J Clin Pathol 1993;46:589-595

ACP Broadsheet No 140

Techniques in pulmonary cytopathology

J A Young

IntroductionThe foundation of pulmonary cytopathologyis generally attributed to Dudgeon andBarrett (1934),1 although physicians haveinspected sputum for evidence of lung diseasesince the earliest days of medicine. There arefew fields of cytopathology in which modemdevelopments in fibreoptics and imagingtechniques have contributed so much, notonly by expanding the variety of specimenswhich can be obtained, but also by increasingthe quantity and quality of the cellular mater-ial recovered. The chief diagnostic applica-tion remains the investigation of patients withsuspected carcinoma of the lung, but throughthe technique of bronchoalveolar lavage, pul-monary cytopathology also contributes to themanagement of benign disease particularly inthe immunosuppressed.

Sampling methods and cytopreparatorytechniquesThe different methods of obtaining cellularmaterial for cytological examination areshown in table 1. In addition, supplementarytechniques such as fine needle aspiration oflymph nodes and examination of pleural fluidcan be used in staging.

This Broadsheet has beenprepared by the author at theinvitation of the Association ofClinical Pathologists whoreserve the copyright. Furthercopies of this Broadsheet maybe obtained from thePublishing Manager, Jtournalof Clinical Pathology, BMAHouse, Tavistock Square,London WCIH 9JR

Department ofPathology, TheMedical SchoolUniversity ofBirmingham,Birmingham B15 2TTJ A YoungAccepted for publication13 October 1992

SPUTUMExamination of sputum is non-invasive anddoes not inconvenience patients or clinicians.It can provide definitive evidence of malig-nancy in many cases but no localisation of thelesion. It is, however, labour intensive as alaboratory investigation both with regard topreparation and microscopy and a percentage

Table 1 Specimen collection methods for pulmonarycytopathology

SputumSpontaneousInduced

Bronchoscopic specimensAspiration of bronchial secretionsBronchial washingBronchial brushingTransbronchial fine needle aspiration (TBFNA)

Percutaneous fine needle aspiration (PCFNA)Bronchoalveolar lavage (BAL)

of the specimens processed will turn out to beunsatisfactory.

Specimens of early morning "deep cough"sputum should be submitted on three consec-utive days. Sputum produced during physio-therapy to the chest is also suitable. Sputumobtained after bronchoscopy may yield valu-able diagnostic information as it is usuallyhighly cellular but must be labelled as such;otherwise the hyperexfoliation of epithelialcells may lead to interpretative difficulties.In patients that cannot expectorate spon-taneously induced sputum can be produced byinhalation of the vapours of a warmed (37'C)mixture of 15% sodium chloride and 20%propylene glycol for 20 minutes. Sputumsamples should be collected into clean, drycontainers with screw-on lids and transportedto the laboratory in biohazard bags. Prior fix-ation with 70% ethyl alcohol is not recom-mended as it produces a rubbery specimenwhich is almost impossible to smear.Preservation of cells in unfixed sputumremains reasonable for 24 hours if the con-tainer is refrigerated.

Saliva and upper respiratory tract mucusare seldom of diagnostic value, but the dis-tinction between these and lower respiratorytract material cannot be reliably made with-out microscopy and all specimens must there-fore be processed. Using a safety cabinet, thesputum is emptied into a Petri dish andinspected. Any thick streaks, bloodstainedareas, or tissue fragments should be selec-tively included in the sampling which is car-ried out with disposable plastic forceps. Asmall blob is placed on a slide, squasheddown with a second slide, and the two drawnapart and immediately placed into 95% ethylalcohol for fixation. A total of four slides isgenerally sufficient. Staining is best achievedby the traditional Papanicolaou methodwhich gives excellent clear morphologicaldetail and permits inspection of the specialtinctorial properties of squamous cells, aninherent part of this technique. The othertraditional cytopathological stain, May-Grunwald-Giemsa (MGG), is not suitable forsputum due to the mucoid background whichobscures the cells. Much less cellular detail,especially in thickly spread areas, can be seenwith haemotoxylin and eosin. If special stains,

July 1993

589

Young

V'

b.

