Fibrosis

Idiopathic pulmonary fibrosis (IPF) is a chronic disorder that is associated with a poorer prognosis than subacute idiopathic interstitial pneumonias (IIPs). IPF can be differentiated from other IIPs on the basis of its histologic pattern of usual interstitial pneumonia (UIP). Although a surgical lung biopsy specimen showing a UIP pattern is required for the definitive diagnosis of IPF, clinical criteria can be used in the absence of a lung biopsy specimen to make a likely diagnosis of IPF. The predictive value of these criteria largely depends on the expertise of the clinician and radiologist, but considerable interobserver variability exists even when evaluations are performed by experts in the field. Moreover, these criteria lead to misdiagnosis in about 25 to 35% of cases. Interobserver variability is reduced and diagnostic accuracy is improved in cases in which a diagnosis is made with a high degree of confidence. Diagnostic accuracy is also higher when the diagnosis is made by a core group of experts rather than by a referring center. The decision on whether or not to perform a surgical lung biopsy is difficult. It is clearly indicated in cases in which clinical or radiologic findings are atypical or when the diagnosis is made with a low degree of certainty.

Key words: idiopathic interstitial pneumonia; idiopathic pulmonary fibrosis; interobserver variability; interstitial lung disease; surgical lung biopsy

Abbreviations: AIP = acute interstitial pneumonia; COP = cryptogenic organizing pneumonia; DIP = desquamative interstitial pneumonia; HRCT = high-resolution CT; IIP = idiopathic interstitial pneumonia. ILD = interstitial lung disease; IPF = idiopathic pulmonary fibrosis; NSIP = nonspecific interstitial pneumonia. UIP = usual interstitial pneumonia; VATS = video-assisted thoracoscopic surgery

**********

The 2002 international consensus classification of the idiopathic interstitial pneumonias (IIPs) recommended by the American Thoracic Society and European Respiratory Society (1) recognizes seven distinct Clinicopathologic entities. In order of frequency, these entities include idiopathic pulmonary fibrosis (IPF), nonspecific interstitial pneumonia (NSIP), cryptogenic organizing pneumonia (COP), acute interstitial pneumonia (AIP), respiratory bronchiolitis-associated interstitial lung disease, desquamative interstitial pneumonia (DIP), and lymphoid interstitial pneumonia. IPF, unlike the other entities, is a smoldering chronic disorder that evolves over many years, resulting in a gradual loss in lung function interspersed with exacerbations or periods of rapid decline that may cause respiratory arrest and death. In contrast, AIP is an aggressive disorder that evolves rapidly and proves fatal in [greater than or equal to] 50% of cases within 1 to 2 months of onset. (1) The other entities are subacute; most follow a clinical course lasting weeks to months, and patients often improve with treatment or smoking cessation. IPF usually proves fatal, with a median survival time of about 3 years from the time of diagnosis. (1) As shown in retrospective analyses of lung biopsy specimens, the histologic pattern found in IPF (ie, usual interstitial pneumonia [UIP]) is associated with significantly poorer survival times than NSIP and other subacute IIPs. (2,)

HISTOLOGIC DIFFERENCES AMONG IIPs

The different histologic patterns can be distinguished from one another by several notable differences. (1,4) The UIP histology found in IPF stands out from the other patterns in that it is a patchy and nonuniform disease, with inflammation and fibrosis that is temporarily and spatially heterogeneous, and exhibits the characteristic presence of honeycomb change (Table 1). (5,6) "Fibroblastic foci" are present, but are not specific and are thought to be a marker of disease activity. In comparison, AIP is associated with diffuse alveolar damage and the presence of hyaline membranes, which are sometimes detected in the organizational phase. COP is technically not an interstitial lung disease (ILD) alone in that there is also a prominent interalveolar component of loose connective tissue proliferation. DIP is not really idiopathic but is related to smoking, with uniform and diffuse intraalveolar macrophage accumulation, minimal fibrosis, and the absence of fibroblastic loci. Lymphoid interstitial pneumonia is distinguished by the presence of monotonous sheets of lymphoplasmacytic cells that expand the interstitium much like pulmonary lymphoma. The cellular pattern of NSIP is distinguished from UIP by the presence of more prominent septal inflammation, with or without fibrosis, in a temporally and spatially uniform pattern. Distinguishing NSIP with fibrosis from UIP can be problematic. Unlike UIP, the fibrotic pattern of NSIP is temporally homogeneous, and fibroblastic loci are generally, but not always, inconspicuous. These IIPs must also to be distinguished from hypersensitivity pneumonias, which are usually recognized by the presence of well or poorly formed granulomas that are often in a perivascular or peribronchiole distribution, in addition to chronic interstitial inflammation.

