AMERICAN THORACIC SOCIETY
American Thoracic Society/European Respiratory Society International Multidisciplinary Consensus Classification of the Idiopathic Interstitial Pneumonias
This Joint Statement of the American Thoracic Society (ATS), and the European Respiratory Society (ERS) was adopted by the ATS Board of Directors, June 2001 and by The ERS Executive Committee, June 2001

	CONTENTS

TOP CONTENTS EXECUTIVE SUMMARY INTRODUCTION RATIONALE FOR A CHANGE... DEVELOPMENT OF A NEW... PRINCIPLES GUIDING THE... IDIOPATHIC PULMONARY FIBROSIS NONSPECIFIC INTERSTITIAL... CRYPTOGENIC ORGANIZING... ACUTE INTERSTITIAL PNEUMONIA RESPIRATORY BRONCHIOLITIS-... DESQUAMATIVE INTERSTITIAL... LYMPHOID INTERSTITIAL PNEUMONIA REFERENCES

Executive Summary

    Objectives

    Participants

    Evidence

    Validation

    Key Messages

Introduction

Rationale for a Change in the Approach to Classification of Idiopathic Interstitial Pneumonias

Development of a New Classification of Idiopathic Interstitial Pneumonia

    Current Classification of IIP

    New ATS/ERS Classification

Principles Guiding the Assessment of Patients with Idiopathic Interstitial Pneumonias

    The Diagnostic Process Is Dynamic

    Clinical Evaluation

    Radiological Evaluation

    Role of Surgical Lung Biopsy

    Unclassifiable Interstitial Pneumonia

    Bronchoalveolar Lavage Fluid Evaluation

Idiopathic Pulmonary Fibrosis

    Clinical Features

    Radiologic Features

    Histologic Features

    IPF: Areas of Uncertainty

Nonspecific Interstitial Pneumonia

    Clinical Features

    Radiologic Features

    Histologic Features

    NSIP: Areas of Uncertainty

Cryptogenic Organizing Pneumonia

    Clinical Features

    Radiologic Features

    Histologic Features

    COP: Areas of Uncertainty

Acute Interstitial Pneumonia

    Clinical Features

    Radiologic Features

    Histologic Features

    AIP: Areas of Uncertainty

Respiratory Bronchiolitis-Associated Interstitial Lung Disease

    Clinical Features

    Radiologic Features

    Histologic Features

    RB-ILD: Areas of Uncertainty

Desquamative Interstitial Pneumonia

    Clinical Features

    Radiologic Features

    Histologic Features

    DIP: Areas of Uncertainty

Lymphoid Interstitial Pneumonia

    Clinical Features

    Radiologic Features

    Histologic Features

    LIP: Areas of Uncertainty

References

Appendix

Despite the considerable progress in the classification of the idiopathic interstitial pneumonias (IIPs), the lack of an international standard has resulted in variable and confusing diagnostic criteria and terminology. The advent of high-resolution computerized tomography, the narrowed pathologic definition of usual interstitial pneumonia (UIP) and recognition of the prognostic importance of separating UIP from other IIP patterns have profoundly changed the approach to the IIPs. This is an international Consensus Statement defining the clinical manifestations, pathology, and radiologic features of patients with IIP. The major objectives of this statement are to standardize the classification of the idiopathic interstitial pneumonias (IIPs) and to establish a uniform set of definitions and criteria for the diagnosis of IIPs. The targeted specialties are pulmonologists, radiologists, and pathologists. A multidisciplinary core panel was responsible for review of background articles and writing of the document. In addition, this group reviewed the clinical, radiologic, and pathologic aspects of a wide spectrum of cases of diffuse parenchymal interstitial lung diseases to establish a uniform and consistent approach to these diseases and to clarify the terminology, definitions, and descriptions used in routine clinical practice. The final statement was drafted after a series of meetings of the entire committee. The level of evidence for the recommendations made in this statement is largely that of expert opinion developed by consensus. This classification of IIPs includes seven clinico-radiologic-pathologic entities: idiopathic pulmonary fibrosis (IPF), nonspecific interstitial pneumonia, cryptogenic organizing pneumonia, acute interstitial pneumonia, respiratory bronchiolitis-associated interstitial lung disease, desquamative interstitial pneumonia, and lymphoid interstitial pneumonia. The need for dynamic interaction between pathologists, radiologists, and pulmonologists to accurately diagnose these disorders is emphasized. The level of evidence for the recommendations made in this Statement is largely that of expert opinion developed by consensus. This Statement is an integrated clinical, radiologic, and pathologic approach to the classification of the IIPs. Use of this international multidisciplinary classification will provide a standardized nomenclature and diagnostic criteria for IIP. This Statement provides a framework for the future study of these entities.

	EXECUTIVE SUMMARY

TOP CONTENTS EXECUTIVE SUMMARY INTRODUCTION RATIONALE FOR A CHANGE... DEVELOPMENT OF A NEW... PRINCIPLES GUIDING THE... IDIOPATHIC PULMONARY FIBROSIS NONSPECIFIC INTERSTITIAL... CRYPTOGENIC ORGANIZING... ACUTE INTERSTITIAL PNEUMONIA RESPIRATORY BRONCHIOLITIS-... DESQUAMATIVE INTERSTITIAL... LYMPHOID INTERSTITIAL PNEUMONIA REFERENCES

The American Thoracic Society and European Respiratory Society sponsored this project to standardize classification of the idiopathic interstitial pneumonias (IIPs), a subset of acute and chronic lung disorders collectively referred to as interstitial lung diseases or diffuse parenchymal lung diseases of unknown etiology. Major progress has been made in our understanding of the clinical, radiological, and pathological manifestations of these disorders. Consequently it was felt that an international multidisciplinary Consensus Statement was needed to establish a uniform set of definitions and criteria for the diagnosis of IIPs.

Objectives

This is an international Consensus Statement defining the clinical manifestations, pathology, and radiologic features of patients with IIP. This Statement has been produced as a collaborative effort from the American Thoracic Society (ATS) and the European Respiratory Society (ERS). The purpose of this Consensus Statement is to provide an integrated clinical, radiologic, and pathologic approach to classification of the clinicopathological entities within the IIP group that has bearing on clinical course and prognosis. The targeted specialties are pulmonologists, radiologists, and pathologists.

Participants

Panel members are expert clinicians, radiologists, and pathologists in adult pulmonary diseases. The supporting associations nominated panel members. The co-chairs were selected by the ATS. Panel members were selected because of special interest and expertise in diffuse parenchymal lung disease and to provide an international range of expertise. The panel was divided into a core group and a reviewer group (see Appendix).

Evidence

The core group was responsible for review of background articles that discussed the existing scientific evidence. Relevant articles from the medical literature were identified by a MedLine search (1966 to December 1998) of English language articles or articles with English abstracts, the bibliographies of the articles retrieved, and the committee members' files.

In addition, this group reviewed the clinical history, chest-imaging studies including chest radiograph and high-resolution computerized tomography (HRCT) scans, and lung biopsy slides from a wide spectrum of cases of diffuse parenchymal lung disease. The core group established a uniform and consistent approach to these diseases and clarified the terminology, definitions, and descriptions used in routine clinical practice. The final Statement was drafted after a series of meetings of the entire committee.