. *t.

't!wt .Ai:= ^* t

sUi_li_

..'_..

....*aF ik.s _ _o3 _:

## w >.: >

.s .:e: 8;

F__t.P,,,, ;;-R-_.i...fx

Figure 1 Adenocarcinoma: sputum. A large threedimensional group of malignant cells (Papanicolaou).

such as Ziehl Neelsen are required, additionalslides can be prepared from the originalspecimen which should be retained untilmicroscopy is complete.

Several more complicated preparatorytechniques exist, but none conveys any realadvantage over the standard simple methodas they are time consuming, result in loss ofmorphological detail, or disrupt the pattern ofcell streaks and groups which are helpful informing a diagnosis. The most widely used ofthese is the Saccomano technique2 in whichpooled sputum is collected into a mixture of50% ethyl alcohol and 2% polyethylene gly-col (Carbowax) and the specimen emulsifiedin a blender in the laboratory. Apart from theinherent risk from aerosol spread of infection,the diagnostic sensitivity is inferior to theexamination of fresh sputum as describedabove.3 Formalin fixation of centrifuged cellblocks which are then processed histo-logically4 is a laborious procedure and frompersonal experience produces a specimenwhich is difficult to interpret.

BRONCHOSCOPIC SPECIMENS FOR DIAGNOSISOF NEOPLASIABronchial secretions and washingsBronchial secretions can be aspirated through aflexible catheter inserted down the broncho-scope. They generally yield only degeneratedebris, mucus, and inflammatory cells andare of little diagnostic value. Bronchial wash-ings are obtained by instilling 3-5 ml of a bal-anced salt solution down the bronchoscopeand then aspirating the fluid. The lavage canbe aimed at an abnormal area or directedalong the bronchus to sample from outsidethe area of visual inspection. The fluid is cen-trifuged and about six slides prepared fromthe cell button. These are immediately fixedin 95% ethyl alcohol. Four are stained by thePapanicolaou method and the remainderretained in case additional material isrequired.

Figure 2 Small cell anaplastic carcinoma: sputum. Oatcells in characteristic streak (Papanicolaou).

Bronchial brushingsBrushings are taken directly from the sus-pected tumour and can contain many wellpreserved cells. If the surface of the lesionappears necrotic two brushings should beobtained as the first may yield only debris.Brushing is best carried out before bronchialbiopsy as the latter may cause bleeding whichcan interfere with the fixation of the cells har-vested by the brush. Immediately the brush isremoved from the bronchoscope it should befirmly rolled on to a series of glass slides, eachslide being placed in a transport container of95% ethyl alcohol as it is prepared. Greatcare must be taken to ensure that the slidesare not allowed to dry before alcohol fixation.Sufficient material is usually available for fourslides. Additional cells can be obtained byagitating the brush in 5 ml of isotonic salineand centrifuging the fluid in the laboratory.

Transbronchialfine needle aspirationTransbronchial fine needle aspiration is usedless frequently than other cytological sam-pling techniques and is time consuming toperform. It is useful for investigating externalbronchial compression or submucosal lesions,however, especially in the upper lobes wherebiopsy forceps can be difficult to manipulate.Hilar, subcarinal, and low right paratracheallymph nodes are particularly suitable forthis type of investigation. Transbronchial fineneedle aspiration may sometimes providediagnostic evidence of malignancy when bothbrushing and biopsy fail.The Wang disposable needle (Millrose

Company, Ohio) is the device most oftenused for transbronchial fine needle aspiration.5 6The apparatus consists of a 120 cm long

590

'f

4."'k-,-*j

Techniques in pulmonary cytopathology

Figure 3 Squamous cellcarcinoma: bronchialbrushings. Welldifferentiated keratinisedmalignant cells(Papanicolaou).

double-lumen retractable needle system. Acentral stylet is surrounded by two polythenesheaths, the inner of which is fitted with a 22gauge, 13 mm long needle. The needle is pro-tected by the outer sheath while it is passeddown the bronchus, the stylet providing therigidity necessary for guidance. When theaspiration site is reached the stylet is retractedand the proximal end of the catheter attachedto a 20 ml syringe. The plunger is withdrawnto produce a vacuum and with this main-tained a series of rapid passes into the lesionare made with the needle. Suction is thenreleased and the catheter and needle with thecellular material in situ are withdrawn fromthe bronchoscope. The needle contents areexpressed on to slides and smeared out in the

K ...

