CHAPTER 2 Respiratory tract

Thomas E. Giles, Julie McCarthy, Winifred Gray

Chapter contents

Preparatory techniques and diagnostic applications

Criteria for assessing adequacy of samples

Accuracy of Cytology for Tumour Diagnosis and Management

The authors are pleased to acknowledge their indebtedness to Dr G. Sterrett, Dr F. Frost and the late Dr Darrel Whitaker, authors of the lung tumour section in the two previous editions of Diagnostic Cytopathology. Their work forms a substantial proportion of the current chapter and it is an honour to have been able to use their material. We are also indebted to Dr G. Kocjan for her additional help in the completion of the chapter.

Introduction

The importance of cytological techniques for investigation of respiratory conditions has been recognised since the earliest days of clinical cytology,¹ as described in the Introduction to this book. The rise in incidence of bronchogenic carcinoma throughout the twentieth century, related to smoking, ensured that examination of sputum and bronchial secretions for malignant cells became a major part of the workload of all routine cytology laboratories.²

Developments in sampling techniques, in particular the advent of fibreoptic endoscopic techniques in the 1960s and the more recent use of fine needle aspiration (FNA) for obtaining material, have changed the practice of respiratory tract cytology, although not completely supplanting more traditional methods. Bronchial brushings and lavage procedures usually yield better diagnostic material than simple exfoliative sampling and can be used for sequential studies.^3,⁴ FNA cytology of the lung in many ways paralleled exfoliative cytology of the respiratory tract, initial reports of detection of lung cancer occurring as early as 1886.⁵ Radiological imaging allows FNA sampling of lesions at virtually any site within the thorax and has improved the safety of these procedures.^6,⁷ The British Society for Clinical Cytology (BSCC) has issued guidelines on the procurement, assessment for adequacy and examination of exfoliative samples and endoscopic specimens⁸ as has the Papanicolaou Society of Cytopathology in the United States.⁹ The last few decades have seen ample demonstration of the sensitivity and predictive value of cytodiagnosis of lung tumours, an acceptance of all cytological modalities as a basis for management, and gradual extension of the range of diagnoses to virtually all neoplastic, and some of the non-neoplastic, processes affecting the lung and mediastinum.¹⁰^–¹⁵

Preparatory techniques and diagnostic applications

Upper respiratory tract

Spontaneous nasal secretions may be collected on a moistened swab, or with a gently abrasive rhinobrush,¹⁶ or simply by nose blowing. May-Grünwald Giemsa (MGG) staining of air-dried preparations is used to assess eosinophil levels in cases of allergic rhinitis, challenging with inhaled allergens to identify the underlying cause.¹⁷ Increased exfoliation of epithelial cells has also been noted in cases of nasal hyper-reactivity.¹⁸

Nasopharyngeal sampling by swab has been used for rapid diagnosis of nasopharyngeal carcinoma, and has been proposed as an effective screening method in areas with a high incidence of this tumour.¹⁹ Pharyngeal and laryngeal samples are usually collected under direct vision by direct scraping of the lesion. FNA is appropriate if there is an intact mucosa covering a tumour such as lymphoma. Imprint cytology of biopsies from laryngeal and pharyngeal biopsies has proved useful, providing a rapid accurate diagnosis for clinical management and excellent correlation with histology.²⁰

Elucidation of pulmonary infiltrates and identification of opportunistic infections in immunocompromised patients are important applications of this procedure as described in Chapter 16. Part of the sample should be submitted for microbiological culture. Papanicolaou staining is combined with other stains for opportunistic infections. Thin layer methods are less suitable as both organisms and inflammatory cells may be selectively lost in the preparation, a risk that can be circumvented by dividing the sample between the commercial cell fixative and saline. This will then allow MGG staining for cell differential counts.

BAL fluid examination is one of the initial procedures in the diagnosis of interstitial lung disease.²⁷ Differential counts on the inflammatory cell population in lavage fluid have been shown to reflect the histological findings in cases of pulmonary fibrosis and non-infectious granulomatous lung disease and serial lavage has found a place in monitoring progress of these conditions.²⁸ BAL also has a limited role in diagnosis of peripheral lung cancer.

Fine needle aspiration (FNA)

FNA material may be obtained either by a transbronchial or transthoracic approach. The procedure is of particular value for tumour diagnosis and staging if bronchoscopy has failed to achieve a diagnosis or is inappropriate. The introduction of ultrasound guided imaging has improved the accuracy of sampling, while use of a fine gauge needle (19–22G) makes the procedure safe and well-tolerated.⁷ Aspiration is performed at bronchoscopy using a flexible metal needle to which suction is applied.⁷

Recent advances in technology have led to the development of the combined endoscope/ultrasound probe which allows direct, real-time visualisation of the needle during aspiration (EBUS-FNA). This is particularly useful in the transcarinal aspiration of mediastinal structures but is of less value for certain pulmonary lesions due to intervening air within the lung tissue.²⁹^–³¹ ⁵⁷¹

The advantages and limitations of FNA in lung tumour diagnosis have been highlighted ³¹^–³³ and the BSCC has issued general guidelines for optimising its use.³⁴ Air-dried and wet-fixed slides can be prepared; spare material obtained by rinsing the needle in normal saline or tissue culture medium can be processed as a cell block or by cytospin for other stains and for immunocytochemistry.³⁵ Alternatively, the entire sample may be processed for liquid-based cytology. Electron microscopy, tumour proliferation studies and cytogenetic analysis are among the additional procedures that can be performed on FNA material.