The level of evidence for the recommendations made in this Statement is largely that of expert opinion developed by consensus. The best evidence is from well-conducted cohort studies. There is no supportive evidence from well-conducted randomized controlled trials.

Validation

The draft document was reviewed by a large and diverse reviewer group, which provided additional expert input. Peer reviewers identified by the ATS and ERS subjected the final document to external review. It was submitted for review and approval to the governing bodies of the ATS and ERS.

Key Messages

1. The idiopathic interstitial pneumonias (IIPs) comprise a number of clinicopathological entities, which are sufficiently different from one another to be designated as separate disease entities. As a group they can be distinguished from other forms of diffuse parenchymal lung disease by clinical methods including history, physical examination, chest radiology and laboratory studies, and pathology.

2. These conditions are rare and few physicians have substantial experience with their diagnosis and management.

3. The new ATS/ERS classification proposed in this document comprises the following clinicopathologic entities in order of relative frequency: idiopathic pulmonary fibrosis (IPF), nonspecific interstitial pneumonia (NSIP), cryptogenic organizing pneumonia (COP), acute interstitial pneumonia (AIP), respiratory bronchiololitis-associated interstitial lung disease (RB-ILD), desquamative interstitial pneumonia (DIP), and lymphoid interstitial pneumonia (LIP).

4. NSIP is an area of uncertainty that requires further definition. The panel recommended that the use of the term NSIP be considered as a provisional diagnosis until there is further clarity on the nature of the corresponding clinical condition. Under the auspices of the ATS, a multidisciplinary panel is reviewing clinical cases of NSIP from around the world. When published (expected in late 2002), this review will be used to better characterize this entity and to determine its relationship to IPF, hypersensitivity pneumonitis, COP, DIP, and LIP.

5. Achieving a correct diagnosis is a dynamic process. During the diagnostic work-up of patients with an IIP, the diagnosis may need to be revised, as more details of history are obtained, when new associations are discovered, or when results of bronchoalveolar lavage, transbronchial biopsy (where appropriate), and surgical lung biopsy become available. It is particularly important to re-evaluate the patient in a search for a specific etiology when NSIP, diffuse alveolar damage (DAD), and LIP are found on lung biopsy. The final diagnosis should be rendered only after the pulmonologist, radiologist, and pathologist have reviewed all of the clinical, radiological, and pathological data obtained from the patient.

6. In order to clarify the relationship between the historical pathologic and clinical terms that have been used for these entities, the new classification defines a set of histologic patterns that provide the basis for a final clinico-radiologic-pathologic diagnosis. Because the histologic patterns seen by pathologists usually allow for better separation of these entities than the imaging patterns seen by radiologists, these histologic patterns provide the primary basis for the various categories of IIP and serve as the foundation for the classification.

7. In the absence of contraindications, surgical lung biopsy is advised in patients with suspected IIP who do not show a classic clinical and HRCT picture of IPF/usual interstitial pneumonia (UIP). The availability of less invasive surgery in the form of video-assisted thoracoscopic lung biopsy has made it more acceptable for clinicians to recommend surgical biopsy to their patients with diffuse parenchymal lung disease. Surgical lung biopsies should be obtained from more than one lobe of the lung.

8. It is recommended that the term pattern be added to the IIP designations when referring to the lung biopsy pathologic pattern, to distinguish it from the clinico-radiologic- pathologic diagnosis (e.g., NSIP, DIP, or LIP).

9. High-resolution computerized tomography (HRCT) scans are indicated for all but a small proportion of patients for whom a specific diagnosis is strongly suggested on the basis of the standard chest radiograph. Careful attention to technique is necessary to assure diagnostic accuracy. The HRCT images should be obtained in accordance with established guidelines and interpreted by a radiologist experienced in the evaluation of diffuse lung diseases.

10. The primary role of HRCT is to separate patients with typical findings of IPF/UIP from those with the less specific findings associated with other idiopathic interstitial pneumonias.

11. Transbronchial biopsies are not useful in the diagnosis of most of the IIPs, with the exception of DAD/AIP, and occasionally organizing pneumonia (OP)/COP. The primary role of transbronchial biopsies is to exclude sarcoidosis and certain infections. Bronchoalveolar lavage is not always required in the assessment of the IIPs.

12. The final diagnosis should be rendered only after the pulmonologist, radiologist, and pathologist have reviewed all of the clinical, radiological, and pathological data obtained from the patient.

13. Trials of therapy should be discouraged until a concerted effort has been made to establish a firm diagnosis based on the integrated approach proposed in this document. These criteria should provide an international standard as the basis for future studies and publications on the subject of IIP.

	INTRODUCTION

TOP CONTENTS EXECUTIVE SUMMARY INTRODUCTION RATIONALE FOR A CHANGE... DEVELOPMENT OF A NEW... PRINCIPLES GUIDING THE... IDIOPATHIC PULMONARY FIBROSIS NONSPECIFIC INTERSTITIAL... CRYPTOGENIC ORGANIZING... ACUTE INTERSTITIAL PNEUMONIA RESPIRATORY BRONCHIOLITIS-... DESQUAMATIVE INTERSTITIAL... LYMPHOID INTERSTITIAL PNEUMONIA REFERENCES

The idiopathic interstitial pneumonias (IIPs) are a group of diffuse parenchymal lung diseases (DPLDs), a group also described as interstitial lung diseases (Figure 1). The IIPs are a heterogeneous group of nonneoplastic disorders resulting from damage to the lung parenchyma by varying patterns of inflammation and fibrosis. The interstitium includes the space between the epithelial and endothelial basement membranes and it is the primary site of injury in the IIPs. However, these disorders frequently affect not only the interstitium, but also the airspaces, peripheral airways, and vessels along with their respective epithelial and endothelial linings (1).

View larger version (26K): [in this window] [in a new window]   Figure 1.   Diffuse parenchymal lung diseases (DPLDs) consist of disorders of known causes (collagen vascular disease, environmental or drug related) as well as disorders of unknown cause. The latter include idiopathic interstitial pneumonias (IIPs), granulomatous lung disorders (e.g., sarcoidosis), and other forms of interstitial lung disease (ILD) including lymphangioleiomyomatosis (LAM), pulmonary Langerhans' cell histiocytosis/histiocytosis X (HX), and eosinophilic pneumonia. The most important distinction among the idiopathic interstitial pneumonias is that between idiopathic pulmonary fibrosis and the other interstitial pneumonias (IPs), which include nonspecific interstitial pneumonia (a provisional term), desquamative interstitial pneumonia, respiratory bronchiolitis-associated interstitial lung disease, acute interstitial pneumonia, cryptogenic organizing pneumonia, and lymphocytic interstitial pneumonia.