Figure 4 Large cell carcinoma: bronchial brushing. Malignant cells showing nodiscernible differentiation (Papanicolaou).

standard manner for fine needle aspirates.7One to two slides should be rapidly placed in95% ethyl alcohol for fixation and subsequentPapanicolaou staining. The remaining couplecan be air dried and then stained with MGG.The latter technique is very suitable for fineneedle aspirates as contamination withbronchial mucus and debris is much less thanwith brushings and washings.

PERCUTANEOUS FINE NEEDLE ASPIRATIONOf all anatomic regions, the lung is probablythe site in which the advent of percutaneousfine needle aspiration (PCFNA) has had thegreatest impact on patient management.Apart from the occasional instance of infec-tious disease it is almost exclusively used forthe diagnosis of malignant disease. Sensitivityand specificity are high and PCFNA avoidsopen lung biopsy as a means of diagnosis.The method is not without its complications,however, chief among which is the risk ofpneumothorax. Tumours seeding in theneedle track is extremely rare.

Accurate localisation of the lesion is neces-sary. Skin markers and plain x-ray picturesare insufficient. Fluoroscopic control, prefer-ably with biplane imaging, computed tomog-raphy, or for very superficial lesions,ultrasound scans, are required. Aspiration iseasiest if the patient is lying down and a verti-cal track is chosen for the needle. The centresof very large tumours are frequently necroticand better preserved diagnostic material iscommonly aspirated from the periphery.

Local anaesthetic is injected down to thepleura and the needle is then inserted andadvanced under fluoroscopic or other type ofimaging control. A 20 to 25 gauge needle isused. Special needles such as the Rotex screwneedle8 are available for sampling hardlesions, such as chondromatous hamartoma.A fresh sterile needle is required for each per-cutaneous pass if aspiration is repeated. Withthe needle in position a 20 ml or 50 mlsyringe is attached and suction is applied.Aspiration of pus or blood mixed with solidparticles is indicative of satisfactory aspira-tion. The needle and syringe are then withdrawn.PCFNA is carried out by the radiologist in

the Department of Diagnostic Imaging but itis advisable for a cytotechnologist (or pathol-ogist if it is considered that viewing the imageduring aspiration will assist in diagnosis) tobe present to receive the needle and syringeto ensure optimal handling of the aspiratedmaterial. The equipment listed in table 2should be laid out ready for use.

Unless a large volume of aspirate isobtained the needle is detached from thesyringe and a small quantity of air drawn in.The needle is then reattached and a drop ofcontents "blown out" on to a slide and thinlyspread with the end of a second slide as ifmaking a blood smear. The procedure isrepeated until the aspirate is fully processed.Half the slides should be placed in the ethylalcohol as soon as they are prepared and theremainder rapidly air dried (preferably with ahairdryer) and placed in the rack. The needle

591

III 0

ip0. It

Young

Table 2 Equipment for slide preparation afterpercutaneous fine needle aspiration

* Supply of frosted end glass slides* Pencil to label slides* Tray for spreading out slides* Containers of 95% ethyl alcohol for wet fixation of slides* Hair dryer to dry slides* Rack for air dried slides* Watch glass* Paper tissues* Disposable plastic forceps* Disposable centrifuge tube containing 5 ml physiological

saline, Hanks's solution, or cell culture medium to rinseout needle and syringe

* Pot of 10% buffered formalin for tissue fragments* Tube of medium if microbiological culture is required* Reagents for Diff-Quik stain (Merz Dade) as in table 3* Request forms

and syringe are then washed out in saline orother transport medium to recover anyremaining material by subsequent centrifuga-tion. If the aspirate is very copious and con-

tains much blood the whole contents of thesyringe can be expelled into a watch glass andthe excess fluid absorbed with paper tissueand smears then prepared from the residualmaterial. Any obvious tissue fragments are

placed in the buffered formalin. If microscopefacilities are available a slide can be stained bythe Diff-Quik method (table 3) and examinedin the Department of Diagnostic Imaging toensure that adequate material has been aspi-rated before the patient returns to the ward.