IPF has been defined as a "specific form of chronic fibrosing interstitial pneumonia limited to the lung and associated with the histologic appearance of UIP on surgical lung biopsy." (4) The UIP pattern, however, is not specific to IPF, as it is also found in collagen vascular diseases, chronic hypersensitivity pneumonitis, and other forms of fibrotic lung disease. (4) These disorders may be distinguished from IPF on the basis of clinical, radiologic, or serologic findings. The spatial heterogeneity in IPF is evident on gross examination, as shown by very abrupt transitions from regions with subpleural fibrosis and honeycomb change to regions of normal lung tissue. The fibrotic pattern recognized in UIP tends to start in the subpleural areas of the lung and shows greater involvement in the lower lobes.

TYPES OF LUNG BIOPSIES

Although in some cases a lung biopsy is required for the definitive classification of IIPs, a clinical diagnosis can be made without biopsy. (1) Surgical lung biopsy, preferably by video-assisted thoracoscopic surgery (VATS) or alternatively by limited open thoracotomy, provides the best tissue samples for differentiating UIP from the other IIPs. (4) Transbronchial biopsies are not helpful in this regard, although they may be useful for excluding lips by confirming an alternative diagnosis, usually infection, malignancy, or sarcoidosis. In such cases, transbronchial biopsy may obviate the need for open biopsy. Percutaneous CT scan-guided needle biopsy is another method for sampling the lung, but it is not relevant to the diagnosis of IIPs.

Several studies (7-10) have claimed that VATS and limited open thoracotomy provide diagnostic yields in excess of 90%. Rena and colleagues (7) reported that VATS allowed a histologic diagnosis in all 58 patients presenting with an ILD of unknown etiology. These patients had previously undergone clinical evaluation, high-resolution CT (Ht/CT) scanning, serologic testing, BAL, and bronchoscopy. In this cohort, the most common diagnoses following VATS were IPF (24%), COP (17%), and sarcoidosis (17%). Postoperative complications were rare, with two patients experiencing prolonged air leakage for > 5 days. Chest tube drainage was needed for a median of 3 days (range, 1 to 7 days). The median hospital stay was 4 days (range, 2 to 7 days). Several nonrandomized comparisons (8,9,11) have shown that VATS is associated with less postoperative morbidity and mortality, and with shorter hospital stays than limited thoracotomy. However, little difference between the two techniques was seen in another study, (10) although patients undergoing VATS returned to their preoperative functional level in a shorter period of time.

Lung biopsy specimens may also be obtained by video-assisted medical thoracoscopy performed by interventional pulmonary endoscopists. Vansteenkiste and coworkers (12) reported a 75% diagnostic yield with this technique in a prospective study of 24 consecutive patients presenting with ILD that had not been specified after clinical and radiologic workups, BAL, and transbronchial biopsy. An IIP was diagnosed in 33% of the patients, including 17% with UIP/IPF. Major complications were not reported, but prolonged air leakage for > 7 days occurred in 29% of patients. Chest tube drainage was needed for a mean time of 5.3 days (range, 0 to 14 days).