Pneumothorax is a potential complication of FNA, although only 4–10% of these patients require a chest drain.^36,³⁷ Patients with emphysema are at greater risk.³⁸ Because of the risk of complications, FNA is contraindicated in unconscious or uncooperative patients and in those with respiratory failure, haemorrhagic diathesis or intractable coughing.

General respiratory tract findings

The respiratory system includes the nasal passages, sinuses and nasopharynx, the oropharynx and larynx, trachea, bronchi and bronchioles, and the air spaces beyond. Transportation of gases is the primary function of the upper respiratory tract and airways, but there is also an important role in warming and moistening inspired air, removing particulate material and providing an initial immunological defence against inhaled microorganisms. Gaseous exchange is carried out within the alveoli and other complex activities take place in the lung parenchyma, including further pulmonary defence mechanisms, some endocrine functions and maintenance of homeostasis.

Not surprisingly, there are many variations in cell structure throughout the respiratory system, and their delicate balance is frequently disturbed by disease. A comprehensive knowledge of the normal findings is therefore necessary to understand the pathological changes encountered in cytological specimens.³⁹

Normal histology of the respiratory tract

Two different types of epithelium form the mucosa of the respiratory tract, their exact distribution varying with age. Stratified squamous epithelium covers areas liable to abrasion, such as the nasal vestibule, nasopharynx, lingual surface of the epiglottis and the vocal cords. Elsewhere a complex layer of glandular cells is found. Squamous mucosa is composed of basal, parabasal, intermediate and superficial cells, and is not keratinised in health. Beneath the basement membrane of this epithelium lies a fibrocollagenous stroma containing blood vessels, lymphatics, nerves and seromucinous glands (Fig. 2.1). Inflammatory cells of the immune system, mainly lymphocytes, plasma cells and macrophages, are also seen migrating into the overlying epithelium. In strategic areas lymphoid cells aggregate into organised tissue masses forming the tonsils and adenoids.

Fig. 2.1 Normal bronchial wall showing the lining mucosa resting on fibrocollagenous submucosa containing seromucinous glands, blood vessels and lymphatic channels (H&E).

The bronchial tree and remainder of the upper airways are lined by specialised respiratory epithelium (Fig. 2.2). This consists of a pseudostratified layer of ciliated tall columnar cells interspersed with mucin secreting goblet cells, which have microvilli on their luminal surfaces. There are approximately five ciliated cells for each goblet cell. Mucin from the goblet cells coats the airways with a sticky layer within which inhaled particles, organisms and cell debris are trapped. The cilia have a metachronous beat which sweeps this material upwards, to be expectorated or swallowed.

Fig. 2.2 Normal respiratory mucosa of bronchus. Note the multilayered pseudostratified columnar epithelium composed mainly of ciliated cells with occasional goblet cells. A distinct single layer of reserve cells can be seen resting on the basement membrane. Deep to this the submucosa includes a few capillaries, lymphatics and inflammatory cells (H&E).

Two further cell types are present in respiratory epithelium. Small reserve cells rest on the basement membrane, forming an undifferentiated stem cell population from which regeneration of bronchial mucosa takes place after injury. Inconspicuous round cells with neuroendocrine properties are also found situated towards the basement membrane. Known as Feyrter or K (Kultschitzsky) cells, they are most numerous in the smaller bronchi where they are grouped around capillaries and nerve fibres forming neuroepithelial bodies. They contain neurosecretory granules producing locally active polypeptide hormones, and belong to the APUD (amine precursor uptake and decarboxylation) cell system.

Bronchioles, the first branches of bronchi without cartilaginous support in their walls, are lined by a single layer of non-ciliated columnar cells interspersed with a few goblet cells. In addition there are tall columnar cells, the Clara cells, producing surfactant. Terminal bronchioles are lined by low columnar epithelium and are involved solely in air conduction. They are continuous with respiratory bronchioles, which mark the commencement of gaseous exchange. Here the lining becomes cuboidal, merging with flattened epithelial cells in the alveolar ducts. These lead into rotunda-like spaces called alveolar sacs. The periphery of each sac is partitioned into alveoli, the main site of gaseous exchange.

The principal cells lining alveoli are known as type I and type II pneumocytes. In addition, there are many macrophages of bone marrow derivation, forming an important component of cytology samples from the lower airways. They adhere to the walls of alveoli, ingesting cellular debris and foreign material, which is then transported to the bronchial tree or to lymphatic channels arising at the level of the terminal bronchioles.

It has been estimated that type I pneumocytes cover approximately 90% of the alveolar wall area, but form only about 40% of the lining cell population. Their cytoplasm is thinly spread out to allow maximal exchange of gas between the alveolar space and the underlying capillaries. Type II pneumocytes comprise 60% of the lining cells numerically, but are bulky and rounded, occupying less than 10% of the alveolar surface area. Their cytoplasm is dense, containing spherical laminated osmiophilic bodies when examined by electron microscopy, composed of the precursors of pulmonary surfactant. These cells are also the progenitors of type I pneumocytes.

General cytological findings in respiratory samples

Cell population

Only a few of the many different cells lining the respiratory tract are seen with any regularity in cytological preparations. The distribution of cells varies considerably with the nature of the sample, but is of importance in assessing specimen adequacy. The appearances to be described for normal and abnormal cells are those seen with Papanicolaou staining unless otherwise specified.