Diffuse lung diseases such as emphysema or chronic obstructive lung disease (COPD), bronchiolitis, and pulmonary hypertension are excluded from this discussion. Some categories of diffuse parenchymal lung diseases such as those associated with occupational or environmental exposures and/or collagen vascular disease, granulomatous lung disorders such as sarcoidosis, and a further group comprising several rare forms of DPLD with distinctive and well-defined clinicopathologic features such as lymphangioleiomyomatosis (LAM), pulmonary Langerhans' cell histiocytosis, and eosinophilic pneumonia are also excluded. The IIPs described herein comprise a number of clinicopathologic entities, which are sufficiently different from one another to be designated as separate disease entities. As a group they can be distinguished from other forms of diffuse parenchymal lung disease by clinical methods including history, physical examination, chest radiology and laboratory studies, and pathology.

Idiopathic indicates unknown cause and interstitial pneumonia refers to involvement of the lung parenchyma by varying combinations of fibrosis and inflammation, in contrast to airspace disease typically seen in bacterial pneumonia. The idiopathic interstitial pneumonias include the entities of idiopathic pulmonary fibrosis (IPF), nonspecific interstitial pneumonia (NSIP), cryptogenic organizing pneumonia (COP), acute interstitial pneumonia (AIP), respiratory bronchiolitis-associated interstitial lung disease (RB-ILD), desquamative interstitial pneumonia (DIP), and lymphocytic interstitial pneumonia (LIP). For the purposes of this document, the following terms are viewed as synonymous: idiopathic and cryptogenic as well as pneumonia and pneumonitis.

	RATIONALE FOR A CHANGE IN THE APPROACH TO CLASSIFICATION OF IDIOPATHIC INTERSTITIAL PNEUMONIAS

TOP CONTENTS EXECUTIVE SUMMARY INTRODUCTION RATIONALE FOR A CHANGE... DEVELOPMENT OF A NEW... PRINCIPLES GUIDING THE... IDIOPATHIC PULMONARY FIBROSIS NONSPECIFIC INTERSTITIAL... CRYPTOGENIC ORGANIZING... ACUTE INTERSTITIAL PNEUMONIA RESPIRATORY BRONCHIOLITIS-... DESQUAMATIVE INTERSTITIAL... LYMPHOID INTERSTITIAL PNEUMONIA REFERENCES

1. The terminology applied to the IIPs has been confusing. Clinicians in different countries have employed varied terms such as idiopathic pulmonary fibrosis (IPF) in the United States and cryptogenic fibrosing alveolitis (CFA) or "lone CFA" (CFA not associated with the presence of collagen vascular disease) (2, 3) in the United Kingdom, or idiopathic interstitial pneumonia in Japan (4). These represent clinically defined disease entities that historically have included a range of histologic patterns (5, 6).

2. In clinical practice, patients are commonly misclassified as having an IIP because of inadequate history taking. In addition, an increasing number of associations between the development of DPLD and occupational, environmental, and drug exposures are being described (7). For these reasons, during the diagnostic work-up of the IIPs, a diagnosis may need to be revised at several stages, as more details of history are obtained, when new associations are discovered, or when results of bronchoalveolar lavage, transbronchial biopsy (where appropriate), and surgical lung biopsy become available.

3. These conditions are rare and few physicians have substantial experience with their diagnosis and management. A reported overall prevalence of interstitial lung disease in New Mexico is 80.9 per 100,000 in males and 67.2 per 100,000 in females, corresponding with annual incidence rates of 31.5 per 100,000/yr in males and 26.1 per 100,000/yr in females (10). The IIPs are rare in children, but increase with advancing age. The prevalence of IPF in different series ranges from 6 to 14.6 per 100,000 persons (8, 10), but in those older than 75 yr the prevalence may exceed 175 per 100,000 (8).

4. Lung biopsies are not frequently obtained from patients with clinical evidence of interstitial lung disease (ILD) (11- 15). This low biopsy rate has been attributed to the invasive nature of surgical biopsies and the fact that many patients with these diseases are viewed as too old or too frail to undergo biopsy. The practice of video-assisted thoracoscopic biopsy has resulted in an increase in surgical lung biopsy in some institutions because of decreased morbidity compared with a formal thoracotomy procedure. In addition, there is a lack of confidence in the diagnostic and predictive value of pathology, leading to a view that knowledge of the pathological findings does not alter the treatment approaches. Therefore, clinicians have commonly preferred to rely on a trial of therapy as a predictor of clinical course and prognosis rather than subject the patient to lung biopsy.

5. Pathology-based classifications have largely been developed from data derived from series of surgical (thoracoscopic or open) biopsies and postmortem examinations of lung tissue. Consequently they may represent only a small proportion of patients with these disorders because most patients with IIP do not undergo surgical lung biopsy (3, 10, 12). Occasionally, pathology case series have lacked sufficient supporting clinical and radiological data or have included patients with histories of exposure to occupational or environmental agents or systemic diseases associated with the development of ILD (16).

6. Clinicians frequently do not have access to the opinion of pathologists who are experienced in examining lung biopsies from patients with IIP. Further, a limited number of experts have been relied on to provide pathological identification of the lesions present.

7. Clinicians have frequently been confused by the descriptions provided in pathology reports, particularly when several patterns are described in a single biopsy.

8. It is not always clear, when the common terms are employed, whether they are being used as a pathologic or clinical term. For example, bronchiolitis obliterans organizing pneumonia (BOOP) is often assumed to indicate the idiopathic entity; however, it could represent a nonspecific histologic reaction that can occur in a wide variety of clinical settings. As a further example, a pathologist may recognize the histologic pattern of diffuse alveolar damage (DAD); however, the diagnosis of AIP requires additional clinical information to exclude potential etiologies (19). Much of the evaluation to exclude specific causes is likely to be completed after the lung biopsy has been signed out. Therefore the best diagnosis many pathologists may be able to make in cases of AIP is "DAD, etiology undetermined" with an added comment that the differential diagnosis may include AIP. For these reasons, the term pattern may be added to the designations that have the same term as that used for the clinicopathologic diagnosis (NSIP, DIP, and LIP) when referring to the lung biopsy pathologic pattern only.

9. Pathologists have generally tended to be "splitters" and clinicians tend to be "lumpers" (17). That is, pathologists have divided the IIPs into separate groups based on the histopathologic pattern found on biopsy. However, clinicians have commonly applied a single term that included several different pathologic patterns; for example, the term IPF has been applied to patients with interstitial lung diseases of unknown cause characterized pathologically by several different histologic patterns including DIP, UIP, and NSIP.

	DEVELOPMENT OF A NEW CLASSIFICATION OF IDIOPATHIC INTERSTITIAL PNEUMONIA

TOP CONTENTS EXECUTIVE SUMMARY INTRODUCTION RATIONALE FOR A CHANGE... DEVELOPMENT OF A NEW... PRINCIPLES GUIDING THE... IDIOPATHIC PULMONARY FIBROSIS NONSPECIFIC INTERSTITIAL... CRYPTOGENIC ORGANIZING... ACUTE INTERSTITIAL PNEUMONIA RESPIRATORY BRONCHIOLITIS-... DESQUAMATIVE INTERSTITIAL... LYMPHOID INTERSTITIAL PNEUMONIA REFERENCES

Several developments have prompted a review of the previous classification systems and encouraged the development of a new comprehensive clinical-radiologic and pathologic classification of the IIPs.