CLINICAL CRITERIA FOR DIAGNOSIS OF IPF

According to the 2002 international consensus Classification, (1) a surgical lung biopsy is required for the definitive diagnosis of IPF. However, a diagnosis of IPF can be considered likely in the absence of a surgical lung biopsy specimen when certain major and minor criteria are met. In such cases, the following four major criteria must be satisfied: (1) exclusion of other known causes of the ILD; (2) abnormal findings of pulmonary function studies showing evidence of restriction and impaired gas exchange; (3) bibasilar reticular abnormalities with minimal ground-glass opacities seen on HRCT scans; and (4) transbronchial lung biopsy specimen or BAL fluid sample showing no features to support an alternative diagnosis. In addition, at least three of the following minor criteria should be present: age > 50 years; insidious onset of otherwise unexplained dyspnea on exertion; duration of illness lasting > 3 months; and bibasilar inspiratory crackles.

INTEROBSERVER VARIATION

Peckham and colleagues (13) recently evaluated the interobserver variability of the clinical criteria in making a diagnosis of IPF. Three board-certified pulmonary physicians without extensive expertise in evaluating IPF reviewed the HRCT scans and clinical data of 26 patients referred for surgical lung biopsy due to ILD. Fourteen patients had a clinical history consistent with IPF and subsequently underwent surgical lung biopsy showing UIP. The other cases included NSIP (n = 5), sarcoidosis (n = 2), malignancy (n = 2), COP, respiratory bronchiolitis-associated ILD, and end-stage fibrosis. The sensitivity and positive predictive value for a diagnosis of IPF based on the clinical criteria identified in the 2002 international consensus classification (1) were 71% and 77%, respectively. The specificity and negative predictive value were 75% and 69%, respectively. Overall, the pulmonary physicians had four false-negative results and three false-positive results (including two patients with NSIP) for an overall accuracy of 73%. Similar results were obtained when only the HRCT scan was used to make a diagnosis.

The interobserver variability of the clinical criteria and HRCT scan was assessed by a [kappa] statistic. High [kappa] values are indicative of low interobserver variability. Values of > 0.80 are considered to be excellent, and those from 0.60 to 0.79 are considered to be good. For reference, the HRCT scan diagnosis of pulmonary embolism has [kappa] values of 0.72 to 0.96, and the CT diagnosis of cystic lung disease has values of 0.77 to 1.0. In the study by Peckham and colleagues, (13) the [kappa] values were 0.53 for the American Thoracic Society/European Respiratory Society clinical criteria and 0.59 for HRCT scans, indicating that these approaches are associated with significant interobserver variability. Moreover, they show that using the clinical criteria in the absence of a surgical lung biopsy may frequently lead to the misdiagnosis of IPF.

Aziz and coworkers (14) evaluated the interobserver variability among 11 thoracic radiologists in using HRCT scans for the diagnosis of 131 patients with diffuse parenchymal lung diseases. Overall, there was only moderate agreement among the radiologists in their diagnosis of the entire cohort ([kappa] = 0.48). It is important to note that the higher the [kappa] value, the stronger the interobserver agreement. The interobserver agreement was higher for cases from regional teaching centers than for those from a referral hospital ([kappa] = 0.60 and 0.34, respectively). IPF was the most frequently offered diagnosis, but there was only moderate agreement among the radiologists ([kappa] = 0.50). Interobserver variability was even greater in cases diagnosed as NSIP and COP ([kappa] = 0.38 and 0.37, respectively). In this study, the radiologists also noted their level of confidence in making the diagnosis. Notably, interobserver variability was substantially higher in cases diagnosed with low confidence than in those in which the confidence was high ([kappa] = 0.28 and 0.68, respectively). These findings also show that HRCT scans will frequently lead to the misdiagnosis of IIPs and suggest that cases in which diagnoses are made with low confidence may require review by an expert panel.