• Squamous cells (Fig. 2.3) are numerically the most common cells in sputum, but are less frequent in other specimens and usually absent from FNA samples. Superficial and intermediate cells predominate. They have small central pyknotic or vesicular nuclei, as seen in squames from other sites

• Bronchial epithelial cells (Fig. 2.4) are profuse in brushings, but less common in sputum as they do not exfoliate readily. Columnar or triangular in shape, the cells lie singly, in short ribbons or in flat sheets which often have a straight ‘anatomical’ edge. They have delicate cyanophilic cytoplasm, bluish grey with MGG stains, tapering at the point of previous anchorage. Nuclei vary considerably in size and shape but are usually basal and rounded or oval (Fig. 2.5A), with open granular or condensed chromatin and a single small nucleolus. Multinucleation may be seen. Cilia are often preserved, arising from a dark stained terminal bar at the broader end of the cell

• Goblet cells are inconspicuous in sputum unless hyperplastic, but are quite often seen in brushings and increase in number with chronic bronchial irritation. They are columnar but are distended centrally by globules of mucin, which overlie or displace the nucleus (Fig. 2.5B)

• Reserve cells (Fig. 2.6) are rarely present in sputum but are sometimes seen in brushings and lavage specimens, mainly in reactive states. They form sheets of small regular cells slightly larger than lymphocytes, with a high nuclear/cytoplasmic ratio, coarse chromatin and a narrow rim of green cytoplasm

• Macrophages (Figs 2.7, 2.8) are the hallmark of a satisfactory sputum sample and are found in the majority of pulmonary specimens. In sputum, the adequacy of a sample is claimed to be directly proportional to the number of alveolar macrophages it contains. Round or oval dissociated cells, they are usually over 10 μm in diameter but vary greatly in size and shape. Their cytoplasm is poorly defined, cyanophilic and often vacuolated. Phagocytosed material, usually carbon, may be present. They have central or eccentric nuclei, rounded or bean-shaped, with coarse chromatin and visible nucleoli. Binucleation is common and the cells may form giant cells with numerous nuclei

• Inflammatory cells, mainly polymorphonuclear leucocytes and lymphocytes, are invariably present in low numbers and are only of diagnostic significance if markedly increased. A preponderance of one inflammatory cell type may, however, be significant, for example, when eosinophils are conspicuous

• Clara cells, Feyrter cells and types I and II pneumocytes are prone to rapid degenerative changes and are not recognisable in respiratory samples unless hyperplastic or neoplastic.

Fig. 2.3 (A) Normal oral squamous cells in a sputum sample consisting mainly of saliva. The cells are of superficial and intermediate type but there is no keratinisation (PAP). (B) Normal mature squames in a liquid based preparation showing nuclear detail (PAP, LBC).

Fig. 2.4 Bronchial epithelial cells in a bronchial brushing sample. These tall columnar cells show the tapering point of anchorage at one end and dark terminal bar, bearing pink cilia at the opposite end of the cell. Nuclei are regular, ovoid in shape and basal in position. Chromatin is finely divided and small nucleoli are visible (PAP).

Fig. 2.5 (A) Bronchial cells with nuclei of variable size and shape, but still within the normal range. Bronchial brushing. (B) Single bronchial epithelial cell and goblet cell, the latter showing distension centrally due to the presence of greenish grey mucin. Bronchial brushing (PAP).

Fig. 2.6 Reserve cells in a crowded group of small cells with dark nuclei and very little cytoplasm, surrounded by columnar epithelial cells. There is a suggestion of nuclear moulding at the right edge of the group. Bronchial brushing (PAP).

Fig. 2.7 (A) Macrophages forming a streak of dissociated cells in sputum. Note the variation in cell size and finely vacuolated cytoplasm with a few particles of ingested carbon. Most of the nuclei are eccentrically placed, varying in shape and size. Nucleoli are visible and several cells show binucleation. (B) Sputum. Multinucleated macrophages aggregated with mononuclear forms (PAP).

Fig. 2.8 Carbon pigment in macrophages varies greatly in amount, but may obscure the nucleus entirely, as in this sputum sample (PAP).

Other components of respiratory samples

Non-cellular material and extraneous elements are seen in many specimens, and should be firmly identified to avoid misinterpretation. The possibilities are virtually unlimited.

• Mucus is almost invariably present in specimens obtained via the airways unless mucolytic agents have been used. It forms a pale translucent background material staining variably, and may be smooth or stringy in texture, usually with streaks of enmeshed inflammatory cells. Densely stained inspissated mucus may obscure cell detail, or give a false impression of cytological abnormality on low magnification. Coils of compressed mucus known as Curschmann’s spirals (Fig. 2.9) are frequently seen in sputum from smokers or patients with obstructive airways disease, especially asthmatics. The spirals are casts of the small bronchioles and vary considerably in structure.⁴⁰