1. The publication of large series of patients with IIP with accompanying pathologic evaluation by experts in lung pathology has provided a clearer picture of the types of histopathologic patterns seen and their relationship to the patient's clinical course and responsiveness to treatment (20).

2. The availability of less invasive surgery in the form of video-assisted thoracoscopic lung biopsy has made it more acceptable for clinicians to recommend surgical biopsy to their patients (24). This has led to an increase in the frequency of surgical lung biopsy in some institutions.

3. The widespread use and improved understanding of the value of high-resolution computerized tomography (HRCT) scans in the evaluation of these diseases has led to improved understanding of the extent and severity of the lesions commonly present (26, 28).

4. The development of several new therapeutic approaches for the management of fibrotic lung diseases has prompted renewed interest in understanding the pathogenesis of these disorders (3, 32, 33).

Current Classification of IIP

Liebow and Carrington provided a landmark histologic classification of the chronic interstitial pneumonias in 1969 (34). Liebow and Carrington described five types of chronic interstitial pneumonia: usual interstitial pneumonia (UIP), bronchiolitis obliterans interstitial pneumonia and diffuse alveolar damage (BIP), desquamative interstitial pneumonia (DIP), lymphocytic interstitial pneumonia (LIP), and giant cell interstitial pneumonia (GIP). The evolution of this classification over time is presented in Table 1.

The classification schemes proposed by Liebow and Carrington (34), Müller and Colby (35), and Katzenstein (36) maintained UIP and DIP as separate types of lung diseases in contrast to the concepts of IPF and CFA, which regarded them as part of a single spectrum (3). However, LIP and GIP were dropped because many of the former were thought to develop into lymphomas and were therefore preferably classified as lymphoproliferative disorders, whereas many of the latter were found to be hard metal pneumoconioses. In addition, newly recognized entities including respiratory bronchiolitis-associated interstitial lung disease (RB-ILD) (16), bronchiolitis obliterans organizing pneumonia (BOOP) (termed cryptogenic organizing pneumonia [COP] in some countries) (37, 38), acute interstitial pneumonia (AIP) (39), and nonspecific interstitial pneumonia (NSIP) (17) have been added to the classification of the IIPs (35, 40, 41).

New ATS/ERS Classification

The new ATS/ERS classification proposed in this document comprises the following clinicopathologic entities in the order of relative frequency: IPF/CFA, NSIP (provisional), COP, AIP, RB-ILD, DIP, and LIP (Table 2) (20, 41). The rationale for choosing these terms and including each entity is discussed in each respective section below. To clarify the relationship between the historical pathologic and clinical terms that have been used for these entities, the new classification defines a set of histologic patterns that provide the basis for a final clinico- radiologic-pathologic diagnosis (CRP diagnosis) (Table 2). Because the histologic patterns seen by pathologists allow for better separation of these entities than the imaging patterns seen by radiologists, these histologic patterns provide the primary basis for the various categories of IIP and serve as the foundation for the classification. However, the final clinicopathologic diagnosis, including the issue of whether the disorder is idiopathic, can be made only after careful correlation with clinical and radiologic features. Thus, the final diagnosis should be rendered only after the pulmonologist, radiologist, and pathologist have reviewed all of the clinical, radiological, and pathological data obtained from the patient.

	PRINCIPLES GUIDING THE ASSESSMENT OF PATIENTS WITH IDIOPATHIC INTERSTITIAL PNEUMONIAS

TOP CONTENTS EXECUTIVE SUMMARY INTRODUCTION RATIONALE FOR A CHANGE... DEVELOPMENT OF A NEW... PRINCIPLES GUIDING THE... IDIOPATHIC PULMONARY FIBROSIS NONSPECIFIC INTERSTITIAL... CRYPTOGENIC ORGANIZING... ACUTE INTERSTITIAL PNEUMONIA RESPIRATORY BRONCHIOLITIS-... DESQUAMATIVE INTERSTITIAL... LYMPHOID INTERSTITIAL PNEUMONIA REFERENCES

Before describing the features of each IIP, it is necessary to consider the principles that apply to the general assessment of patients with IIPs. The new classification of IIP is based on clinical, radiological, and pathological criteria.

The Diagnostic Process Is Dynamic

The process of achieving a diagnosis in a patient with IIP is dynamic, requiring close communication between clinician, radiologist, and pathologist. For example, a pathologist is at a disadvantage if asked to interpret a lung biopsy without a relevant history of clinical presentation, radiologic findings, occupational exposure, smoking status, and associated diseases. Also, once a pathologist has recognized a histologic pattern such as NSIP, the clinician needs to go back to the patient and check carefully for antigen exposure that could account for hypersensitivity pneumonitis, laboratory or clinical features of collagen vascular disease, and possible drug or toxic exposure. The practice of observing clinical and radiologic deterioration before obtaining biopsy is not helpful because it delays diagnosis, reduces the likelihood that the disease will be correctly identified, and not infrequently results in patients receiving unnecessary or inadequate treatment. For the same reasons, trials of therapy should be discouraged until a concerted effort has been made to establish a firm diagnosis based on this integrated approach.

Clinical Evaluation

The approach to patients with diffuse parenchymal lung disease begins with a careful history followed by physical examination, routine chest radiographs, and pulmonary function testing (Figure 2) (42). The assessment of the clinical history should include the nature of the first symptoms (usually breathlessness or cough), their progression, clinical course, and in particular the presence of comorbid disease such as collagen vascular disease or immunodeficiency disorders such as infection with the human immunodeficiency virus (HIV). A record of environmental exposures including smoking status, drug use, and detailed occupational exposures with dates, duration of exposure, and a detailed description of work activities is essential. A history of previous malignancy and treatment for malignancy or a family history of lung disease may also be relevant.

View larger version (36K): [in this window] [in a new window]   Figure 2.   The diagnostic process in diffuse pulmonary lung diseases (DPLDs) begins with a clinical evaluation that includes a history, physical examination, chest radiograph, and lung function tests. On the basis of this information, the patients may be divided into two groups: cases that do not represent idiopathic interstitial pneumonia (IIP), owing to recognition of associated conditions or underlying exposures, and cases that could represent IIP. Patients in the latter category typically receive a high-resolution computerized tomography (HRCT) scan. This generally results in four categories of patients: (1) those with distinctive features that allow for a confident diagnosis of idiopathic pulmonary fibrosis (IPF)/usual interstitial pneumonia (UIP) in the appropriate clinical setting, (2) those with atypical clinical or CT features for IPF, (3) those with features diagnostic of another DPLD such as pulmonary Langerhans' cell histiocytosis (PLCH), and (4) those with suspected other forms of DPLD. Although many patients will go directly to surgical lung biopsy, some patients may undergo transbronchial biopsy (TBBx) or bronchoalveolar lavage (BAL). If these findings are nondiagnostic a surgical lung biopsy may be necessary to separate the various IIPs from non-IIP DPLD.

