper compares the terminology of the latter categories in current and previous classifications and briefly discusses the pathological basis for the classifications of ILDs in general, and for the idiopathic interstitial pneumonias (IIPs) in particular. The difference between high versus low morphological specificity determines the pathological classifications. The classification of IIPs relies upon a pattern recognition taking temporal and spatial distribution into consideration.
The last section of this paper discusses recent research opposing the conventional pathological approach, analogous to the mechanical two-compartment model of the lung, in which a discontinuity is considered between these two compartments, and thus, a distinction is made between interstitial lung diseases with and without bronchiolitis. In the recent "fractal" concept, the continuity of the lung architecture is emphasized: the lung is a so-called fractal tree with noninteger dimensions. In this fractal model, an interstitial lung disease effects a peripheral part of the pulmonary fractal tree and this may or may not include bronchioles.
Chronic interstitial pneumonias were originally divided into five categories by Liebow and Carrington 1. These categories and their current designations are shown in table 1
. Table 2 shows the classification of idiopathic interstitial pneumonias according to Katzenstein and Myers 2 in 1998, and according to the proposals of the American Thoracic Society (ATS) and the European Respiratory Society (ERS). The clinical and pathological definitions of idiopathic pulmonary fibrosis/cryptogenic fibrosing alveolitis have recently been reported 3.
|
|
As recently defined, usual interstitial pneumonia (UIP) is the histological pattern that is associated with idiopathic pulmonary fibrosis (IPF), but UIP may also be seen in conditions other than IPF. The morphological definition of UIP has not been modified since its original description and is basically characterized by temporal and spatial inhomogeneity of the interstitial inflammation and fibrosis, and by fibroblast foci, in the absence of histological features that point to a specific diagnosis 4.
Desquamative interstitial pneumonia (DIP) is characterized by panlobular involvement and diffuse mild-to-moderate interstitial fibrosis, homogenous type II cell hyperplasia and massive alveolar filling with macrophages. Respiratory bronchiolitis-associated interstitial lung disease (RBILD) represents an exaggerated reaction to smoking with an accumulation of macrophages, which is, in contrast to DIP, confined to the centrilobular area around respiratory bronchioles. The condition is also associated with only mild inflammatory fibrotic and metaplastic changes. The morphology of RBILD overlaps with DIP. Respiratory bronchioles (RBs) refer to the smoker's bronchiolitis, as described in 1974 by Niewoehner et al. 5. The accompanying interstitial lung disease (ILD) is mild and has a good prognosis 6. However, RBILD associated with severe dyspnoea, exertional hypoxaemia and clubbing has recently been reported in a few patients 7. Overlapping histomorphology between RBILD and DIP suggests that both entities may represent different ends of the spectrum of the same disease, but this is still controversial. On high-resolution computed tomography (HRCT), there is an overlap between the findings of respiratory bronchiolitis, RBILD and DIP. This led Heyneman et al. 8 to conclude that these entities represent different degrees of severity of small airway and parenchymal reaction to cigarette smoke. Consequently, it has been suggested that DIP is replaced by RBILD because it is anatomically more correct, and because it emphasizes the putative role of cigarette smoking or other environmental exposures in the pathogenesis of this condition 2. However, DIP also occurs in nonsmokers. Of the 40 cases reported by Carrington et al. 4, 10% were nonsmokers. RBILD has also been reported in a nonsmoking patient with an occupational exposure to fumes of solder flux 9. Furthermore, DIP appears not to be uncommon in children with ILD 10. For these reasons, it would seem sensible to retain both terms.
Bronchiolitis obliterans with interstitial pneumonia (BIP) has been classified as cryptogenic organizing pneumonia (COP, synonymous with idiopathic bronchiolitis obliterans with organizing pneumonia (BOOP)). In contrast to BIP, COP was immediately and widely accepted. The term BIP focused on the inflammatory obliteration of bronchioles, associated with a variable mixture of ILD; both chronic lesions of UIP and more acute lesions of diffuse alveolar damage (DAD).