• Inorganic material of various types may be seen and can have diagnostic significance. Charcot–Leyden crystals (Fig. 2.10), derived from the breakdown products of eosinophil granules, appear in conditions evoking pulmonary eosinophilia as orange, yellow or pinkish stained diamond- or needle-shaped crystals.⁴¹ Calcific blue bodies and corpora amylacea are similar in routine preparations. The former consist largely of calcium carbonate and show central birefringence (Fig. 2.11).⁴² Corpora amylacea are non-calcified rounded structures composed of glycoproteins including amyloid. They stain pale pink, are Congo-red positive and exhibit birefringence. Both of these are seen in various chronic lung diseases. Psammoma bodies (calcospherites) are laminated non-refractile calcified concretions sometimes found in the presence of malignancy, although not necessarily closely associated with tumour cells.⁴² Isolated psammoma bodies may be seen in the absence of any tumour formation (Fig. 2.12). Ferruginous bodies (see Fig. 2.64) are formed when filamentous dust particles such as asbestos become coated with protein and iron in the lung parenchyma. They vary from 5 to 200 μm in length, are light brown in colour and stain blue with Perl’s stain for iron.⁴³

Fig. 2.9 Curschmann’s spiral composed of a compressed cast of mucus from a small bronchiole. The form, size and staining of these structures are extremely variable. Sputum (PAP).

Fig. 2.10 Charcot–Leyden crystals are bright orange or yellow and needle shaped. Eosinophils are often present since the crystals form from breakdown products of their granules. Sputum (PAP).

Fig. 2.11 Calcified blue bodies are birefringent inorganic concretions seen here on (A) routine light microscopy and by (B) polarised light. Sputum (PAP).

Fig. 2.12 Psammoma body in sputum showing the typical laminated structure. No evidence of malignancy was found in this patient (PAP).

Contaminants

These may be added to respiratory samples at any stage from collection to microscopy. Following this time sequence, they include:

• Food particles, seen in association with saliva, especially in samples from elderly patients. Meat fibres are elongated and may show cross-striations. Vegetable cells usually have thick straight cellulose walls although of variable appearance. They can be confused with benign or malignant respiratory tract cells,⁴⁴ but have refractile cell walls and a repetitive structure, lacking the pleomorphism of tumour cells and the true nuclear features of malignancy (Fig. 2.13)

• Colonial growth of normal oral flora, including bacteria and yeasts, frequently occurs when there is delay in preparation of samples. The organisms obviously do not elicit any inflammatory reaction. Parasitic or saprophytic flora from the mouth, such as non-pathogenic entamoebae, have been described (see Fig. 2.46).⁴⁵ Mites have been found in respiratory specimens and are occasionally significant clinically, since they may be associated with an actual infestation of diseased airways⁴⁵

• Aerial contaminants are acquired during specimen collection, in transit, or at the time of preparation in the laboratory. Pollen is an obvious external contaminant and must be distinguished from fungal spores. Fungal species may be seen, originating from air, soil or water. The water-borne plant pathogen Alternaria is not uncommon, forming light brown conidia 30 μm in length, with an internal segmented structure (Fig. 2.14). Although potentially pathogenic in immunosuppressed patients, they are present only in small numbers when contaminants⁴⁶

• Carryover of cells from another sample can occur during processing, usually leaving the cells ‘out of context’, often in a slightly different plane of focus or at one edge of the slide.⁴⁷ Cross-contamination by tumour cells is a potential source of false positive diagnosis of malignancy.

Fig. 2.13 A pair of vegetable cells with thick cellulose walls and lack of nuclei (PAP).

Fig. 2.14 Alternaria sp. organism in sputum, probably an aerial contaminant (PAP).

Criteria for assessing adequacy of samples

A sample providing enough cells for confident accurate diagnosis can be regarded as adequate. However, misleading reports are sometimes given if the specimen does not include appropriate material confirming the origin of the sample, or if there is insufficient abnormal material to ensure correct interpretation. Hence, it is one of the prime tasks of the cytologist to assess whether a specimen is suitable for diagnosis or whether the test should be repeated. Furthermore, when tumour cells are found, localisation of their site of origin may not always be possible. This question mainly arises with sputum samples where cells from upper or lower respiratory tract tumours may exfoliate into the sputum.

Sputum specimens are judged adequate when plentiful pulmonary macrophages can be identified.⁴⁸ The presence of columnar cells is ambiguous since they may be from the nasal passages or upper airways. Macrophage counts have been used to quantify the adequacy of sputum specimens,⁴⁹ and to relate these findings to smoking status,⁵⁰ but the procedures are too time-consuming for routine laboratory work. All samples irrespective of their apparent quality should always be screened fully as malignant cells are occasionally found.

Bronchial brushings usually show a profusion of epithelial cells or tumour cells since the material is obtained under direct vision. Poor fixation and air-drying of the cells lead to swelling of nuclei and loss of chromatin detail. Bronchial washings are generally less cellular than brushings, but are usually well-fixed, since preparation is undertaken in the laboratory. Alveolar macrophages provide unequivocal evidence of appropriate sampling.

Bronchoalveolar lavages contain many macrophages and, provided the first aliquot is discarded, should be virtually free from any cells from the upper airways. In a study of over 1500 lavage samples prepared by filtration and cytocentrifugation, however, Chamberlain et al. reported an unsatisfactory rate of 30% as judged by: fewer than 10 alveolar macrophages/high- power field; fewer macrophages than cells from the airways; a mucopurulent exudate; cellular changes due to degeneration; or the presence of laboratory artefacts.⁵¹ These criteria are important when inflammatory cells are to be quantified as in the investigation of interstitial lung diseases.⁵² Adequate sampling is also essential in evaluating specimens for opportunistic infections as discussed in Chapter 16.

