Guiding the Pulmonologist's Hand: What They Need to Know About Lung Pathology and What is Lost in Translation

Acute respiratory distress syndrome (ARDS) is a leading cause of morbidity and mortality in pulmonary medicine practice. It is also the largest source of economic strain on the budget of hospitals because patients often require ventilator and hemodynamic support and prolonged hospitalization.¹  Although ARDS is a syndrome that can be seen because of a variety of causes, most cases show the pathology of diffuse alveolar damage (DAD).²  The indications for biopsy vary among centers and are generally limited to cases in which the diagnosis or causation is in doubt. Whereas the diagnosis can be made by transbronchial biopsy, few patients are sufficiently clinically stable to tolerate that procedure, and video-assisted thoracoscopic surgery (VATS) biopsy is generally required. Assuming that the diagnosis is clear, does pathologic assessment yield information that can assist the clinician with decision-making?

One important point has to do with hemodynamic management. There is a tendency on the part of pulmonologists to think of ARDS as a “wet lung” and, consequently, about how to mobilize free water via diuresis while monitoring central venous or pulmonary arterial pressures. However, pathologists know from the appearance of the lungs in DAD at autopsy that the pulmonary edema is firmly bound. The lungs generally show a combined weight of more than 2000 g, which, in part, reflects the proliferation of cellular elements in the fibroproliferative phase of disease, but most of the lung weight can be accounted for by the presence of water entrapped within a fibrin gel.³  Procoagulation is a critical factor in DAD, as evidenced by the presence of fibrin-rich hyaline membranes, intra-alveolar fibrin, and in situ vascular thrombosis. Mobilization of water from the lungs, therefore, ultimately depends on the efficacy of fibrinolytic pathways and effective lymphatic drainage.

Most patients with ARDS who have a biopsy are in the fibroproliferative phase of the disease, which begins approximately 72 hours following an inciting insult. The pattern of fibrosis is predominantly intra-alveolar and septal and resembles confluent “fibroblastic foci.” The degree of intra-alveolar fibrin varies, in part, on the timing of the biopsy and the host response. Approximately 20% of lungs show a substantial degree of obliterative bronchiolitis (OP) as part of the biopsy (DAD-OP), and it is important to distinguish this from organizing pneumonia.⁴  Although the literature suggests that organizing pneumonia can be a cause of ARDS, in my experience these cases invariably prove to be DAD with prominent elements of organizing pneumonia. In these cases, hyaline membranes can be identified, and multiple foci of intra-alveolar fibrin are present. The presence of in situ intravascular thrombi is another important clue to the presence of extensive lung injury. A diagnosis of OP in a patient with diffuse bilateral infiltrates requiring ventilation is likely incorrect and should trigger a careful search for evidence of acute alveolar injury.

Why does this matter? Although it is too early to generalize accurately from preliminary data, it does appear that patients who have DAD-OP may represent a subset of patients with ARDS who have a better prognosis and may respond, like OP alone, to corticosteroids. The reason for this is uncertain, but these cases show a significant difference in luminal patency lymphatic channels compared with patients with DAD, and this may indicate an advantage with respect to lymphatic drainage.⁴ 

What other information can be extracted from the biopsy of patients with ARDS? In recent years, there has been an increased recognition that DAD can represent a form of accelerated, chronic interstitial pneumonia. Whereas usual interstitial pneumonitis (UIP), nonspecific interstitial pneumonitis (NSIP),^5,6  and asbestosis⁷  have been shown to be substrates for accelerated disease, it has also been my impression that OP and smoking-related interstitial fibrosis cases are also at risk for this complication, a hypothesis that requires further investigation. In some cases of the accelerated phase, the pathology is that of DAD; however, both DAD-OP and pure OP may be seen, and prognosis appears, again, to depend on the relative amounts of OP versus fibrin in the biopsy.⁸ 