Acute interstitial pneumonia (AIP) 11 and nonspecific interstitial pneumonia (NSIP) 12 have recently been included among the IIP's. AIP is a very rapidly evolving disease with the histological appearance of DAD. Classically, the lesions of DAD are morphologically divided into an exudative phase that ends with the formation of the characteristic hyaline membranes along respiratory bronchioles and alveolar ducts, and the fibroproliferative phase, characterized by fibroblast proliferation, collagen deposition and resolution of the hyaline membranes. The fibroproliferative phase may start as early as 3–4 days after the onset, but, in clinical practice, AIP may develop over a wider time span, ranging from a few days to several months. By the time open lung biopsy is performed, fibroproliferative lesions are almost invariably present. AIP is, by definition, an idiopathic entity. Hence, the diagnosis excludes the numerous other causes of DAD, namely infections, inhalants, ingested agents, including drugs, shock, radiation, and miscellaneous other conditions.
NSIP represents a heterogeneous group of interstitial pneumonias that do not fit into the other entities (i.e. AIP, UIP, DIP or COP). This allows pathologists to limit the diagnosis of UIP to its original morphological description; consequently, they are no longer forced to use UIP as a "wastebasket" for various types of fibrosing ILD. To a certain point, NSIP has now become the new wastebasket, although it has distinctive features that distinguish it from UIP: temporal homogeneity, whether cellular or fibrotic, and few fibroblast foci.
It is important to emphasize that the term "idiopathic" cannot be used by the pathologist. Hence, UIP is a pathological pattern, but not diagnostic of IPF. Similarly, the pathological patterns DAD and organizing pneumonia (OP), are not synonymous with the idiopathic entities of AIP and COP.
An ATS/ERS committee has been given the task of aligning the most recent histopathological definitions with idiopathic clinical disease diagnosis.
Lymphoid (or lymphocytic) interstitial pneumonitis (LIP) is a diffuse lymphocytic interstitial infiltrate. This disorder falls under the heading of lymphoproliferative diseases, which have been the subject of considerable debate as to whether they are reactive or neoplastic. Most experts no longer include LIP among the IIPs because of its tendency to progress to lymphoma and because it is documented in association with collagen vascular diseases and immunodeficiency states, including acquired immune deficiency syndrome (AIDS). If immunohistochemistry or gene amplification reveals monoclonality, the diagnosis of lymphoma seems to be justified. There is still much debate about whether pseudolymphoma should be regarded as a premalignant lesion or part of a spectrum of low-grade malignant lymphoma. Pulmonary pseudolymphoma is a primary, usually solitary, lymphoid mass, and the majority of patients are asymptomatic. In most cases, no recurrence is observed after surgical removal of the mass.
Giant cell interstitial pneumonia (GIP) is related to cobalt sensitivity, especially in hard metal 13, and is no longer considered idiopathic.
|
The pathological basis for the classification of interstitial lung disease: high versus low morphological specificity |
|---|
|
TOPAbstractThe pathological basis for...The pathological basis for...Bronchiolitis in interstitial...References |
|---|
Table 3
schematizes the pathological basis for the classification of ILDs. The entities with morphological features characteristic of the disease have a high morphologic specificity. Entities that lack typical histology have a low morphological specificity. Morphological specificity is not related to a knowledge of the cause of the disease. The aetiologies of Langerhans' cell granulomatosis, lymphangioleiomyomatosis or alveolar lipoproteinosis remain enigmatic, but these lesions have a very high morphological specificity. It must be emphasized that highly specific lesions, such as an eosinophilic granuloma, may lose their morphological specificity during corticosteroid therapy or after healing. Alternatively, some pneumoconioses, such as asbestosis, have a low morphological specificity; for the pathologist, the aetiological diagnosis relies entirely on finding a sufficient number of asbestos bodies in the biopsy.