FNA samples from solid lung nodules pose no problem for assessment of quality when the cells obtained are interpretable and in keeping with the site aspirated. Scanty preparations with a few normal lung parenchymal cells, macrophages and occasional fragments of bronchial epithelium are all that will be seen if the needle misses the lesion. Mesothelial cells in pavemented sheets are sometimes aspirated on traversing the pleural cavity.

Heavily blood-stained aspirates with rather scanty cells are often poorly fixed due to the presence of blood. Even in these smears, however, diagnostic fields may be found. Blood staining is not usually a problem in liquid-based cytology preparations. Inadequate sampling can be minimised by immediate assessment of the specimen by a cytologist and repeating the procedure if necessary.

NON-NEOPLASTIC PULMONARY CONDITIONS

Non-specific reactive changes in cytological preparations

The majority of respiratory specimens from patients with conditions other than tumours show findings reflecting non-specific host responses. These include the effects of damage by environmental agents, as well as changes seen in many of the more common respiratory infections and chronic chest diseases. Appropriate clinical details and judicious use of special stains may enable the cytologist to confirm the diagnosis in some cases.

• Reactive squamous cells from the upper respiratory tract are seen particularly in sputum. They have slightly enlarged hyperchromatic nuclei and have come to be known as ‘Pap cells’ since according to legend they were noted by Dr George Papanicolaou in his own sputum following laryngitis. They have little diagnostic significance

• Anucleate keratinised squamous cells are only noteworthy in sputum if present in large numbers, when they suggest an area of hyperkeratosis due to a focus of chronic irritation. The surface cells from keratotic lesions exfoliate readily, whether the underlying process is benign or malignant. The specimen must be assessed carefully to determine whether it is of upper or lower respiratory tract origin, and whether there are any cells with nuclear features suspicious of malignancy. Cells from an area of benign hyperkeratosis are usually mature anucleate squames compared with the bizarre-shaped deep orange cells shed from the keratinised surface of a well-differentiated squamous carcinoma

• Hyperplasia of bronchial epithelial cells can be induced by many different noxious agents. In simple repair processes, sheets of actively regenerating cuboidal to columnar cells are seen, with enlarged nuclei, vesicular chromatin and prominent nucleoli (Fig. 2.15).⁵³ Multinucleation is a frequent finding and has been described in cells adjacent to malignant neoplasms.⁵⁴ Papillary clusters of epithelial cells are sometimes shed, mimicking clumps of adenocarcinoma cells. Their nuclei are difficult to examine when the clusters are dense, but they should retain a normal chromatin pattern (Fig. 2.16). Nevertheless, these cell groups are a classical diagnostic pitfall in respiratory tract cytology

• Reserve cell hyperplasia (Fig. 2.17) is less easily recognised in cytology specimens. Groups of small cohesive crowded cells with a high nuclear/cytoplasmic ratio and dense chromatin are typical. Nuclear moulding may be seen, hence the cells can be confused with a small cell carcinoma, as described later in this chapter (see p. 67). Absence of necrosis and cell dissociation, and the uniformity of nuclear size and shape are helpful features

• Hyperplasia of type II pneumocytes and bronchiolar cells (Fig. 2.18) usually occurs as a result of specific toxic effects on the lung and the regeneration that follows. The cells are polygonal or rectangular, occurring singly or in twos and threes or sometimes in larger clusters, and they may show cytoplasmic vacuolation. The nuclei are swollen, with prominent nucleoli and pale or dense chromatin. Such cells are not easy to identify with confidence unless an accurate history is available to alert the cytologist to their origin. They differ from hyperplastic bronchial cells in lacking cilia and in their dissociated cell pattern, with no associated columnar or goblet cells. Distinction from bronchioloalveolar cell carcinoma may be virtually impossible.⁵⁵ There is usually less profuse shedding of cells in reactive hyperplasia than in neoplasia and the cells are in flatter groups, without such a striking depth of focus as is required to inspect the nuclei of malignant cells

• Squamous metaplasia (Fig. 2.19) is one of the commonest responses of bronchial epithelium to persistent injury.⁵⁶ It is preceded by reserve cell hyperplasia and is a frequent finding in smokers.⁵⁷ Small fragmented groups of polygonal cells of parabasal size, with regular large bland looking nuclei are found in the early stages. Dissociation increases as the metaplasia progresses to become atypical or dysplastic, and keratinisation develops in longstanding atypical metaplasia. A more detailed description of atypical metaplasia is given on p. 62 and Figures 2.83 and 2.84.

• Ciliocytophthoria (Fig. 2.20) is a degenerative process affecting bronchial epithelium, whereby columnar cells fragment into rounded cytoplasmic remnants, some of which still show tufts of cilia, while others contain pyknotic nuclear material. The change was first observed by Papanicolaou in 1956 in association with viral infections, but can be seen in a range of acute and chronic pulmonary diseases, and in the presence of bronchial carcinoma.⁵⁸

Fig. 2.15 Hyperplasia of bronchial epithelial cells in repair is associated with disorganisation of cells, nuclear enlargement and pleomorphism and prominent nucleoli, as seen in this group of cells in BAL fluid from a transplant patient (PAP).

Fig. 2.16 Hyperplastic bronchial epithelial cells forming a papillary cluster, with cilia visible on the surface of the group. The nuclei are enlarged but bland-looking. Their depth of focus requires careful examination at high magnification. Bronchial brushing (PAP).