On physical examination the presence of crackles and finger clubbing, although varying in frequency in different forms of IIP, provides a useful clue to the presence of an IIP. Joint swelling or tight skin may suggest collagen vascular disease. After obtaining the clinical history, a chest radiograph, and pulmonary function tests, the clinician should be in a position to decide whether the patient has a DPLD and whether a form of IIP needs to be considered (Figure 2). If the answer to the latter question is yes, an HRCT scan is usually indicated.

Radiological Evaluation

HRCT has become an integral part of the evaluation of the patient with idiopathic interstitial pneumonia (Figure 2 and Table 3). HRCT is indicated for all but a small proportion of patients for whom a specific diagnosis is strongly suggested by the standard chest radiograph. Careful attention to technique is necessary to assure diagnostic accuracy. HRCT images should be performed in accordance with established guidelines (43, 44) and interpreted by a radiologist experienced in the evaluation of diffuse lung diseases (45).

When interpreting the HRCT scan of a patient with diffuse lung disease, the radiologist must first determine the presence or absence of a pattern typical of UIP (Figure 2). In more than 50% of cases suspected to be IPF/UIP, the presence of typical clinical and HRCT features of UIP, when identified by expert clinicians and radiologists, is sufficiently characteristic to allow a confident diagnosis and eliminate the need for surgical lung biopsy (31, 42). The HRCT may also provide clues to non-IIP disorders such as sarcoidosis (46), hypersensitivity pneumonitis (47), lymphangioleiomyomatosis (48), Langerhans' cell histiocytosis (48), and pulmonary alveolar proteinosis (49), and may prompt the selection of bronchoscopy (usually with both bronchoalveolar lavage and transbronchial biopsy) in preference to proceeding to a surgical lung biopsy. Therefore, the primary role of HRCT is to separate patients with UIP from those with non-UIP lesions or those with less specific findings associated with other idiopathic interstitial pneumonias (NSIP, RB-ILD, DIP, and AIP).

A confident radiologic diagnosis of UIP on HRCT is based on a bilateral, predominantly basal, predominantly subpleural, reticular pattern, associated with subpleural cysts (honeycombing) and/or traction bronchiectasis (31, 42). The abnormality decreases gradually in extent on serial scans from the base to the apex of the lungs. Consolidation and nodules are absent. When the radiologic diagnosis of UIP is based on these findings, it is correct in more than 90% of cases (30, 31, 42). However, when the CT findings are "atypical" (e.g., upper lobe or peribronchovascular predominance, predominant ground glass abnormality or micronodules), or when there is one or more "atypical" clinical features (e.g., young age, inconclusive exposure history, lack of dyspnea, absence of restrictive lung function defect, or presence of marked lymphocytosis on bronchoalveolar lavage [BAL]), then biopsy is indicated. The primary role of HRCT is to separate patients with typical findings of IPF from those with the less specific findings associated with other idiopathic interstitial pneumonias.

Role of Surgical Lung Biopsy

A surgical lung biopsy is necessary for a confident clinicopathologic diagnosis except in cases with a typical clinical-radiological picture of UIP/IPF. This is not to say that a biopsy is always necessary to make a clinical diagnosis (Table 4). Pathology is least helpful when obtained late in the course of the illness or after commencement of treatment.

The benefits of obtaining a surgical lung biopsy can be briefly summarized as follows:

1. Establishment of a firm clinicopathologic diagnosis allows the patient and clinician to make more informed decisions about therapy.

2. Almost all of the current treatments for the IIPs have potentially serious risks and side effects, and it is not reasonable to expose patients to these risks in the presence of diagnostic uncertainty.

3. Detection of fibrotic processes related to specific exposures can have important compensation implications for the patient, and important public health consequences for the community; for example, asbestosis.

Although a highly probable diagnosis of IPF can be made without a lung biopsy, a definitive diagnosis of IPF and of the other forms of IIP can be established only with the aid of a surgical lung biopsy (Figure 2). In most cases the biopsy provides definitive classification of patients into the recognized histologic patterns of UIP, NSIP, OP, DAD, DIP, respiratory bronchiolitis (RB), and LIP. It also allows for confirmation or exclusion of an alternative diagnosis such as sarcoidosis, hypersensitivity pneumonitis, LAM, or lymphangitic carcinoma, or suggests the presence of an occupational disease such as hard metal disease. In lung biopsy specimens with moderate or marked acute and/or chronic inflammation, it is useful to perform special stains to exclude infectious organisms.

Several issues relating to lung biopsies for diagnosis of IIPs need to be considered.

1. The role of transbronchial biopsies in the diagnosis of the IIPs in most cases is to exclude sarcoidosis, neoplasms, and certain infections. In some cases with classic clinical and radiologic features of COP (50, 51) or AIP the histologic diagnosis of an OP or DAD pattern, respectively, may be viewed as confirmatory on a bronchoscopic biopsy specimen (see below).

2. It can be helpful to have surgical lung biopsies from more than one lobe of the lung (52, 53). In addition, if the lung shows severe fibrosis with honeycombing the biopsy specimen should not be taken from the worst-looking areas because these frequently show nonspecific changes. However, if the lung does not show severe fibrosis or honeycombing grossly the surgeon should take the biopsy from the abnormal areas of the lung. HRCT scanning may guide the surgeon to the most optimal biopsy sites (54). Ideally the lung biopsy should include the full spectrum of the gross appearance, including honeycomb foci, because these confirm severe fibrosis, one of the criteria for UIP. The surgeon must also attempt to avoid deflation of the specimen through clamping, as this complicates interpretation of histological findings. Specimen atelectasis can be corrected by inflating lung biopsies with formalin either by injection with a syringe or by gently shaking thin slices of the lung biopsy specimen in the specimen container before paraffin processing (55). An effort should be made to communicate these issues to the thoracic surgeon.

3. In a small proportion of cases the pathologic diagnosis may need to be revised in the light of an unexpected clinical course, identification of a potential cause for lung fibrosis, or response to treatment. Periodic review should include re-examination of the original lung biopsy and radiologic material in addition to the data from the clinical follow-up.

4. In patients with biopsies from multiple lobes, sometimes the second or third lobe may show a different pattern than the first, for example, a UIP pattern in the lower lobe and an NSIP pattern in the upper lobe. An article by Flaherty and coworkers indicates that if a UIP pattern is present in one of the lobes and an NSIP pattern is present in one or more of the other lobes, the clinical behavior is similar to that of IPF (53). For this reason UIP is the default pattern in such a case. Although uncommon, one lobe may show a pattern of ill-defined fibrosis simulating UIP and another lobe may show a "specific" lesion such as a granulomatous reaction or asbestos bodies. In such a case, the diagnosis is that most consistent with the specific histologic finding, that is, for the above-described examples, sarcoidosis or hypersensitivity pneumonitis or asbestosis.