|
Entities with high morphological specificity are exclusively classified according to histopathological criteria. Their diagnosis is less subject to clinicopathological correlation. Entities with low morphological specificity of unknown cause should be classified clinicopathologically. The final diagnosis thus relies on clinical and pathological features. Without exception, the IIPs listed in table 2
belong to this category. |
The pathological basis for the classification of idiopathic interstitial pneumonias |
|---|
|
TOPAbstractThe pathological basis for...The pathological basis for...Bronchiolitis in interstitial...References |
|---|
Diagnosis and classification rely upon a pattern recognition, according to which the pathologist tries to reconstruct the disease process within time and space. The corresponding descriptive entities should include two parameters: temporal distribution (heterogeneity versus homogeneity) and spatial distribution. Examples are shown in table 4
.
|
Reconstruction in time
DAD/AIP and DIP are temporally homogenous and are morphologically distinct from UIP and NSIP because of the associated alveolar filling in DIP and the DAD characteristics in AIP.
UIP is temporally heterogeneous. In contrast, NSIP is temporally homogeneous 12 and the changes appear to occur over a single, relatively narrow time span. NSIP, therefore, describes a disorder that presents morphologically as a kind of cellular interstitial pneumonia (CIP), characterized by widespread interstitial infiltrates of lymphocytes and usually numerous plasmocytes, as are seen in diseases such as polymyositis/dermatomyositis. The term "lymphoplasmocytic interstitial pneumonitis" 14 has also been used for these conditions.
Originally, NSIP was considered to be a wastebasket term for all idiopathic non-UIP, non-DIP/RBILD and non-AIP ILD 12. Now, however, the concept of NSIP has evolved from a broad all-inclusive term into a specific form of IIP with clear histopathological features. Patients with NSIP, at least the cellular type, tend to fare better than patients with UIP or AIP. Temporal homogeneity is the dominant feature 11, 15. The diagnosis of fibrotic NSIP, however, should only be made when there is a diffuse, uniform interstitial process in which chronic inflammation is mixed with fibrosis. Regardless of the pattern of involvement, and whatever the disease entity concerned, the inflammatory activity and the amount of collagenous fibrosis are the predominant prognostic indicators for reversibility and residual lung function.
Reconstruction in space
At the acinar level, the light microscopic level, and with exclusion of the vasculitis syndromes, only four basic acinar distribution patterns can be recognized for ILD, whether idiopathic or not: centriacinar (fig. 1), lymphatic tracking (fig. 2), random (fig. 3), and panacinar (fig. 4). Spatial distribution patterns of the IIPs are presented in table 4. The figures illustrating these patterns have been specifically chosen in honour of A. Liebow: typical examples were chosen from his original classification, as listed in table 1 1.
|
|
|
|
The terms "diffuse" and "patchy" should be avoided as they are nonspecific. At the acinar level, only the panacinar pattern is diffuse. At the interacinar level, use of the term "uniform" for conditions in which all neighbouring acini are affected by the same acinar pattern is recommended. If not, then the interacinar pattern is "not uniform".
It must be emphasized that temporal inhomogeneity may result in different basic acinar patterns between neighbouring acini, due either to remodelling (such as loss of units or acini in parallel) or to more severe, more fibrotic, or even end-stage honeycombing in some areas. This may blur the original basic acinar pattern, thus rendering the reconstruction of the disease more difficult or even impossible in end-stage disease.
As seen in table 4, DAD/AIP and DIP have a panacinar spatial pattern. In contrast, RBILD and BOOP/OP are centriacinar.
Reconstruction of NSIP within space is not unambiguous. From the 64 patients originally described by Katzenstein and Fiorelli 12, 45% (n=29) had a diffuse pattern and 55% (n=35) had a "patchy" distribution of the lesions, 28% (n=18) of which had bronchiolocentric accentuation of the infiltrate. Today, the concept of NSIP has evolved to a "diffuse" disease 2. Apparently, the distribution pattern in general, and for NSIP in particular, is not as relevant for prognosis, but it is with regard to the aetiopathogenesis. The lack of a specific pattern is in keeping with the wide variety of clinical conditions NSIP is associated with 12, including various connective tissue diseases, toxic reactions and hypersensitivity pneumonitis. Further, it follows from the lack of a specific NSIP distribution pattern that fibrotic stages of the latter will be difficult to separate from UIP. However, the fibrotic variant in NSIP is associated with a much worse prognosis than the cellular variant 15, thus reducing the clinical relevance of the differential diagnosis from UIP.