Fig. 2.17 Reserve cell hyperplasia in bronchial brushings from a patient with a squamous carcinoma of bronchus. Note the enlarged active nuclei in this disorganised group of small crowded cells, the high nuclear/cytoplasmic ratio and the narrow rim of cytoplasm (PAP).

Fig. 2.18 Alveolar cell damage and hyperplasia in BAL from a leukaemic patient receiving cyclosporin treatment. The cells at the centre are enlarged and rectangular with dense green cytoplasm, swollen nuclei and prominent nucleoli. Note lack of cilia. A damaged epithelial cell is present at top left (PAP).

Fig. 2.19 Squamous metaplasia in sputum from a middle-aged smoker. The group is composed of cohesive small polygonal cells with regular darkly stained nuclei. Cytoplasmic staining is variable but there is no keratinisation (PAP, LBC).

Fig. 2.20 Ciliocytophthoria in sputum from a patient with squamous carcinoma. Scattered fragments of cytoplasmic remnants present, some ciliated (arrows), others with pyknotic nuclei (arrowheads) (PAP).

Bacterial infective diseases

Infection by microorganisms is the most common cause of inflammation in the lung. In the majority of such cases, cytology has little to offer in the investigative sequence, since the findings are very often non-specific, the diagnosis depending on firm identification of the organism, usually bacterial, by culture. Certain types of infection, however, can be recognised cytologically: fungal elements may be seen, and some viral infections produce diagnostic cytopathic effects. An important additional role for the cytologist lies in helping to exclude an underlying tumour in cases of unexplained or recurrent chest infection.

Bacterial pneumonia

The term pneumonia generally denotes acute inflammation of lung parenchyma due to invasion by microorganisms. This is in contradistinction to pneumonitis where physical agents are involved, or alveolitis, which is due to allergic or fibrosing inflammatory reactions within the alveoli.⁵⁹ Although the incidence of pneumonia and its mortality have fallen substantially with the advent of antibiotics, the disease remains an important cause of morbidity and death, especially at the extremes of life and in debilitated or immunosuppressed patients.

The usual sequence of events once organisms have lodged in the alveoli or distal airways is an immediate acute inflammatory response, with outpouring of oedema fluid, fibrin, neutrophil polymorphs and red blood cells (Fig. 2.21A). Some organisms produce a rapidly spreading infection, involving the entire lobe in a process of consolidation. Other organisms are subject to host defence mechanisms limiting spread to a more patchy distribution. These two processes are known as lobar pneumonia and bronchopneumonia, respectively. The distinction is by no means absolute, but remains valid in many cases. Certain viruses and mycoplasma organisms induce interstitial pneumonia involving inflammation of alveolar walls rather than alveolar spaces.

Fig. 2.21 (A) Lobar pneumonia in histological section of lung. The alveoli are filled with a dense exudate of neutrophil polymorphs producing consolidation of the parenchyma, most marked on the right (H&E). (B) Sputum from a patient with bronchopneumonia, showing many polymorphs and other inflammatory cella (PAP, LBC).

Appropriate treatment and adequate host responses lead to complete resolution of inflammation in many cases, but other outcomes such as abscess formation or fibrosis may supervene in adverse circumstances. The nature of the infectious agent influences the type of inflammatory reaction, creating a more indolent chest infection in some cases. In other cases, the immune system may contribute to progression of the disease, as happens in tuberculosis.

There are marked variations in incidence of different types of pulmonary infection throughout the world, but the increasing number of patients with impaired immunity and the comparative ease of international travel have led to changes in traditional epidemiological patterns. It is, therefore, important to have a full history when assessing cases of respiratory infection.

Cytological findings: bacterial pneumonia

• Grossly purulent or bloodstained sputum

• Predominance of neutrophil polymorphs

• Cell debris and degenerate epithelial cells

• Exfoliation of epithelial cell groups

• Causative organisms are sometimes identified.

Sputum samples are the commonest specimens to be submitted for cytology in cases of pneumonia, often prompted by the need to exclude a bronchial carcinoma rather than to establish the diagnosis of chest infection. Sputum production may be poor, especially in immunosuppressed patients; induced sputum or bronchoalveolar lavage is then more appropriate. FNA is unlikely to be attempted unless an abscess has developed, which may necessitate exclusion of malignancy.

Macroscopically, sputum is often purulent or noted to be rusty due to the presence of blood. Neutrophil polymorphs dominate the microscopic picture (Fig. 2.21B), often at the expense of pulmonary macrophages, obscuring all other cells in some cases. The specimen may be deemed unsatisfactory for cytological assessment if epithelial cells are totally obscured.

Cell debris is prominent in the early stages, whatever sampling method has been used, and degenerative changes such as cytoplasmic vacuolation or ciliocytophthoria may be seen. Clusters of bronchial epithelial cells often show hyperplastic or atypical changes, such as enlarged hyperchromatic nuclei and prominent nucleoli. These features can result from pre-existing lung disease, such as chronic bronchitis or bronchiectasis.

In most cases of pneumonia, the causative organisms cannot be identified cytologically. In cases of immunosuppression, special stains including Gram, Ziehl Neelsen, PAS and Grocott should be performed.