Finally, interobserver variability among 10 pathologists who were members of the UK Interstitial Lung Disease panel was evaluated retrospectively in 133 lobar specimens obtained from patients with diffuse parenchymal lung diseases. (15) The pathologists were provided only with information about the age and sex of the patient and the site of biopsy, and were asked to assume that infection had been excluded. The pathologists provided their first choice of diagnosis and their level of confidence in that diagnosis to the nearest 5%. Overall, there was only fair agreement among the pathologists in their first choice of diagnosis ([kappa] = 0.38), but it increased somewhat when multiple biopsy specimens were taken ([kappa] = 0.43) and in cases diagnosed with a high level of confidence (ie, > 70% likelihood; [kappa] = 0.50). The level of agreement was marginally higher in cases of UIP (overall, [kappa] = 0.42; those with multiple biopsy specimens, [kappa] = 0.49; in cases with high confidence, [kappa] = 0.58) and considerably higher in cases of sarcoidosis ([kappa] = 0.76, 0.82, and 0.86, respectively). Substantial interobserver variability, however, was seen in the diagnosis of NSIP ([kappa] = 0.29, 0.32, and 0.31, respectively), particularly when it had to be discriminated from UIP. In addition, discrimination between end-stage lung disease and both UIP and NSIP was another important source of disagreement among the pathologists. These findings show that even pathologists who routinely evaluate ILDs often disagree about the histologic diagnosis, even when the diagnosis is based on findings from multiple biopsy specimens. The fact that potentially important clinical and radiologic information was withheld from these evaluators could have confounded the results and underscores the importance of integrating the histologic findings with both the HRCT scan and clinical data when assessing these patients.

ACCURACY OF CLINICAL DIAGNOSES OF IPF AND OTHER ILDs

The accuracy of a clinical diagnosis of IPF and other ILDs was assessed by Raghu and colleagues (16) in a prospective study of 59 consecutive patients who were referred to a tertiary medical center with recognized expertise in the management of ILDs. The clinical diagnosis was made independently by an ILD expert after a thorough evaluation of clinical findings including HRCT scanning and bronchoscopy. A radiologic diagnosis was also made independently by a thoracic radiologist after a review of the chest radiograph and HRCT scan. All patients subsequently underwent surgical lung biopsy, and then the clinical and radiologic diagnoses were compared to the histologic diagnosis. The diagnosis of IPF was made correctly on the basis of the clinical evaluation and HRCT scan assessment in 62% and 76% of cases, respectively. Although the specificity of making a diagnosis of new-onset IPF on the basis of clinical or radiologic findings was very high, the sensitivity was low (Fig 1). Notably, subpleural honeycomb cysts were identified by HRCT scans in only 9 of the 29 patients with UIP histology, illustrating that histologic evaluation is more sensitive than radiologic analysis in IPF. This study shows that new-onset IPF can be diagnosed by clinical and radiologic criteria, hut the diagnosis will be missed in about one third of the cases despite evaluation by experts in the disease.

The value of clinical and radiologic findings in the diagnosis of 91 patients who were suspected of having IPF was assessed in a prospective blinded study by Hunninghake and colleagues. (17) A clinical core group of three pulmonologists reviewed clinical data provided by eight referring centers including chest radiographs and HRCT scans. A core group of four chest radiologists reviewed only the HRCT scans. Each member of the group independently rated their certainty of the diagnosis of IPF (ie, certain, uncertain, or unlikely) and provided an overall diagnosis. After the patients underwent surgical lung biopsy, a core group of three pathologists independently evaluated the pathology slides without seeing any of the clinical information. Overall, the probability of agreement within the pathology core group for a diagnosis of IPF as well as for a specific diagnosis of an ILD was higher than that in the clinical or radiology core groups. Of the 91 patients, 54 received a pathologic diagnosis of UIP/ IPF. The clinical and radiologic core groups were accurate in making the IPF diagnosis in 77% and 75% of the cases, respectively, but in cases with a confident diagnosis (ie, excluding those rated as "uncertain"), the accuracy increased to 86% and 90%, respectively (Table 2). In both analyses, the accuracy was higher among members of the clinical and radiologic core groups than that for the referring center. This finding shows that the value of clinical and radiologic criteria in diagnosing IPF largely depends on who is performing the evaluation, and supports the role of expert panels in reviewing eases and helping in the diagnosis. This ]nay be particularly relevant in multicenter clinical trials. Moreover, this study indicates that surgical lung biopsy is most helpful when clinical and radiologic data result in an uncertain diagnosis or when patients are thought to not have IPF.