In the laboratory the prepared slides are

stained with Papanicolaou (alcohol fixed) or

MGG (air dried), retaining several for specialstains if these should be required. Additionaluseful material can often be obtained fromthe centrifuged transport medium used torinse the needle and syringe. If there are tis-sue fragments these are embedded and sec-tioned as routine for small histological biopsyspecimens.

Diagnosis ofneoplasiasINVESTIGATIVE PROTOCOLSome controversy exists as to the relativeroles that sputum, endoscopic sampling(cytopathology and biopsy), and PCFNA

Figure 6 Small cell anaplastic carci'noma: percutaneous

fine needle aspirate. Two groups of oat cells (MGG).

should have in the investigation of patients

with suspected malignant disease. Protocols

vary from insistence on a minimum of three

sputum examinations plus at least one bron-

choscopy before resorting to PGFNA, to a

recommendation that for reasons of speed,

accuracy, and cost effectiveness PGFNA

should be the primary approach in the diag-

nostic work-up. An intermediate position

based on the clinical findings in each individ-

ual patient including the size and location of

the lung mass and assessment of the relative

risk of the various procedures seem most

practical and the best use of resources. The

various investigations are complementary,

and while PGFNA is the single most sensitive

technique less invasive methods are fre-

quently fully effective.

ASSESSMENT OF SPECIMEN ADEQUACYWith bronchial brushings and fine needle aspi-rates (TBFNA and PCFNA), this is straight-forward and depends on the presence ofsatisfactory numbers of viable cells. For aspecimen of sputum to be considered ade-quate, alveolar macrophages must be present.The identification of ciliated columnar cells is

Table 3 Diff-Quik stainingfor rapid evaluation offineneedle aspirates1 Make smear2 Rapidly air dry with hair dryer3 Pot 1* 15 dips4 Pot 2* 15 dips5 Pot 3* 25 dips6 Rinse with distilled water7 Air-dry as in No 2 above8 Dip in xylene9 Coverslip using DPX resin mountant

Figure 5 Hamartoma: percutaneous fine needle aspirate. Mesenchymal material andcartilage cells (Papanicolaou).

*Commercial fixation and staining reagents supplied byMerz-Dade.

592

.,!.`Ol

Techniques in pulmonary cytopathology

taI

Sll

&L

4

4.

Ih

. ntS* d

Iti.4 %~

Figure 7 Large cell carcinoma: percutaneous fine needleaspirate. Large poorly differentiated malignant cells(MGG).

not sufficient evidence of lower respiratorytract sampling as these may be derived fromthe upper airway.

REPORTINGAll available material should be examined asdiagnostic cells (figs 1-7) may be irregularlydistributed throughout the prepared slides.Particular attention is needed at the edgesand tails of smears, especially of sputum, asthese are the areas where malignant cells aremost likely to be identified. Guidance on theinterpretation of cytomorphology is availableelsewhere.9 10 In addition to a description andconclusion, all reports must include a recordof the number of slides examined, in case offuture review.

Bronchoalveolar lavageBronchoalveolar lavage (BAL) is mainly usedto investigate pulmonary infiltrates inimmunosuppressed patients (figs 8 and 9)and to determine disease activity andresponse to treatment in cases of interstitiallung disease. It may be helpful in suspectedasbestosis (fig 10) or in a few uncommon dis-orders such as lipoid pneumonia and alveolarproteinosis, and can also be used to retrievecells from diffuse or localised, suspected neo-plasia not amenable to diagnosis by brushingor biopsy.

LAVAGE TECHNIQUEBAL is usually performed via the transnasalroute and a 4 9 mm fibreoptic bronchoscope.Fluid is instilled by gravity or under gentlepressure with a syringe and then reaspiratedby syringe or mechanical suction. A total of100-300 mls isotonic saline in 50 ml aliquots

Figure 8 P carinii infection: bronchoalveolar lavage.Two characteristic alveolar casts (Papanicolaou).

is usually obtained, although in severely illimmunosuppressed patients smaller volumescan be used. The lavage is performed at thesite of maximum radiological opacity or themiddle lobe in cases of diffuse disease. About40-60% of the volume of fluid instilled isrecovered. The first aliquot generally onlysamples bronchial contents and alveolarmaterial may not be obtained until subse-quent specimens. A sample of fluid isreserved for the Department of Microbiologyand the remainder for cytopathology andallied studies.

LABORATORY TECHNIQUESCell preservation is poor in saline and fluidfor cytopathology should be processed within30 minutes of BAL. Cytopreparatory meth-ods vary according to the clinical conditionsunder investigation.

Pulmonary infiltrates in immunosuppressedpatientsAliquot volumes are often small and all mate-rial should be processed.

Centrifuge fluid at 3500 rpm for 5 min-utes.

Prepare slides from deposit and stain withPapanicolaou, MGG, Gram, ZN, andGrocott. With the exception of thePapanicolaou stain which requires alcohol fix-ation, all the other stains are performed onair-dried slides. Additional techniques-forexample, fluorescent or immunocytochemicaldetection of Pneumocystis carinii, cyto-megalovirus, or herpes simplex virus-arealso possible. The morphology of the cellspresent and any background debris shouldbe considered in addition to examining the

593

4F .:

-ML %.% .,

q".. "i

'.,6 &.

k:. iI':i.h.,'T'

Young

Table 4 Diffuse pulmonary infiltrates inimmunosuppressed patients: diagnoses possible from BALfluid

Opportunistic infectionBacteria

Gram positive bacteriaMycobacterium tuberculosis (M avium,M intracellulart)

FungiCandida spAspergillus spZygomycetesCryptococcus neoformansHistoplasma capsulatumPneumocystis carinii

VirusesCytomegalovirusHerpes simplex virus (varicella zoster)Non-specific viral inclusions (measles,adenovirus)

ParasitesStrongyloides stercoralis

Malignant infiltrationLeukaemiaLymphomaLymphangitic carcinomatosisKaposi's sarcoma

HaemorrhagelipoprotemosisPneumonitis, non-specific

(Sepsis, drugs, irradiation, oxygentreatment, graft versus host disease)

specimen for evidence of specific infection.The multifactorial nature of disease in theimmunosuppressed patient cannot be overemphasised, and several of the conditionsshown in table 4 may coexist. The morpho-logical changes produced by pneumonitis,especially those caused by drugs or radiation,are severe and care must be taken not to con-fuse these with malignant infiltration.

Diagnosis of specific benign lung diseaseThe application of cytopathology is very lim-

0

ited but some special stains may be helpfulprovided interpretation is closely integratedwith clinical findings. For example, Oil red 0(lipoid pneumonia), Perns's (haemosiderin-laden macrophages, ferruginous bodies),diastase-periodic acid Schiff (alveolar protein-osis). If electron microscopy is required (alve-olar proteinosis, histiocytosis X) 25 ml of 5%glutaraldehyde in cacodylate buffer is addedto 25 ml BAL fluid as an electron microscopyfixative. Langerhans' cells in BAL react withanti-CD1, but small numbers may be foundin conditions other than histiocytosis X.

Interstitial lung diseaseThe role ofBAL is to determine disease activ-ity and response to treatment. There is con-siderable variation in the results obtainedwhich weakens the case for definitive mea-surement by BAL despite the initial enthusi-asm of some centres." 12 Total cell countshave little value." A suitable method forcarrying out differential cell counts is thatdescribed by Turner-Warwick and Haslam."I

Separate the cells from the fluid by slowcentrifugation and resuspend the cells inbuffered tissue culture medium to2 x 106 cells/ml.Use 100 aliquots of this suspension to

prepare a set of slides by cytocentrifugation(Shandon Cytospin 450 rpm for 4 minutes).Stain with MGG and differentially count 300to 500 cells by random point counting withx 40 objective.