Higher bacteria, such as Actinomyces organisms (Fig. 2.22) have a more defined appearance, forming colonies of radiating filamentous Gram-positive bacteria which may be visible in macroscopic samples as ‘sulphur granules’.⁶⁰ The related organism, Nocardia (Fig. 2.23) stains faintly pink by the PAP method, exhibits negative staining with MGG and is well demonstrated by Grocott’s silver stain. Nocardia pneumonia is a recognised cause of infection after cardiac transplantation, often producing a solitary nodule, which may be subjected to FNA.⁶¹

Fig. 2.22 Actinomycotic organisms presenting in colonies of filamentous organisms in sputum with no significant inflammation. The patient was immunosuppressed and had evidence of chest infection but the colonies may represent overgrowth from the oropharynx in this setting (PAP).

Fig. 2.23 (A) BAL fluid from a transplant patient with respiratory collapse. Filamentous organisms tangled amongst numerous polymorphs are seen (PAP, LBC). (B) Gram staining of the same material highlights strands of Nocardia asteroides. This was confirmed on culture (Gram stain, LBC).

Legionella organisms, the bacteria causing Legionnaires’ disease, have been described in sputum, bronchial samples and FNA material.⁶² They are tiny Gram-negative bacilli which can be demonstrated by silver stains and by immunofluorescence.

Diagnostic pitfalls: bacterial pneumonia

In some circumstances epithelial atypia may be extreme and difficult to distinguish from malignancy. Close liaison with clinical staff is needed in these cases, with judicious use of other investigations, and where necessary, the adoption of a wait-and-see policy.

Stains for organisms must be interpreted with care so that overgrowth of commensals from the oropharynx is not mistaken for an infection. It is important to note the context of the organisms. They are unlikely to be significant if mixed in type and found mainly in areas of saliva without accompanying inflammatory cells.

Pulmonary tuberculosis

The incidence of infection by Mycobacterium tuberculosis fell dramatically in developed countries during the twentieth century, due to improvements in public health and the advent of effective chemotherapy. Nevertheless, tuberculosis remains one of the major causes of morbidity and mortality throughout the world,⁶³ and is again occurring more frequently in Western countries. This is partly attributable to the increasing numbers of disadvantaged groups within affluent societies but also due to the emergence of resistant strains of the organism and because conditions associated with immunosuppression are becoming more common.⁶⁴ The latter group of patients are also susceptible to infection with atypical mycobacteria such as M. avium-intracellulare, M. kansasii and M. fortuitum (see Ch. 16).

The natural history and pathogenesis of pulmonary tuberculosis were expounded by Rich, in 1951.⁶⁵ The causative organism was isolated in 1852 by Koch, and antibiotic treatment has been available from the 1940s. Awareness of the pathology and cytological findings is important to ensure early diagnosis and treatment.

Primary infection usually occurs in childhood by droplet spread. The organisms are localised in the lung parenchyma and the draining hilar lymph nodes, forming a primary tuberculous complex. Macrophages and lymphocytes mount a defence reaction; with persistence of organisms and their breakdown products macrophages take on an epithelioid appearance. After about 1 week, some of the epithelioid cells fuse to form Langhans giant cells, with many nuclei arranged in an arc at one pole of the cell. Lymphocytes accumulate and the whole circumscribed focus of inflammation is known as a granuloma (see Box 2.1). Within about 2 weeks, the centre of the granuloma starts to undergo caseation necrosis of a characteristic soft cheesy consistency (Fig. 2.24). Epithelioid histiocytes and Langhans giant cells tend to form palisades around the edge of the caseous material and it is in this area that acid-fast tubercle bacilli are most often found.

Box 2.1 Some conditions associated with pulmonary granulomata

Occupational exposure

• Organic dusts (extrinsic allergic alveolitis)

Iatrogenic causes

• Drug toxicity

• Radiation exposure

• Oxygen therapy

Other conditions

• Foreign body reaction

• Tumour-related granulomas

Fig. 2.24 Section of bronchial wall with discrete epithelioid giant cell granulomata in the submucosa, a less common site for these lesions than in the lung parenchyma. Note the intact mucosa over the surface, impeding direct sampling by cytology. (H&E)

(Courtesy of Dr MS Dunnill, Oxford, UK.)

As immunity develops, resolution occurs, leaving a peripheral lung scar and calcified draining lymph nodes. If the number of infecting organisms is large, however, or the patient is debilitated, infection may spread elsewhere in the lung or via the bloodstream to other organs. Tuberculous bronchopneumonia and miliary tuberculosis develop in this way. Healing may take place but infection may occur at any time, usually by reactivation of a dormant focus of organisms in the lung. This secondary or adult infection takes the form of progressive granulomatous bronchopneumonia with caseation, cavitation and extensive lung destruction.

Awareness of this stage of the disease is important in cytology. The patient has a cough productive of sputum. Bronchoscopy may be undertaken for the collection of washings and brushings to exclude malignancy and to obtain material for culture. Effusions are common. Localised pulmonary lesions occur, simulating malignancy radiologically and inviting FNA sampling. Thus, a variety of methods of cytological investigation may be employed. The findings are well-documented in sputum and bronchial secretions.⁶⁶ The characteristic features of granuloma formation are best seen in FNA material.⁶⁷

Cytological findings: pulmonary tuberculosis

• A mixed inflammatory exudate

• Epithelioid histiocytes

• Langhans giant cells are infrequently found

• Fragmented granulomas may be seen in FNA material

• Caseation is inconspicuous

• Organisms may be demonstrated, mainly in FNA samples.