Unclassifiable Interstitial Pneumonia

There is a small subset of patients with interstitial pneumonia that remains unclassifiable after extensive clinical, radiologic, and/or pathological examination. This guideline has resisted the creation of an additional nonclassifiable category because this is clinically unhelpful. However, it is standard in tumor classifications proposed by the World Health Organization to have an unclassifiable category, because there are cases when a specific pathologic diagnosis cannot be made (56). However this has not been fully addressed in the classification of interstitial pneumonia (57, 58). For interstitial lung diseases this situation often exists when some critical piece of data is unavailable or when there is a major discrepancy between the clinical, radiologic, and/or pathologic information. Some examples of reasons or circumstances under which a case may be unclassifiable include the following.

1. Inadequate clinical information.

2. Inadequate radiologic data.

3. An inadequate or nondiagnostic biopsy (e.g., because of small size or poor sampling).

4. The existence of a major discrepancy between clinical, radiologic, and pathologic findings.

5. Previous therapy resulting in alterations in the radiologic or histologic findings.

6. Discrepancy between histologic findings in different lobes that is not resolved after correlation with clinical and radiologic data. The issue of NSIP in one lobe and UIP in another lobe has been addressed elsewhere in this document. Another example is the coexistence of multiple histologic patterns; for example, a biopsy specimen from one lobe may show a UIP or NSIP pattern and other areas may show features of eosinophilic pneumonia, organizing pneumonia, or acute lung injury with fibrin and/or hyaline membranes. In such rare instances, one should rely on clinical findings and the most prominent radiologic findings to determine which of these possibilities appears to be the major/predominant lesion. This careful analysis will often result in a more specific clinico-radiologic-histologic diagnosis.

In summary, we propose that cases having any of the above-described unresolved issues to be called unclassifiable interstitial pneumonia. In this circumstance, the clinician would then need to embark on treatment based on the most probable diagnosis after detailed clinico-radiological-pathological case discussion with the pathologist and radiologist. Importantly, this category designation should not be used for cases of clearly defined NSIP or cases in which the distinction between the UIP and fibrosing NSIP patterns is difficult. In such cases, one should make the best possible diagnosis given the available information, realizing the differential diagnosis may be a challenge. Finally, the purpose of the concept of unclassifiable interstitial pneumonia is not to create an entity from which clinical studies might derive, but to emphasize the requirement for adequate clinical, radiologic, and pathologic information for classification and to acknowledge that uncertainty remains in individual cases.

Bronchoalveolar Lavage Fluid Evaluation

BAL is not always required in the assessment of the IIPs. However, if, as is commonly the case, it has been performed in the diagnostic work-up of diffuse parenchymal lung disease to exclude infection or tumor, the results may assist in the decision to perform a surgical biopsy and in distinguishing different forms of IIP (59). Although not diagnostic in IIP, a "typical" BAL cell pattern strengthens the clinical diagnosis and may contribute to the clinico-radiologic-pathologic assessment in difficult cases (1, 60). In rare instances, diagnostic features may be found, for example, in pulmonary alveolar proteinosis (64). BAL may also be suggestive of Langerhans' cell histiocytosis (1). The presence of hemosiderin-laden macrophages suggests alveolar hemorrhage. In addition, lipid-laden macrophages can be seen in a variety of settings including aspiration of material from the stomach or upper airway, oily liquids, and oil-based nasal instillation as well as lipid emboli or amiodarone therapy (1).

	IDIOPATHIC PULMONARY FIBROSIS

TOP CONTENTS EXECUTIVE SUMMARY INTRODUCTION RATIONALE FOR A CHANGE... DEVELOPMENT OF A NEW... PRINCIPLES GUIDING THE... IDIOPATHIC PULMONARY FIBROSIS NONSPECIFIC INTERSTITIAL... CRYPTOGENIC ORGANIZING... ACUTE INTERSTITIAL PNEUMONIA RESPIRATORY BRONCHIOLITIS-... DESQUAMATIVE INTERSTITIAL... LYMPHOID INTERSTITIAL PNEUMONIA REFERENCES

The terms UIP and IPF have become more narrowly defined since they were originally proposed several decades ago (5, 65). The relationship between historically defined IPF (or lone CFA) and UIP has been described in an ATS Consensus Statement entitled "Idiopathic Pulmonary Fibrosis: Diagnosis and Treatment" (60). According to the current definition, IPF is a distinctive type of chronic fibrosing interstitial pneumonia of unknown cause limited to the lungs and associated with a surgical lung biopsy showing a histologic pattern of UIP (60). In the presence of a surgical biopsy showing a UIP pattern the diagnosis of IPF requires (1) exclusion of other known causes of interstitial lung disease including drug toxicities, environmental exposures, and collagen vascular diseases, (2) characteristic abnormalities on conventional chest radiographs or high-resolution computed tomography (HRCT) scans as described below (60), and (3) abnormal pulmonary function studies showing restriction (reduced total lung capacity [TLC], or reduced vital capacity [VC] with a normal or increased FEV₁/FVC ratio) and/or impaired gas exchange [increased P(A-a)O₂ (alveolar-arterial pressure difference for O₂), decreased Pa_O2 with rest or exercise, or decreased DL_CO (diffusing capacity of the lung for CO)].

Definitive histologic diagnosis of IPF requires a surgical lung biopsy. The criteria for diagnosis of IPF in the absence of a surgical lung biopsy are summarized in Table 4 (60). In such cases the diagnosis of IPF is likely but not as certain as when a surgical lung biopsy also shows a UIP pattern. However, lung biopsy occasionally may also not be definitive. As mentioned above, this may arise when there is histologic heterogeneity in different lobes of the lung in IPF (53). So after correlating all the clinical, radiological, and pathological information, the final diagnosis may still be IPF in a patient with typical clinical-radiological IPF, even though a lung biopsy shows a fibrosing NSIP pattern.

Clinical Features

Onset of symptoms is usually gradual, with dyspnea the most prominent and disabling symptom (3). A nonproductive cough is usual and may be paroxysmal (60). It is often refractory to antitussive agents. The patient's age at onset is usually greater than 50 yr and IPF is slightly more common in males (3, 11). Constitutional symptoms are unusual. Digital clubbing develops in 25 to 50% of patients (3, 11), and Velcro-type fine end-inspiratory crackles that are initially confined to the basal areas are found on chest auscultation (60). These progress gradually to involve the entire lung. Features of right heart failure and peripheral edema develop only in the late stages. Most patients exhibit a restrictive pattern of ventilatory defect with a decrease in DL_CO and low resting Pa_O2, which falls on exercise. Pulmonary function or chest radiographs may be normal or near normal in the early phase of IPF (66) In smokers and ex-smokers with IPF, coexistent chronic obstructive pulmonary disease may result in relatively higher lung volumes compared with never-smoking patients with IPF (60).

Clinical course In most patients, symptoms have been present for more than 6 mo before presentation (3, 67). The clinical course is invariably one of gradual deterioration. Median length of survival from time of diagnosis varies between 2.5 and 3.5 yr (60, 68, 69). Occasionally, periods of rapid decline are recognized (70). These may represent accelerated disease, intercurrent viral infection with the development of organizing pneumonia, or diffuse alveolar damage (41). Improvement in lung physiology and radiologic abnormalities is rare (22).