Cells present include alveolar macro-phages, lymphocytes, neutrophils, eosinophilsand mast cells. Subsets of lymphocytes canbe identified by appropriate immunocyto-

4i

e,.

_

Figure 9 P carinii infection: bronchoalveolar lavage.Cystic forms within alveolar cast (Grocott's stain).

A_ALFigure 10 Ferruginous body: bronchoalveolar lavage. Acurledferruginous body against a background of alveolarmacrophages (Perls's stain).

594

r.

"..z .11., ".., ':-

.r-- "I S. .... ".

Techniques in pulmonary cytopathology

chemical studies. Epithelial cells should bevery few. Excessive numbers of ciliatedcolumnar cells or abundant mucopurulentdebris are indicative of predominantlybronchial rather than alveolar material andcell counts will be unreliable. Differential cellcounts should not be carried out in isolationand the cytomorphology of the cell popula-tion should be examined on Papanicolaou aswell as MGG stained slides and by additionalspecial stains if necessary.

It is difficult to give guidance on results ofdifferential cell counts due to the variability ofpublished figures, each centre establishing itsown range. Details of these can be found intabulated form in Stanley et al.'2 Overall,idiopathic pulmonary fibrosis is associatedwith alveolar neutrophilia, prognosis isadversely affected by the presence ofeosinophilia, and corticosteroid treatmentis most likely to benefit patients with lavagelymphocytosis. Increased helper T lympho-cytes are associated with sarcoidosis andincreased T suppressor cells with hyper-sensitivity pneumonitis.

1 Dudgeon LS, Barrett HR. The examination of fresh tis-sues by the wet-film method. BrJ Surg 1934;22:4-22.

2 Risse EKJ, van't Hof MA, Laurini RN, Vooijs PG.Sputum cytology by the Saccomanno method in diag-nosing lung malignancy. Diagnost Cytopathol 1985;1:286-91.

3 Perlman EJ, Erozan YS, Howdan AH. The role of theSaccomanno technique in sputum cytopathologic diag-nosis of lung cancer. AmJ7 Clin Pathol 1989;91:57-60.

4 Bocking A, Biesterfeld S, Chatelain R, Gien-Gerlach G,Esser E. Diagnosis of bronchial carcinoma on sectionsof paraffin-embedded sputum. Acta Cytol 1992;36:37-45.

5 Wang K-P, Haponik EF, Britt EJ, Khouri N, Erozan Y.Transbronchial needle aspiration of peripheralpulmonary nodules. Chest 1984;86:819-23.

6 Wagner ED, Ramzy I, Greenberg SD, Gonzalez JM.Transbronchial fine-needle aspiration. Am J Clin Pathol1989;92:36-50.

7 Young JA. Technique. In: Young JA, ed. Fine needleaspiration cytopathology. Oxford. Blackwell ScientificPublications Ltd, 1993:6-12.

8 Nordenstrom BEW. Technical aspects of obtaining cellu-lar material from lesions deep in the lung. A radiolo-gist's view and description of screw-needle samplingtechnique. Acta Cytol 1984;28:233-42.

9 Young JA. Colour atlas ofpulmonary cytology. Oxford colouratlases of pathology. London: Harvey Miller Publishersand Oxford University Press, 1985.

10 Young JA. The lung, pleura, and chest wall. In: YoungJA, ed. Fine neede aspiration cytopathology. Oxford:Blackwell Scientific Publications Ltd, 1993:97-12 1.

11 Tumer-Warwick DM, Haslam PL. Clinical applicationsof bronchoalveolar lavage. Clin Chest Med 1987;8:15-26.

12 Stanley MW, Henry-Stanley MJ, Iber C. Bronchoalveolarlavage. New York: Igaku-Shoin, 1991.

595