An analysis undertaken in Brazil by Tani et al.⁶⁸ in 1987 of over 100 tuberculous cytology samples other than FNAs revealed increased numbers of macrophages in 100%, excess neutrophils in 98% and increased lymphocytes in 85% of the specimens. Epithelioid cells were present in 56% and giant cells in only 40% of the samples.

Epithelioid histiocytes are elongated macrophages with pale cytoplasm devoid of any tingible ingested material such as carbon pigment. Their nuclei are drawn out and indented or folded, producing a variety of footprint-like shapes. The chromatin is finely divided and nucleoli are usually inconspicuous. The cells are arranged in loose aggregates in sputum or washings, but may be aspirated in ragged clumps by FNA (Fig. 2.25). Macrophages of more usual type can also be seen.

Fig. 2.25 Ragged fragment from an epithelioid granuloma in FNA sample from a patient with pulmonary tuberculosis. Pale histiocytic cells with elongated nuclei can be seen forming the granuloma; lymphocytes and other inflammatory cells are present in the debris at the periphery (MGG).

Langhans giant cells are characteristic of the disease but are not often seen in cytology samples, apart from FNA material. They are 2–10 times the size of mononuclear macrophages and may contain up to 100 or so ovoid nuclei, typically distributed at one pole of the cell (Fig. 2.26). This feature and absence of ingested carbon help to distinguish them from other multinucleated pulmonary macrophages. The cytoplasm is amphophilic with Papanicolaou staining or pale blue with MGG.

Fig. 2.26 Langhans giant cell with abundant amphophilic cytoplasm containing many rounded or oval nuclei grouped at one pole. Note absence of any ingested material. Bronchial washing (PAP).

Caseation necrosis is suggested by the presence of pale amorphous material, not easily seen in PAP-stained sputum or bronchial secretions unless liquid-based methods of preparation are used, but recognisable in FNA samples as faintly granular light blue stained debris in MGG or H&E preparations, sometimes speckled darker blue if calcification has occurred. It is within the caseous material that tubercle bacilli are most likely to be found.

Lavage fluid from patients with AIDS who have tuberculosis typically contains many lymphocytes and enlarged foamy macrophages, but it is unusual to find a definite granulomatous picture in these samples. A thin purulent background may be seen.

Mycobacterium tuberculosis may be demonstrated, especially in FNA material using Ziehl Neelsen stain, which reveals a beaded magenta pink straight or slightly curved slender bacillus 1–4 μm in length. Fluorescent methods, such as Rhodamine-auramine staining are quicker to screen if there are large amounts of material. The organisms are sometimes visible as negative staining images within caseous material in FNA preparations stained by MGG (Fig. 2.27) or in sputum when there is a high load of acid fast bacilli.

Fig. 2.27 Caseous material obtained on FNA of a lung nodule subsequently confirmed as tuberculous. Within the amorphous debris negative images of the tubercle bacillus can be made out (arrows) (H&E).

(Courtesy of Dr G. Sterrett, Perth, Western Australia.)

Atypical mycobacteria differ slightly in morphology, but cannot be firmly distinguished without culture. When no spare material is available, slides can be decolourised and re-stained successfully. Nevertheless, culture is essential in all cases, submitting as much material as possible.

Diagnostic pitfalls: pulmonary tuberculosis

The combination of epithelioid histiocytes and Langhans’ giant cells is highly suggestive of tuberculosis but is not pathognomonic since either or both cell types can be seen in other conditions with granulomatous inflammation (Box 2.1).

Caseation necrosis may closely resemble tumour necrosis. A careful search for evidence of granuloma formation or for remnants of malignant nuclei should resolve this problem.

Atypical strains of mycobacteria causing lung infections are morphologically similar to M. tuberculosis and can only be distinguished by culture.

Viral infections

In contrast to pneumonia due to bacteria, viral infections frequently induce specific cytopathic changes in epithelial cells and alveolar macrophages, enabling the pathologist to give an indication as to the causative agent. This is particularly important since other methods of diagnosis may take longer to complete, may not be available, or may not be as accurate. It is important, however, to obtain confirmation by culture whenever possible if a viral origin is suspected on cytology.

The cytopathic effects are noted mainly in sputum and bronchial secretions and can be seen on PAP staining. Immunostaining will provide firm identification of the virus. Non-specific inflammatory, reactive and degenerative changes are also often present, providing a background to the diagnosis.

Non-specific cytological findings: viral infections

• Necrosis and inflammation

• Ciliocytophthoria

• Bronchial and alveolar cell hyperplasia.

Inflammatory cells and necrotic debris are a frequent finding in the early stages, especially in those infections caused by influenza and parainfluenza viruses.⁶⁹ Ciliocytophthoria with fragmented ciliated and pyknotic nucleated remnants is a variable feature. The phenomenon was first described in adenovirus infections, but is not specific, occurring in other infections and also in association with neoplasia and radiotherapy.⁵⁸

Bronchial and alveolar cell hyperplasia may be seen, producing clusters of enlarged epithelial cells with swollen hyperchromatic nuclei and prominent nucleoli, as found in a variety of inflammatory states. These changes are easily confused with malignancy but the cohesiveness of the groups in only small numbers and absence of single dissociated abnormal cells favour a reactive process.⁷⁰