Bronchoalveolar lavage features Bronchoalveolar lavage fluid contains an excess of neutrophils, the proportions of which correspond to the extent of reticular change on HRCT (71). There may also be a mild or moderate increase in the percentage of eosinophils (1, 62). BAL cell counts, although correlating with severity of disease, do not predict prognosis (72). When eosinophils represent more than 20% of the count, consideration should be given to an eosinophilic lung disease (73). Lymphocytosis is not a feature of UIP, and counts above 15% should alert to an alternative diagnosis such as NSIP, COP, hypersensitivity pneumonitis, sarcoidosis, or other granulomatous lung disease (60).

Radiologic Features

The commonest chest radiographic abnormality in patients with IPF is peripheral reticular opacity, most marked at the bases, and often associated with honeycombing and lower lobe volume loss (Figure 3A and Table 3) (74, 75). In patients with associated upper lobe emphysema, the radiographic lung volumes may be normal or even increased. Chest radiographs may occasionally be normal in patients with IPF (66).

View larger version (165K): [in this window] [in a new window]   View larger version (157K): [in this window] [in a new window]   View larger version (145K): [in this window] [in a new window]   Figure 3.   Idiopathic pulmonary fibrosis. (A) Chest radiograph shows typical peripheral reticular opacity, most marked at the bases, with honeycombing. Lower lobe volume loss (not present in this case) is also common. Chest radiographs may occasionally be normal in patients with IPF. (B and C) CT images show basal predominant, peripheral predominant reticular abnormality with traction bronchiectasis and honeycombing, typical of IPF.

CT features UIP is characterized on CT by the presence of reticular opacities, often associated with traction bronchiectasis (Table 3 and Figure 3B and 3C). Honeycombing is common (Figure 3B and 3C) (76). Ground glass attenuation is common, but is usually less extensive than reticular abnormality (26, 77). Architectural distortion, reflecting lung fibrosis, is often prominent. Lobar volume loss is seen with more advanced fibrosis. The distribution of UIP on CT is characteristically basal and peripheral, although often patchy (26, 78).

View larger version (131K): [in this window] [in a new window]   View larger version (149K): [in this window] [in a new window]   Figure 4.   Progression of IPF in a 65-yr-old man. (A and B) CT scans obtained 20 mo apart show progression of lung fibrosis and honeycombing. Areas of ground glass opacity have progressed to reticular abnormality. Honeycomb cysts have enlarged, and the extent of disease has increased.

CT-pathologic correlation Reticular abnormality on CT correlates with fibrosis on histopathologic examination (26, 77). Honeycombing on CT correlates with honeycombing on biopsy. When ground glass attenuation is associated with reticular lines, traction bronchiectasis, or bronchiolectasis, it usually indicates histologic fibrosis. Isolated ground glass attenuation may correlate with evidence of interstitial inflammation, airspace filling by macrophages, patchy fibrosis, or a combination of these (26, 84, 85).

Radiologic differential diagnosis The CT pattern of UIP due to IPF can be indistinguishable from that found in UIP due to asbestosis and to collagen vascular disease. The presence of pleural plaques or diffuse pleural thickening helps to distinguish asbestosis from IPF. Patients with chronic hypersensitivity pneumonitis (47, 86), or with end-stage sarcoidosis (87), may uncommonly develop a CT pattern similar to that of UIP. Hypersensitivity pneumonitis should be considered if poorly defined fine micronodules are seen, or if there is sparing of the lung bases. Sarcoidosis should be suspected if the cysts are large, or if peribronchovascular nodules are present.

Histologic Features

The key histologic features of the UIP pattern are architectural destruction, fibrosis often with honeycombing, scattered fibroblastic foci, patchy distribution and involvement of the periphery of the acinus or lobule (Figure 5A) (23, 40). It has a heterogeneous appearance at low magnification, with alternating areas of normal lung, interstitial inflammation, fibrosis, and honeycomb change (Table 5) (23, 40). The histological changes affect the peripheral subpleural parenchyma most severely. Interstitial inflammation is usually mild to moderate, patchy, and consists of an alveolar septal infiltrate of lymphocytes, plasma cells, and histiocytes associated with hyperplasia of Type II pneumocytes. The fibrotic zones show temporal heterogeneity with dense acellular collagen and scattered fibroblastic foci (Figure 5B and 5C). Areas of honeycomb change are composed of cystic fibrotic airspaces, which are frequently lined by bronchiolar epithelium and filled with mucin. Smooth muscle hyperplasia is commonly seen in areas of fibrosis and honeycomb change (88). Areas of relatively normal lung should be present in surgical biopsy specimens in order to exclude the presence of active lesions of other interstitial disorders. Otherwise the UIP pattern may be difficult to recognize and a pathologist may only be able to diagnose "severe fibrosis with honeycomb change." In some patients with a UIP pattern on lung biopsy, specimens from a second or third lobe of lung may not fulfill the histologic criteria for UIP and suggest other patterns such as NSIP. However, in such a setting the default pathologic diagnosis is UIP.

View larger version (160K): [in this window] [in a new window]   View larger version (157K): [in this window] [in a new window]   View larger version (155K): [in this window] [in a new window]   Figure 5.   (A) Usual interstitial pneumonia pattern. Patchy fibrosis with remodeling of the lung architecture shows a striking subpleural distribution. Interstitial chronic inflammation is mild with a few lymphoid aggregates. Areas of "normal" lung are present that lack active lesions of other interstitial lung disorders. (B) There is marked fibrosis consisting of dense collagenous scarring with remodeling of the lung architecture and cystic changes. (C) The dense collagenous scar is juxtaposed with a fibroblastic focus of loose organizing connective tissue.

Patients who are biopsied during an accelerated phase of their illness may show a combination of UIP pattern and a variety of acute lesions. These include infection, prominent organizing pneumonia, DAD, and capillaritis. If no cause can be determined this may represent "accelerated decline of IPF" or acute exacerbation of IPF (70). A pattern of interstitial inflammation and fibrosis nearly indistinguishable from that seen in UIP can occur in patients with collagen vascular diseases, certain drug-induced lung diseass, chronic hypersensitivity pneumonitis, asbestosis, and familial IPF (Table 6). There is no single histologic finding that has shown a consistent correlation with treatment response or prognosis in IPF.

Histologic differential diagnosis The differential diagnosis of the IIPs must be approached in two ways: histologically and clinically. In interpreting lung biopsies, the pathologist must address the differential diagnosis on the basis of the histologic pattern (89). A search should be made for histologic clues to a potential cause such as asbestos bodies, infectious agents, or other exogenous agents. The clinician must address most of the etiologic possibilities and in most cases ultimately determines whether the process is idiopathic.

IPF: Areas of Uncertainty

Because the revised definition of IPF is more restrictive, previously reported series of cases need to be reviewed to establish the proportions of cases that would now be considered as non-UIP (and especially the cases that would be currently reclassified as NSIP).

Do patients with histologic UIP who have an atypical CT pattern have different clinical features or clinical course? The true clinical course of confirmed IPF and the impact (if any) of treatment need to be defined.

Is fibrosis in IPF another predisposing factor for lung carcinoma in addition to smoking?

The pathogenesis of UIP needs to be defined.

Histologic predictors of prognosis need to be sought.

The characteristics and causes of accelerated IPF require study.

The interobserver variability of pathological interpretation, particularly among general pathologists, needs to be defined.

Is there a difference in the clinical features and prognosis of UIP in patients with IPF compared with those with known causes such as collagen vascular disease?

	NONSPECIFIC INTERSTITIAL PNEUMONIA

TOP CONTENTS EXECUTIVE SUMMARY INTRODUCTION RATIONALE FOR A CHANGE... DEVELOPMENT OF A NEW... PRINCIPLES GUIDING THE... IDIOPATHIC PULMONARY FIBROSIS NONSPECIFIC INTERSTITIAL... CRYPTOGENIC ORGANIZING... ACUTE INTERSTITIAL PNEUMONIA RESPIRATORY BRONCHIOLITIS-... DESQUAMATIVE INTERSTITIAL... LYMPHOID INTERSTITIAL PNEUMONIA REFERENCES

The recognition that lung biopsy samples from some patients with idiopathic interstitial disease do not fit into any well-defined histologic patterns of idiopathic interstitial pneumonia led to proposals of the terms "unclassified interstitial pneumonia" by Kitaichi in 1990 (41) and NSIP by Katzenstein and Fiorelli in 1994 (17). The concept of NSIP has helped to identify a group of interstitial lung disorders with a more favorable prognosis and that need to be distinguished from IPF but that also differ from DIP, AIP, and COP (17, 18, 20, 21, 23, 35, 90). However, the term NSIP had also since the 1980s been used previously for noninfectious interstitial pneumonitis in HIV-infected patients (91).

Katzenstein and Fiorelli divided NSIP into three major subgroups based on the amount of inflammation and/or fibrosis in the lung biopsies: Group I, primarily with interstitial inflammation; Group II, with both inflammation and fibrosis; and Group III, primarily with fibrosis (17). In this study 39% of the patients with NSIP as a lung biopsy finding had a broad spectrum of associated conditions, such as collagen-vascular diseases (16%), slowly resolving DAD, and a variety of exposures (17).

In more recent publications the term NSIP has evolved from its original use, which was intended to indicate a histologic pattern with a variety of etiologies (17). Now it is almost exclusively used to identify a form of IIP (18, 20, 21, 35, 36, 40, 90, 94). However, the concept of an idiopathic form of NSIP presents a problem for the clinician because there is no recognized and distinctive clinical description for patients presenting with this histologic pattern on lung biopsy. Although these patients have a better prognosis than those with IPF, the clinician does not know this in advance. This improved prognosis has been observed in several studies and appears to correlate with differences in the dominant pathology, whether a cellular or fibrotic pattern of NSIP is present and dominates (17, 18, 20, 21, 23, 36, 40, 94, 95). Further subclassification may become necessary, but this remains an issue for further study. It is possible that specific occupational exposures may give rise to this pattern.

Importantly, the finding of an NSIP pattern on biopsy should prompt the clinician to redouble efforts to find potentially causative exposures. NSIP may be the presenting manifestation preceding the diagnosis of collagen vascular disease by several months or several years. The NSIP pattern may also be the lone histologic feature in a patient with hypersensitivity pneumonitis. Therefore, care should be taken to search for serological and other markers of the connective tissue diseases and a careful search for potential exposures is essential. It is possible that specific occupational exposures may give rise to this pattern.

After considerable debate regarding the best clinical term for patients with this histologic pattern, it was decided that use of the term NSIP was acceptable as a provisional measure until there is further clarity on the nature of the corresponding clinical condition. Although there are several reasons to be critical of the term NSIP, one advantage is that the name implies the uncertainty that prevails.

Clinical Features

The clinical features of patients with an NSIP pattern on surgical lung biopsy are poorly defined. NSIP probably represents a heterogeneous group of disorders and subsets of patients with different clinical courses are being recognized, but at this time there is no consensus about these.

The mean age of patients at onset of NSIP is a decade or more younger than patients with IPF (median age of onset is 40 and 50 yr in different case series) (18, 90) and unlike IPF may occur in children (17). There is neither sexual predominance nor association with cigarette smoking (18). Onset is usually gradual, but a minority of patients may have a subacute presentation. The median duration of symptoms before diagnosis is 18 and 31 mo in different series, but may be as short as 6 mo or as long as 3 yr (18, 90). Breathlessness, cough, and fatigue are usual symptoms, and almost half present with a history of weight loss (mean, 6 kg) (18). Constitutional symptoms such as fever are present in a minority and finger clubbing, which occurs in 10 to 35% of patients, is less common than it is in IPF. Crackles are initially predominantly basal but may be widespread. Inspiratory squeaks are found in some (18). Other clinical features are similar to those found in IPF. Lung function tests show similar but milder physiological abnormalities than those found in IPF; that is, a restrictive ventilatory defect in more than 90% of patients, mild airflow limitation in a minority, and reduced DL_co in all. Some patients may show moderate to severe physiological abnormalities. The KCO (CO transfer coefficient) is normal in about 50%. More than two-thirds develop hypoxemia during exercise.

Clinical course The prognosis of NSIP is more variable than in IPF and appears to depend on the extent of fibrosis (17, 21, 23). Some patients experience almost complete recovery, and most of the remainder stabilize or improve on treatment. Relapse may occur (18). A minority of patients progress and die (17, 21, 23, 90).

Bronchoalveolar lavage features Unlike in UIP, increases in the percentages of lymphocytes occur in about 50% of cases, and similar proportions also have increased numbers of neutrophils and/or eosinophils (18, 21, 97, 98). The presence of bronchoalveolar lavage lymphocytosis strengthens the suspicion of NSIP in conjunction with other findings, including HRCT and pulmonary function test results.

Radiologic Features

Chest radiographic features NSIP typically shows bilateral pulmonary infiltrates and the lower lung zones are more frequently involved, although no large detailed analysis of the radiographic appearances of patients with NSIP exists. Of 97 reported cases, the chest radiograph was abnormal in 91 (94%) (17, 18, 94). Patchy parenchymal opacity was the most common radiographic feature in one study (94), but interstitial abnormalities have also been described (Table 3) (17).

CT features Analysis of a total of 85 patients from three studies (18, 95, 99) shows that ground glass attenuation is the predominant finding in the majority of cases (Figure 6) and is the sole abnormality in about one-third of cases. It is most commonly bilateral and symmetrical with subpleural predominance. Irregular linear or reticular opacities are seen in approximately half of all cases, and may be associated with traction bronchiectasis. In general, honeycombing and consolidation are relatively infrequent. Differences among the three studies in the prevalence of honeycombing, consolidation nodules, and bronchovascular bundle thickening may reflect different study populations or different criteria for the inclusion of cases as NSIP. Fibrosing NSIP may be associated with HRCT evidence of honeycombing, and in such cases only the pathologist can make the distinction from the UIP pattern (21).