The key in establishing the diagnosis of an accelerated phase of disease is to recognize the presence of a preexisting fibrotic disorder. Histologically, such a disorder may be identified in the periphery of the lung, based on the presence of dense, hyalinized, type-I collagen; smooth muscle hyperplasia; and bronchiolectasis. All lobes should be sampled, and determinations based on biopsies of the right middle lobe or lingula are best avoided. Once histologic evidence of chronic scarring is identified, a prompt re-examination should be made of previous chest imaging and pulmonary function tests to determine whether a preexisting, interstitial abnormality may have been present. Currently, there are no reliable statistics, to my knowledge, concerning the frequency of accelerated phase in occult chronic interstitial disease.

A potentially confounding issue is the diagnosis of acute interstitial pneumonitis.^9,10  These cases generally have a subacute presentation and show the pathology of the proliferative phase of DAD, although there are cases that show overlapping DAD-OP. Because of the stuttering, but progressive, course of this disorder, areas of established, chronic changes may be found that mimic those of the accelerated phase of chronic interstitial lung disease. For this reason, clinical history, radiographic, and functional progression are critical pieces of information in determining both diagnosis and outcome.

Repetitive bouts of acute lung injury should raise the possibility of gastroesophageal reflux disease, especially if the findings are most pronounced in the lower lobes of the lung or in dependent areas of the lungs. A history of reflux symptoms, food or ethanol ingestion before bedtime, diminished neurologic status, laryngeal dysfunction, or a computerized tomography appearance of a dilated, flatulent, or fluid-filled esophagus or a hiatal hernia can assist in determining whether aspiration is the likely cause of ARDS. Therefore, both pathologists and clinicians need to recognize that gastroesophageal reflux disease can be a dynamic, functional syndrome, which is exacerbated by foods and emotional factors, so that the absence of reflux on barium swallow exam or pH probe analysis, in a single determination, does not exclude this disorder. Indeed, there is a growing consensus amongst pulmonologists that gastroesophageal reflux disease may be an important factor in the accelerated phases of many cases of chronic interstitial pneumonia.^11,12 

Hypersensitivity pneumonitis can also account for recurrent interstitial disease, although there is generally upper lobe predominance. The computerized tomography scan shows centrilobular predominance, and air-trapping can be seen on expiratory phase by high-resolution computerized tomography.

The presence of centrilobular pulmonary inflammation and fibrosis on a biopsy has a limited differential, and even in small biopsies, important information can often be relayed to the clinician. The most common causes include airway insults from cigarette smoking or other inhaled toxins; aspiration; Langerhans cell histiocytosis; rheumatoid disease, in particular Sjögren syndrome; and hypersensitivity pneumonitis. The entity of airway-centered interstitial fibrosis that is characterized by bronchiolocentric airway epithelial metaplasia (lambertosis) has not been adequately characterized but appears in most cases to have a guarded prognosis.^13,14 

The literature suggests that the diagnosis of UIP carries an extremely poor prognosis.¹⁵  Although that is true, a subset of patients with the diagnosis of UIP/idiopathic pulmonary fibrosis may show clinically protracted disease with slow devolution of their diffusing capacity for carbon monoxide (DlCO), and some do appear to stabilize, at least for some time, with the administration of corticosteroids. Nevertheless, it is important to be conservative in making this diagnosis because of its implications and because of the current aura of therapeutic nihilism that surrounds it. Patient age is a useful element in arriving at the diagnosis in those biopsies in which the findings are suggestive, but not clearly diagnostic, because the disease is far more common in the elderly.¹⁶  Heterogeneity in the biopsy findings in this disorder is well recognized, but not necessarily widely known, by clinicians.^17–21  For this reason, when I encounter a case that might best be characterized as a fibrotic form of NSIP in an elderly patient, I warn my clinical colleagues that this may well represent UIP that has been inadequately sampled.