|
Bronchiolitis in interstitial lung disease: a fractal approach |
|---|
|
TOPAbstractThe pathological basis for...The pathological basis for...Bronchiolitis in interstitial...References |
|---|
In a conventional approach, and analogous to the mechanical two-compartment model of the lung, i.e. a resistive and an elastic element in series 16, 17, the lung has been divided morphologically into an air conducting system, the bronchial tree, and a gas exchanging system, the alveolated parenchyma. This conventional approach emphasizes a discontinuity between the two compartments. Furthermore, by light microscopy, and thus in a two-dimensional (2D) histological plane, the conducting airways have distinct morphological features and are easily recognized as a dichomotomous branching system. In contrast, the alveolated parenchyma appears "amorphous" on transsection, as the internal organization of this compartment cannot be interpreted by simple visual inspection.
From this conventional approach, it logically follows that there are ILDs (for practical purposes here defined as affecting the parenchyma) with and without bronchiolitis. This can be seen in previous and current classifications of ILD and IIP, such as the BIP originally described by Liebow and Carrington 1, BOOP/OP, and RBILD.
Recent research work, however, favours a more modern approach to the lung architecture, which stresses the continuity, not only of the gross "anatomical" compartments, such as cartilagenous bronchi, noncartilageneous membranous bronchioles and alveolated parenchyma, but also, more subtly, of the various tissue components, particularly the force-bearing fibrous system of the lung 18. This approach conceptualizes the lung as a so-called "fractal tree".
Much can be said about fractal design. Fractal geometry deals with the geometry of hierarchies and such structures are too complex to be described in terms of classic three-dimensional Euclidean geometry. Therefore, they appear "amorphous" on 2D transsection. Mandelbrot 19 dealt with this problem by using "fractal" dimensions, i.e. noninteger dimensions, from which the name is derived.
Unique to the fractal tree is the self-similarity of each following order and the absence of a characteristic scale. In keeping with this design, in the lung, each further airway generation supplies a corresponding parenchymal unit. Some of these units have an anatomical name, such as the "lung" or the "lung lobe", but also the "segment", the "secondary lobule" and the "acinus" (in fact, only the first two can be recognized anatomically because they are more or less enwrapped by a septum, the pleura).
In terms of fractal design, however, there is no difference between a parenchymal unit supplied by a central cartilageneous bronchus or by a peripheral bifurcating alveolar duct. Thus, in this model, a parenchymal unit of the lung supplied by each further hierarchical airway order is architecturally identical to the preceding one, smaller than it, and located within it.
Further, the human lung is characterized by a gradual transition of the morphological features characterizing cartilagenous bronchi, membranous bronchioles (MBs) and alveolated air passages. Indeed, a gradual loss of cartilage and a progressive increase of the number of alveoli is observed along the fractal tree 20. By definition, membranous bronchioles are airways that do not bear alveoli or cartilage in their wall. There are, on average, two to three generations of MBs in a normal lung. However, it has been shown that due to the gradually changing morphology, at least five to six generations of airways figure as "membranous" bronchioles in a histological section 20. This may represent one-quarter to one-half of all airway generations, depending on the pathway. Consequently, bronchiolitis may indicate proximal disease, close to bronchi, and distal disease, which is also close to the alveolar ducts. This cannot be judged from 2D histological sections.
In rodents and sheep the gradual alveolization is absent or very limited. Further, the acinar airway tree has fewer generations of alveolar ducts. Although this does not change the fractal design pattern, the distance between alveolar lumen and bronchiolar mucociliary escalator is shorter than in humans.
The definition of an ILD in a fractal tree model is different from the definition in a conventional model. In the fractal model, an ILD affects a peripheral part of the pulmonary fractal tree. The affected segments along the tree may or may not include bronchioles. Also, the pathological changes do not necessarily extend towards the most peripheral ramifications in the alveolated parenchyma. In the latter condition, the acinar periphery is spared, and consequently, the pattern is basically centriacinar.