Flaherty and colleagues (18) considered the impact of histologic variability on the diagnosis of UIP and NSIP in 109 patients with suspected IPF on the basis of clinical and radiologic findings. Surgical lung biopsy specimens were collected from multiple lobes and then were evaluated by three expert pulmonary pathologists. Patients with histologic patterns other than UIP and NSIP were not included in the analysis. Overall, 51 patients had UIP in all lobes evaluated (concordant UIP), and 30 patients had NSIP in all lobes, including 2.5 with a fibrotic pattern and 5 with a cellular pattern. Notably, the remaining 28 patients (26% of the total) had UIP in at least one lobe and NSIP in another lobe (discordant UIP). Patients with NSIP had significantly better survival times than those with concordant or discordant UIP (p < 0.0001), whereas both groups of UIP patients showed similar survival rates (p = 0.16). Similar results were reported in a study by Monaghan and coworkers, (19) who evaluated 64 patients who had undergone multiple biopsies and had a histologic pattern of either UIP or NSIP. Discordant UIP was identified in 13% of the cases, and again this group had significantly poorer survival times than those with NSIP (p < 0.04), but similar to those with concordant UIP (p = 0.48). Taken together, these studies show that histologic variability is common in patients with IIPs and can lead to a misdiagnosis in up to one quarter of eases if only one lobe is sampled. Moreover, the results provide the basis for using UIP as the default diagnosis in patients with UIP in at least one lobe, because they have the same poor prognosis as those with concordant UIP. Therefore, biopsy specimens from more than one lobe should be obtained.

On gross examination, there is considerable variation between IPF patients in the extent of honeycomb change. Some patients may have most of their lungs replaced by honeycomb change and have little normal lung remaining, others may show a classic pattern with honeycomb change adjacent to areas of normal lung, and still others may have an entire lobe of honeycomb change and another lobe with normal lung tissue. These differences raise questions about the lung area that should be biopsied, as it can significantly impact the histologic diagnosis. Ideally, biopsy specimens should be taken from the junction between abnormal and normal lung tissue.

SUMMARY

Surgical lung biopsy is needed for the definitive diagnosis of UIP/IPF, but clinical and radiologic criteria can establish a likely diagnosis in the absence of a biopsy specimen. The predictive value of these criteria largely depends on the expertise of the clinician and radiologist. Studies (13,16,17) have shown that clinical criteria and HRCT scans provide an accurate diagnosis in up to three quarters of cases of ILDs or suspected IPF. Nevertheless, considerable interobserver variability is seen in the clinical, radiologic, and histologic diagnoses, even when performed by experts in the field, (13-15) Interobserver variability is improved, however, in cases in which the diagnosis is made with a high degree of confidence. Moreover, the accuracy of the diagnosis made by a core of experts is typically higher than that made by a referring center. (17) This suggests that cases with an uncertain diagnosis should be referred to a panel of experts. The decision on whether or not to perform a surgical lung biopsy is difficult. Clearly, a biopsy is not needed in all patients. Therefore, understanding the current data and clinical recommendations will assist in making an informed decision on whether or not to obtain a biopsy specimen in your patients.

REFERENCES

(1) American Thoracic Society, European Respiratory Society. American Thoracic Society/European Respiratory Society International Multidisciplinary Consensus Classification of the Idiopathic Interstitial Pneumonias. This joint statement of the American Thoracic Society (AT8), and the European Respiratory Society (ERS) was adopted by the AT8 board of directors, June 2001 and by the ERS Executive Committee, June 2001. Am J Respir Crit Care Med 2002; 165:277-304

(2) Bjoraker JA, Ryu JH, Edwin MK, et al. Prognostic significance of histopathologic subsets in idiopathic pulmonary fibrosis. Am J Respir Crit Care Med 1998; 157:199-203

(3) Travis WD, Matsui K, Moss J, et al. Idiopathic nonspecific interstitial pneumonia: prognostic significance of cellular and fibrosing patterns: survival comparison with usual interstitial pneumonia and desquamative interstitial pneumonia. Am J Surg Pathol 2000; 24:19-33

(4) American Thoracic Society. Idiopathic pulmonary fibrosis: diagnosis and treatment; international consensus statement-American Thoracic Society (ATS), and the European Respiratory Society (ERS). Am J Respir Crit Care Med 2000; 161:646-664

(5) Katzenstein ALA, Meyers JL. Idiopathic pulmonary fibrosis: clinical relevance of pathologic classification. Am J Respir Crit Care Med 1998; 157(suppl):1301-1315

(6) King TE Jr. Bronchiolitis obliterans. In: Schwarz MI, King TE Jr, eds. Interstitial lung diseases. 3rd ed. Hamilton, ON, Canada: BC Decker, 1998; 645-684

(7) Rena O, Casadio C, Leo F, et al. Video thoracoscopic lung biopsy in the diagnosis of interstitial lung disease. Eur J Cardiothorac Surg 1999; 16:624-627

(8) Bensard DD, McIntyre RC Jr, Waring BJ, et al. Comparison of video thoracoscopic lung biopsy to open lung biopsy in the diagnosis of interstitial lung disease. Chest 1993; 103:765-770

(9) Ferson PF, Landreneau RJ, Dowling RD, et al. Comparison of open versus thoracoscopic lung biopsy for diffuse infiltrative pulmonary disease. J Thorac Cardiovasc Surg 1993; 106:194-199

(10) Molin LJ, Steinberg JB, Lanza LA. VATS increases cost in patients undergoing lung biopsy for interstitial lung disease. Ann Thorac Surg 1994; 58:1595-1598

(11) Carnochan FM, Walker WS, Cameron EW. Efficacy of video thoracoscopic lung biopsy: an historical comparison with open lung biopsy. Thorax 1994; 49:361-363

(12) Vansteenkiste J, Verbeken E, Thomeer M, et al. Medical thoracoscopic lung biopsy in interstitial lung disease: a prospective study of biopsy quality. Eur Respir J 1999; 14:585-590

(13) Peckbam RM, Shorr AF, Helman DL Jr. Potential limitations of clinical criteria for the diagnosis of idiopathic pulmonary fibrosis/cryptogenic fibrosing alveolitis. Respiration 2004; 71: 165-169

(14) Aziz ZA, Wells AU, Hansell DM, et al. HCT diagnosis of diffuse parenchymal lung disease: interobserver variation. Thorax 2004; 59:506-511

(15) Nicholson AG, Addis B J, Bharucha H, et al. Inter-observer variation between pathologists in diffuse parenchymal lung disease. Thorax 2004; 59:500-505

(16) Raghn G, Mageto YN, Lockhart D, et al. The accuracy of the clinical diagnosis of new-onset idiopathic pulmonary fibrosis and other interstitial lung disease. Chest 1999; 116:1168-1174

(17) Hummingdale GW, Zimmerman MB, Schwartz DA, et al. Utility of a lung biopsy for the diagnosis of idiopathic pulmonary fibrosis. Am J Respir Crit Care Med 2001; 164: 193-196

(18) Flaherty KR, Travis WD, Colby TV, et al. Histopathologic variability in usual and nonspecific interstitial pneumonias. Am J Respir Crit Care Med 2001; 164:1722-1727

(19) Monaghan H, Wells AU, Colby TV, et al. Prognostic implications of histologic patterns in multiple surgical lung biopsies from patients with idiopathic interstitial pneumonias. Chest 2004; 125:522 526

* From the Department of Pathology and Laboratory Medicine, David Geffen School of Medicine at UCLA, Los Angeles, CA. Reproduction of this article is prohibited without written permission from the American College of Chest Physicians (www.chestjournal. org/misc/reprints.shtml).

Correspondence to: Michael C. Fishbein, MD, Professor of Pathology and Medicine, Chief, Autopsy Service David Geffen School of Medicine, UCLA, Los Angeles, CA 90095; e-mail: mfishbein@mednet.ucla.edu

COPYRIGHT 2005 American College of Chest Physicians
COPYRIGHT 2005 Gale Group