In many cases, the diagnosis of idiopathic pulmonary fibrosis is currently made by computerized tomography scan based on the presence of bibasilar disease with architectural distortion and extensive honeycomb changes. The sine qua non for diagnosis is the presence of end-stage, honeycomb lung, together with active fibroblastic foci, preferably at the leading edge of the scarring, with preeminent importance on the former. The diagnosis is often apparent at low power because of the characteristic subpleural patchy distribution and distortion of the lung architecture.

The diagnosis, however, is only one important feature that should be included as part of the final pathology report, in all cases. Simply establishing a diagnosis tends to short circuit the thinking process in medical lung disease, or it can lead to an overactive imagination on the part of the clinicians caring for the patient. For this reason, all the following information should be conveyed: the diagnosis, the frequency of fibroblastic foci in the biopsies, the degree of honeycomb change, the extent and character of the pulmonary inflammation, and the grade of pulmonary hypertensive arteriopathy, when present. Even if a definitive diagnosis cannot be achieved, this information will generally help the clinician determine the best course of treatment or further testing.

Certain collagen vascular disorders, including rheumatoid disease, scleroderma-70 disease (antitopoisomerase I), and scleroderma can show patterns of UIP, with possibly improved responses to anti-inflammatory interventions.²²  In some cases, a simple screen for antinuclear antibody and rheumatoid factor can miss a more esoteric diagnostic marker of an autoimmune process. It is also important to recognize that low titers of antinuclear antibody are common in UIP and do not indicate an associated collagen vascular disorder.

Chronic hypersensitivity pneumonitis is suggested by the presence of centrilobular inflammation; poorly formed, and often rare, nonnecrotizing granulomas in the biopsy; and the suggestion of alternating areas of air-trapping and atelectasis, ideally together with a known sensitizing agent or serologic evidence to support the presence of an inciting antigen.⁵  Upper lobe predominance is a useful feature but is not always present. Evidence of air-trapping during expiration on a high-resolution computerized tomography scan also suggests the small airways abnormality that is part of this disorder. However, the prognosis by the time this pattern of disease has developed appears to be similar to that of idiopathic pulmonary fibrosis, even if an offending agent is identified and removed or corticosteroids initiated.

This disorder is generally suggested by clinical and radiographic evidence of interstitial infiltrates with ground-glass opacities. Honeycomb lung should not be present, although it can rarely be a minor feature and must be distinguished from proximal areas of traction bronchiectasis, which can be difficult both radiographically and by histopathology. In my opinion, unless contraindicated, VATS biopsy should be pursued in all cases of presumed NSIP. Although it may be argued that it may not influence treatment, when NSIP does not respond well or recurs, it is helpful to have an established diagnosis from which to guide treatment. Most cases show either cellular, or mixed cellular, fibrotic patterns of relatively homogeneous disease.²⁰  The presence of lymphoid follicles or increased plasma cells, as well as vascular and pleural inflammation, points toward the presence of an underlying collagen-vascular disease. Hypersensitivity pneumonitis and drug reactions also need to be considered in the differential diagnosis. If prominent areas of desquamative interstitial pneumonia-like changes are seen with macrophages that lack anthracotic smoking pigment, one should consider the possibility of an underlying myelodysplastic syndrome or drug-induced pneumonitis. Cases of NSIP with evidence of desquamative interstitial pneumonia-like change and foci intra-alveolar proteinosis should be examined by electron microscopy, immunohistochemistry, and genetic testing to exclude an adult form of congenital surfactant protein deficiency.²³ 

The fibrotic pattern of NSIP appears to be less frequently diagnosed than in past years. Part of this is due to an increased awareness of the heterogeneity of UIP in biopsy samples. In addition, the recognition of smoking-related fibrosis may have accounted for an uncertain percentage of cases with NSIP, fibrotic type.

This disorder is far more common than previously recognized.¹⁴  It is often present in tumor resections of patients with lung cancer, and its presence raises the question as to whether it may be an independent factor for developing smoking-related lung cancer.¹⁹  The appearance of the collagen is characteristic; it is predominantly subpleural, paucicellular, and “ropey” appearing, and it may be associated with emphysema. In some areas, the scarring can assume a stellate appearance that has traditionally been attributed to “burnt-out” Langerhans cell histiocytosis (eosinophilic granuloma). The clinical implications of these old scars are usually minimal, and unless there is evidence of active Langerhans cell histiocytosis in other areas of the biopsy or a centrilobular nodule and cystic changes on computerized tomography scan to suggest active disease, I will not conjure up the diagnosis of Langerhans cell histiocytosis because pulmonologists tend to have a different idea as to what that disease represents.

Another area of interest with respect to smoking-related fibrosis is its potential impact on the clinical diagnosis of pneumoconiosis, especially asbestosis. Current American Thoracic Society criteria require a b-reading of profusion 1/0 in the presence of a history of asbestos exposure with appropriate latency to diagnose asbestosis.²⁴  This minor degree of radiographic abnormality reflects the exclusion of most asbestos-containing materials from the workplace and stricter governmental regulation of permissible limits. However, because a sizable proportion of patients with asbestos exposure were current or past cigarette smokers, it is currently virtually impossible to predict, based on radiographic appearance, whether parenchymal infiltrates are due to smoking or asbestos. It is possible that both contribute to the development of interstitial fibrosis in the absence of histologic evidence of pulmonary asbestosis. For this reason, it is currently my opinion that the only reliable way to definitively establish a diagnosis of pulmonary asbestosis in a smoker with minimal pulmonary fibrosis on chest radiographs is to biopsy the lung to confirm the presence of both fibrosis and asbestos bodies.

Idiopathic organizing pneumonia or bronchiolitis-obliterans with organizing pneumonia is one of the most common changes encountered on lung biopsy. It is also perhaps one of the least specific.²⁵  Clinically, patients with organizing pneumonia may have cough, dyspnea, and pulmonary consolidations or, alternatively, may be asymptomatic, presenting with either a solitary pulmonary nodule or multifocal disease. The biopsy should ideally be made by VATS but can at times be made reliably by the transbronchial approach. The finding of a pattern of organizing pneumonia requires a diligent pathologic and clinical search to exclude infection, aspiration, drug-induced injury, collagen vascular disease, and renal failure. Most cases show intraluminal obliteration of branching respiratory bronchioles with a surrounding area of interstitial pneumonitis. In some cases, the NSIP pattern may predominate, and those cases have been referred to as bronchiolitis interstitial pneumonia.²⁶ 

The fibroplasia of organizing pneumonia may include intra-alveolar fibrin in the biopsy, but hyaline membranes should not be present. A subset of patients with organizing pneumonia show a prominent degree of intra-alveolar filling by fibrin and this has been termed acute fibrinous pneumonia-organizing pneumonia. In such cases, what needs to be determined is the extent of disease and the physiologic status of the patient. In the intensive care unit setting, most of these cases appear to be a variant of acute lung injury and have a guarded prognosis. However, in practice, it appears that most have relatively localized disease and show a good initial response to corticosteroids, much like the nonfibrinous variant of organizing pneumonia. We have recently demonstrated that the prognosis with respect to the recurrence of treated organizing pneumonia varies inversely with the amount of fibrin in the biopsy, and I have, for some time now, counseled my colleagues to maintain this subset of patients on prolonged courses of corticosteroids. The presence of alveolar fibrin in the biopsy correlates directly with prognosis and is independent of age, smoking status, or the presence of gastroesophageal reflux disease.²⁷ 

Usually a straightforward disease pathologically, some cases of sarcoidosis can be challenging both clinically and pathologically. Hilar and mediastinal adenopathy in a young, asymptomatic woman may be diagnosed and followed radiographically. The pathology of sarcoidosis can show a variety of features that can lead to concern for another diagnosis. I confidently diagnose the disease in lymph node biopsies when granulomas are nonnecrotizing, compact, and show evidence of hyalinizing fibrosis. However, focal fibrinoid necrosis is common, especially in patients in an active phase of disease. Extensive fibrinoid necrosis can occur in patients with Löfgren syndrome, and most pulmonary pathologists can recount cases where the degree of central necrosis in granulomas appeared most consistent with infection but proved to be sarcoidosis based on microbiologic testing. In lymph nodes, one must also be careful to exclude granulomatous forms of Hodgkin disease, which can mimic both sarcoidosis and mycobacterial infection but, on detailed examination, reveals diagnostic Reed-Sternberg cells focally in the tissue.

As sarcoidosis tends to find a conduit along the lymphatic pathways of the lung, the diagnosis can be reliably established by sampling via transbronchial biopsy or by endobronchial ultrasound-guided biopsies. Occasional cases of hypersensitivity pneumonitis and lymphoid interstitial pneumonia can mimic the histologic appearance of sarcoidosis, so that history, serologic testing, and radiographic appearance may be important in making these distinctions. Pulmonary beryllium disease, although rare, is indistinguishable from sarcoidosis, and clinicians sometimes need to be reminded to consider that possibility. The diagnosis may be inferred by work with beryllium, a positive lymphocyte transformation test to beryllium salts in blood or lavage lymphocytes, or by tissue quantitation.²⁸ 

Two other points are worthy of mention with respect to sarcoidosis. In 35 years of practice, I have only once seen a positive histochemical stain for mycobacteria or fungi in compact, nonnecrotizing granulomas, and this accounts for thousands of biopsies. Although still routinely ordered, in most settings, the cost effectiveness of this approach is exceedingly low.

Resections for lung cancer not infrequently show either small, nonnecrotizing granulomas or similar findings in the lung and draining lymph nodes. In my reports, I always indicate that they most likely reflect a hypersensitivity response to tumor, unless their frequency or a clear perilymphatic distribution suggests concomitant sarcoidosis. For many clinicians who are not conversant with the subclinical changes that can occur in lung biopsies, the diagnosis of nonnecrotizing granulomas can trigger an unnecessary workup for sarcoidosis.

There are many areas that I have left untouched in this short essay. There is, however, one critical point I want to convey. My pulmonology colleagues are outstanding clinicians, but few of them have a deep understanding of lung pathology. This can result in a certain diagnostic tension that most pathologists have experienced. There are times when the lung pathology does not fit well into their conception of what the most likely diagnosis is for their patients. Most nonneoplastic diagnoses should generate a differential clinical diagnosis, and it is here that clinicopathologic correlation is most helpful in ruling in or out various possibilities. At times, the clinicians are right, and the diagnostic pathologist must carefully entertain their opinions. Often, however, they are not, and it behooves pathologists not to be unduly influenced by their ideas. One must, therefore, endeavor to make certain that the clinician fully recognizes the implication of the pathologic findings, even if it means reigning in their active imaginations because failure to do so may relieve tension but can compromise clinical care.

Certain terms used by pathologists trigger difficulties in translation. For example, pathologists have a rather broad understanding of what they mean by granulomatous inflammation, ranging from compact sarcoidal granulomas to diffuse polymorphic inflammation with macrophages, lymphocytes, and fibroblastic inflammation. The distinction can be lost on clinicians, who, as a consequence, may reflexively include Wegener or Churg-Strauss granulomatosis in their differential diagnosis in response to hearing the term granuloma.

Finally, those who practice medical pathology should take a lesson from what is occurring in the domain of personalized therapies for neoplasia. The medical lung biopsy, even in the absence of genetic information, continues to hold important and untapped details that may be pertinent to diagnosis, prognosis, treatment, and pathogenesis. Pathologists should play an active role in any research endeavors in these disorders if progress is to be made based on the elements of observable disease.