In summary, in the fractal approach, an ILD without bronchiolitis corresponds to a "distal ILD", whereas bronchiolar involvement points to a "proximal ILD", or at least, a proximal extension of the disease.
It is not clear whether proximal and distal ILD should a priori be considered different entities. Smoker's (respiratory) bronchiolitis, RBILD, and DIP are considered different spectra of the same disease. If so, this suggests that the more distal the disease is, the more panacinar the pattern will be. Also, follicular bronchiolitis (FB) and LIP have been regarded as part of an overlapping spectrum of reactive lymphoid hyperplasia 21. The entity described as either COP or idiopathic BOOP/OP can be explained by more proximal or distal involvement of the fractal tree by the same disease. BIP originally described by Liebow and Carrington 1, should be regarded as a "proximal ILD of whatever (histological) category".
More illustratively, proximal and distal presentations can also be documented in ILDs of known causes. Asbestosis will now be considered. By definition, asbestosis refers to a bilateral diffuse interstitial fibrosis of the lungs caused by the inhalation of asbestos fibres. This is considered to be the only condition that should be labelled asbestosis 22, and this has medico-legal implications. However, it has been documented that workers with heavy occupational asbestos exposure may manifest severe and widespread airway lesions without evidence of diffuse fibrosis 23. According to the conventional two-compartment model of the lung, asbestos-induced airways disease may be regarded as a separate entity 24. Conversely, in a fractal approach, they represent the proximal and distal ends of the spectrum along the fractal tree.
Hypersensitiviy reactions to inhaled antigens, such as extrinsic allergic alveolitis (EAA), are another example of both proximal and distal presentations. Basically, EAA is characterized by a lymphohistiocytic interstitial infiltrate together with poorly formed epithelioid cell granulomas. The topographical distribution is in a spectrum between exclusively centrilobular involvement of the walls of respiratory brochioles, hence the term "microgranulomatous bronchiolitis" as a descriptor for these conditions 14, and panlobular involvement. Moreover, and not infrequently, EAA shares morphological features of bronchiolitis obliterans and even organizing pneumonia, thus completing the entire spectrum along the distal fractal tree.
Why should the same disease present either as a proximal bronchiolar disease or as a distal alveolar disease? With regard to inhalation pathology, experimental studies of humans inhaling inert dust suggest that marked variations in individual rates of particle clearance rely primarily on individual differences in lung structure. The latter differences are only poorly, if at all, studied. Basic research is required in order to try to understand the morphology of lung function in order to increase the accuracy of clinicopathological correlations.
|
References |
|---|
|
TOPAbstractThe pathological basis for...The pathological basis for...Bronchiolitis in interstitial...References |
|---|
- Liebow AA, Carrington CB. The interstitial pneumonias. In: Simon M, Potchen EJ, LeMay M, editors. Frontiers of Pulmonary Radiology 1. New York, Grune and Stratton, 1969; pp. 102––141.
- Katzenstein AA, Myers JL. Idiopathic pulmonary fibrosis. Clinical relevance of pathologic classification. Am J Respir Crit Care Med 1998;157:1301–1315.
[Free Full Text] - King TE, Costabel U, Cordier J-F, et al. Idiopathic pulmonary fibrosis: diagnosis and treatment. Am J Respir Crit Care Med 2000;161:646–664.
[Free Full Text] - Carrington CB, Gaensler EA, Coutu RE, et al. Natural history and treated course of usual and desquamative interstitial pneumonia. N Engl J Med 1978;298:801–809.[Abstract]
- Niewoehner DE, Kleinerman J, Rice DB. Pathologic changes in the peripheral airways of young cigarette smokers. N Engl J Med 1974;291:755–758.[Web of Science][Medline] [Order article via Infotrieve]
- King TE. Respiratory bronchiolitis associated interstitial lung disease. Clin Chest Med 1993;14:693–698.[Web of Science][Medline] [Order article via Infotrieve]
- Sodikot RT, Johnson J, Loyd JE, Christman JW. Respiratory bronchiolitis associated with severe dyspnea, exertional hypoxemia, and clubbing. Chest 2000;117:282–285.
[Abstract/Free Full Text] - Heyneman LE, Ward S, Lynch DA, Remy-Jardin M, Johkoh T, Müller NL. Respiratory bronchiolitis, respiratory bronchiolitis-associated interstitial lung disease, and desquamative interstitial pneumonia: different entities for part of the spectrum of the same disease process? Am J Radiol 1999;173:1617–1622.[Abstract]
- Moon J, du Bois RM, Colby TV, Hansell DM, Nicholson AG. Clinical significance of respiratory bronchioles on open lung biopsy and its relationship to smoking related interstitial lung disease. Thorax 1999;54:1009–1014.
[Abstract/Free Full Text] - Nicholson AG, Kim H, Corrin B, et al. The value of classifying interstitial pneumonitis in childhood according to defined histological patterns. Histopathology 1998;33:203–211.[CrossRef][Web of Science][Medline] [Order article via Infotrieve]
- Katzenstein AA, Meyers JL, Mazur MT. Acute interstitial pneumonia. A clinicopathologic, ultrastructural, and cell kinetic study. Am J Surg Pathol 1986;10:256–267.[Web of Science][Medline] [Order article via Infotrieve]
- Katzenstein AA, Fiorelli RF. Nonspecific interstitial pneumonia/fibrosis. Am J Surg Pathol 1994;18:136–147.[Web of Science][Medline] [Order article via Infotrieve]
- Demedts M, Gheysens B, Nagels J, et al. Cobalt lung in diamant polishers. Am Rev Respir Dis 1984;130:130–135.[Web of Science][Medline] [Order article via Infotrieve]
- Mark EJ. Interstitial disease. In: Lung biopsy interpretation. Baltimore, Williams & Wilkins, 1984; pp. 113––130.
- Travis WD, Matsui K, Moss J, Ferrans JV. Idiopathic nonspecific interstitial pneumonia: prognostic significance of cellular and fibrosing patterns. Am J Surg Pathol 2000;24:19–33.[CrossRef][Web of Science][Medline] [Order article via Infotrieve]
- Mead J. Contribution of compliance of airways to frequency-dependent behaviour of lungs. J Appl Physiol 1969;26:670–673.
[Free Full Text] - Otis AB, McKerrow CB, Bartlett RA, et al. Mechanical factors in distributions of pulmonary ventilation. J Appl Physiol 1956;48:427–447.
- Verbeken EK, Cauberghs M, Van de Woestijne K. Membranous bronchioles and connective tissue network of normal and emphysematous lungs. J Appl Physiol 1996;81:2468–2480.
[Abstract/Free Full Text] - Mandelbrot BB. The fractal geometry of nature. New York, Freeman, 1983.
- Verbeken EK, Cauberghs M, Lauweryns JM, Van de Woestijne K. Anatomy of membranous bronchioles in normal, senile, and emphysematous human lungs. J Appl Physiol 1992;72:2343–2353.
[Abstract/Free Full Text] - Nicholson AG, Wotherspoon AC, Dis TC, et al. Reactive pulmonary lymphoid disorders. Histopathology 1995;26:405–412.[Web of Science][Medline] [Order article via Infotrieve]
- American Thoracic Society. Diagnosis of nonmalignant diseases related to asbestos. Am Rev Respir Dis 1986;134:363–368.[Web of Science][Medline] [Order article via Infotrieve]
- Wright JL, Churg A. Severe diffuse small airways disease in long term chrysotile miners. Br J Indust Med 1985;42:556–559.[Web of Science][Medline] [Order article via Infotrieve]
- Churg A. Nonneoplastic disease caused by asbestos. In: Churg A, Green FH, editors. Pathology of occupational lung disease. Baltimore, Williams & Wilkins, 1998; pp. 277––338.
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |