ot_bib_bo_diagnosticcorrelates

DOI 10.1378/chest.99.4.809 1991;99;809-814Chest

 G Buccheri, P Barberis and M S Delfino of 1,045 bronchoscopic examinations.correlates in bronchogenic carcinoma. A review Diagnostic, morphologic, and histopathologic

  http://chestjournal.chestpubs.org/content/99/4/809

can be found online on the World Wide Web at: The online version of this article, along with updated information and services 

) ISSN:0012-3692http://chestjournal.chestpubs.org/site/misc/reprints.xhtml(without the prior written permission of the copyright holder.reserved. No part of this article or PDF may be reproduced or distributedChest Physicians, 3300 Dundee Road, Northbrook, IL 60062. All rights

ofbeen published monthly since 1935. Copyright1991by the American College is the official journal of the American College of Chest Physicians. It hasChest

 © 1991 American College of Chest Physicians by guest on October 28, 2010chestjournal.chestpubs.orgDownloaded from

CHEST/99/4/APRIL, 1991 809

Diagnostic, Morphologic, andHistopathologic Correlatesin Bronchogenic Carcinoma*A Review of I ,045 Bronchoscopic ExaminationsGianfranco Buccheri, M. D. , F. C. C.P; Tholo Barberis, M.D.;

and Maria S. Delfino, M.D.

Information on the correlation between bronchoscopicallyvisible aspects, histopathologic classification, and diagnosticyield is very scarce. To contribute to the knowledge of thesubject, we reviewed the bronchoscopic charts of 1,045patients with lung cancer who were seen in the years from1983 to 1989 at the Bronchology Service of the A. Carle

Hospital. Tumors were more often located centrally and

superiorly. No preference as to side was found. Squamous

carcinomas were, by far, the most frequent cell type.Forceps biopsies, brushings, and washings were positive in

79 percent, 38 percent, and 32 percent of the obtained

specimens, respectively. Bronchoscopically, squamous andsmall-cell carcinomas were more often visualized as central

tumor-like lesions, which were better diagnosed by forceps

biopsies. Adenocarcinomas, on the contrary, were morefrequently peripheral and showed infiltrative, compressive,

or aspecific findings. In these latter tumors, cytologic studieswere more fruitful. Large-cell anaplastic carcinomas hadan intermediate behavior. Cell type, endoscopic appear-ance, and diagnostic success are interrelated features.Visible characteristics at bronchoscopy can therefore antic-ipate the more likely histotype and guide the diagnosticapproach. (Chest 1991; 99:809-14)

AS = aspecific findings; SC squamous carcinoma; 5CCsmall-cell carcinoma; ACadenocarcinoma; LCC large-cellcarcinoma

F or the diagnosis of bronchogenic carcinoma, no

method has proven more valuable than endoscopic

examination of the tracheobronchial tree. The intro-

duction ofthe fiberoptic bronchoscope (FB) in the late

1960s further enhanced the diagnostic potential of the

examination. The flexibility of the new instrument

ameliorated both the acceptability to patients and the

security of the procedure.2 Bronchoscopy, which in-

cludes both visual inspection and collection of speci-

mens for cytohistologic studies, is now an essential

part of the routine work-up of any patient suspected

of having lung cancer.

Fifteen years ago, Sackner� reviewed the clinical

application of flexible bronchoscopy. Since then, mod-

ifications in the use ofthe FB, such as lung transbron-

chial biopsies,� lymph node transbronchial needle

aspirations,5’6 or alveolar lavages for peripheral le-

sions,7 have sometimes been adopted. Nevertheless,

even today, the majority ofdiagnoses are substantiated

by biopsy, brushing, and washing specimens obtained

during a routine fiberoptic bronchoscopy.

Several studies have investigated the overall accu-

racy of the diagnostic techniques used in association

*Fmm the A. Carle Hospital ofChest Diseases, Cuneo, Italy.Presented in part at the 8th Congress of the European Society ofPneumology, Freiburg, West Germany, Sept 10-14, 1989, and atthe XVI World Congress on Diseases of the Chest, Boston, Oct

30-Nov3, 1989.

Manuscript received April 25; revision accepted July 17.Reprint requests: Dt� Buccheri, Via Repubblica 10/C, Roccavione(CN), Italy 1-12018

with the FB.�” Other studies have compared them to

other methods ofcytohistologic sampling (for example,

to percutaneous needle 12 More recently,

the capability of the same techniques in making a

correct diagnosis of cell type has been the object of

diverse 1214

Unfortunately, most of the previously mentioned

information is oflittle value to the bronchoscopist who

is examining a new patient. As a matter of fact, the

bronchoscopist has no knowledge of what the patho-

logic diagnosis will be; and he is faced, at best, with

recognizable endoscopic signs of malignancy; so he

needs to know, above all, what the diagnosis could be,

what is the histotype, and what is the diagnostic

potential for that given bronchoscopic finding. Accord-

ingly, we thought it important to examine the relation-

ship existing between visible aspects and the other

diagnostic elements ofthe bronchoscopic examination.

In this study, we reviewed seven years of experience

and 1 ,045 bronchoscopically examined cases of lung

cancer, laying particular emphasis on endoscopic mor-

phology� pathologic classification, and diagnostic ac-

curacy.

MATERIALS AND METHODS

The patients of this series were selected from 3,292 whounderwent fiberoptic bronchoscopy in the years from 1983 to 1989

at the Bronchology Service ofthe A. Carle Hospital. Examinations

were carried out at the request of physicians of the medical units ofboth the A. Carle Hospital and the surrounding hospitals. Nearly

 © 1991 American College of Chest Physicians by guest on October 28, 2010chestjournal.chestpubs.orgDownloaded from

vs

810 A Review of 1 045 Bronchoscopic Examinations in Carcinoma (Buccheh� Barberis, Deffino)

all patients were referred because of abnormal chest x-ray findings

or hemoptysis with normal roentgenograms. Eligible for this study

were patients with a carcinoma ofthe lung pathologically confirmed,

whatever the tissue biopsied and the way used to access; however,

because of the existence of diverse practical difficulties in obtaining

a copy of the pathologic reports not already kept at the service, we

limited the analysis of histotypes to the sole available diagnoses, ie,

those made bronchoscopically or with posthronchoscopic specimens

of sputum.

Endoscopic procedures were performed by three bronchoscopists

(including one of the authors), all having undertaken more than 300

examinations at the beginning of the study. A classification of the

visible bronchial abnormalities has been used by our team since

the early 1980s. This classification was developed as a simplified

version of that proposed by Ikeda” in an attempt to give the best

description to the physician responsible for the patient. The

following categories are used: (1) tumor: an endobronchial mass

presenting either as an irregular cauliflower-like vegetation or as a

lobulated fleshy growth; (2) necrosis: white necrotic coating of either

flat mucosal lesions or frankly polypoid tumors; (3) infiltration: an

irregular area ofthe mucosal surface, often swollen and surrounded

by a halo of redness and engorgement of blood vessels; (4)

compression: any form of extrinsic compression, deformation, and

narrowing of bronchi, often associated with fixity and reduced

motility, and with an apparently well-preserved mucosa; (5) aspecific

findings (AS): localized redness and vessel engorgement, without

swelling and irregularity of the muc()sal layer; and (6) normal:

absence of localized abnormalities. Paradigmatic examples of the

first four categories are shown in Figure 1 . The visual inspection

and the description of endoscopic findings were always the duty of

at least two operators, who both signed the final report. For this

study, only one type of abnormality (the most significant on the

basis of the description-report) was considered for each patient.

Bronchoscopic examinations were performed in a large, well-

equipped biopsy room using the FB (Olympus BF1T1O or BFP1O

or BF2O). Patients who had no absolute contraindications to the

procedure, such as severe bleeding diathesis, were premedicated

with atropine after an overnight fast; topical anesthesia was accom-

plished with a rhino-oropharyngeal spray of a 4 percent solution of

benoxinate (oxybuprocaine) hydrochloride (Novesine), followed by

additional instillations of the same medication through a curved

cannula and then through the suction channel ofthe same FB. Thetransnasal approach was used in nearly all patients. After a complete

inspection of the visible tracheobronchial tree, samples were taken

from areas that appeared to be abnormal. Any visible lesion was

biopsied, by brush or forceps, through the channel of the instru-

ment. In 25 cases, biopsies were obtained blindly from segments

corresponding to radiologic opacities, since there were no endo-

bronchial abnormalities. Biopsies were performed in all patients,

unless life-threatening bleeding was feared. The number of biopsy

specimens obtained for each patient was usually multiple (three to

six), unless significant bleeding followed taking the first specimen.

Brush biopsies were performed when it was impossible to obtain

reliable forceps biopsies. Then bronchial washings were taken from

the biopsy site using a saline solution. Washings were done in nearly

all patients, independently of the prior type of biopsy. Following

bronchoscopy, food and drink were withheld for two to three hours.

Bronchial washings were processed electively in the local histo-

pathologic laboratory. They were cytocentrifuged, and the sediment

was used to prepare three smears. Smears were immediately fixed

with 95 percent alcohol and stained with the standard Papanicolaou’s

technique. Bronchial brushings were obtained with the unsheathed

brush (Olympus). The brush was then withdrawn from the FB and

rubbed onto a limited area of a microscope slide with a circular

motion. Rubbing was limited to a few seconds to minimize air-

drying artifacts. Slides were then immersed in ethanol fixative for

transport to the laboratory, where staining and interpretation were

performed. Bronchial biopsies were done using a forceps (Olympus).

With cups open, the forceps was carefully placed on the more

abnormal mucosal area and pressed firmly, before closing the cups

FIGURE 1. Representative examples of tumor

(top left), necrosis (top right), infiltration (bot-

torn left), and compression (bottom right). See

text for description.

 © 1991 American College of Chest Physicians by guest on October 28, 2010chestjournal.chestpubs.orgDownloaded from

k�’�1� 1� � fi

t�iiI� � 1� � �: � H

... R:)t1 � 1 �n

�;Mh � “ 1

- I, (iCh,’(&

� . 11 � � ,‘#{149}1� � �.: -. � MIt �t1

� � I� �7iH? � �;i

� i� �

.-

600.0

500.0

400.0

3000

200.0

100.0

0

.. �

CENTRAL

SC

LCC

FIGURE 3. Distribution of histotvpes among 782 lung cancers

diagnosed using bronchial was� In parentheses is numlwr of caseswithin each histotype.

PERIPHER AL

AC5CC

other

CHEST I 99 I 4 I APRIL 1991 811

DIAGNOSTIC YIELD(no_ ofprocedure�)

t000_o � �

1sooz

I�oa -

�

. 0

FIGURE 2. Overall diagnostic yield of three bronchoscopic routine

techniques. FB, Forcep biopsies; BB, brush biopsies; and BW,

bronchial washings.

and withdrawing the forceps. If l)lee(ling occurred after taking the

first specimen, the lesion was washed with physiologic saline

solution �LII(l 0.05 percent epinephrine instilled to avoid a biopsy of

overlying clotted blood. Each biopsy sI)eciIflefl was fixed in forinal-

dehyde solution (forrnalin), enihedded in paraffin, and sectioned.

Four sections were prepared at four (lifferent levels within each

specimen and were routinely stained with hematoxylin-eosin. Ad-

ditional preparati�I1s for mucin and cytokeratin were done when

indicated.

In accordance with the 1981 UICC classification,’’ the following

histologic categones were recognized: squamous carcinoma (SC);

sniall-cell carcinoma (5CC); adenocarcinoma (AC); and large-cell

carcinonia (LCC). Adenosquannus carcinomas, ‘ along with non-

classified carcinomas, were incorporated in the waste category of

others. Although several cvtologists and pathologists participated

in the microscopic slide reading, the opinion of one cytopathologist

(SR.), wlu� reviewed nu)st specimens (including all the dubious

ones), c�nstittites the basis of the data presented in this reprt.

Cytologic specimens were classified a.s positive only when thecellular IflOq)hOlOgV contained frankly malignant characteristics;

for this report, all other interpretations, including marked atypia,

were pooled together as “negative” data.

The statistical significance of dissimilar frequencies was tested

using the x2methd. A value ofp<O.Ol was considered significant.

sCC 631i1AC I4�l

/1_CC li/I

FIGURE 4. Distribution of bronchial location among 1,045 lungcancers. LMB, Left main bronchus; LUB, left upper bronchus; LB,

lingular bronchus; LLB, left lower l)ronchus; LSB, left segmentalOI� subsegmental bronchi; RMB, right main l)ronchus, RUB, right

tipper l)ronchus; RMLB, right middle lobar bronchus; RLB, rightlower bronchus; and RSB, right segmental or sul)segmental bronchi.

Iii parentheses is number ofcases withiii each site.

RESU LTS

The study population (1 ,045 patients who under-

went bronchoscopy and ultimately were proven to

have lung cancer) had the following anthropometric

characteristics: median age, 63 years; range, 34 to 87

years; and M/F sex ratio, 8/1 (930 male patients and

115 female patients). In all, 782 pathologic diagnoses

were available.

Figure 2 shows the overall diagnostic accuracy of

the diverse techniques of sampling: forceps biopsies

were positive in 79 percent of the 841 performed

biopsies (64 percent of the entire sample); brushings

were positive in 38 percent of372 biopsies (14 percent

of all patients); and washings were positive in 32

percent of 1,009 biopsies (31 percent of 1,045). More

than one positive finding occurred in 309 bronchos-

copies out of 1 ,045 (30 percent).

Histologically, there was a very high prevalence of

SCs (Fig 3). The study of the neoplastic distribution

SC

.othor I9�l

FIcuRr. 5. Number of carcinomas of same histotype in either

central (trachea; main and lobar bronchi) or peripheral (segmentalOr subsegmental bronchi) locations. Differences were statistically

significant by x2 test (p<O.OOl).

 © 1991 American College of Chest Physicians by guest on October 28, 2010chestjournal.chestpubs.orgDownloaded from

40() 0

..SO() 0

�

2000�

Hth�1�

10ooj� ti

� �TUMOR IP’WlL.TR AS

NECROSiS COMPRESS M�RMAI.

Sc scc AC L�GC other

800.0

600.0

400.0

200.0

CENTRAL

TUMOR

INFILTR.

AS

PERIPHERAL

NECROSIS

COMPRESS.

I NORMAL

812 A Review of 1 .045 Bronchoscopic Examinations in Carcinoma (Buccheri, Barberis� De!fino)

FIcuIw 6. Number of carcinomas of same histotype in different

categories of endobronchial morphology Infiltr, infiltration; and

compress, compression. See text for description of morphologic

categories. Differences were statistically significant by � test

(i<O.001).

into the tracheobronchial tree (Fig 4) showed a definite

prevalence of tumors arising from the upper lobes

(436 vs 167 of the middle lobe and lingula and 227 of

the lower lobes; p<O.OOl). Also prevalent were loca-

tions in main and lobar bronchi (688 central tumors

vs 337 peripherally located; p<O.OOl). On the con-

trary, tumors of both sides were equally represented

(492 in the right lung vs 489 of the left; p = NS).

Histotypes influenced significantly the type of lo-

calization in the bronchi (Fig 5): indeed, 87 percent

of the SCCs and 71 percent of the SCs were located

centrally, as compared to 27 percent ofthe ACs. Large-

cell carcinonms had an intermediate behavior, being

found centrally in 65 percent of the cases. Histotypes

also influenced the endoscopic tumor morphology (Fig

6): both SSCs and SCs tended to show tumor-like

appearance; on the contrary, ACs were often associated

with no bronchial abnormalities, infiltration and aspe-

FIculu: 7. Numlwr of carcinomas of same morphologic category in

either central or peripheral locations. Infiltr, infiltration; and

compress, compression. See text for description of morphologic

categories. Differences were statisticall� significant h� x2 test

(p<O.OO1).

Table 1-Diagnostk Yield ofThree Bronchoscopk RoutineTechniques by Bronchial Location*

Positiv ity 1 Positi vity 2

Diagnostic ‘� - ,� �

Technique No. Percent No. Percent

Forceps biopsies

Central 471/586 80 4711708t 67

Peripheral 191/255 75 191/337t 57

Total 662/841 79 002/1,045 63

Brush biopsies

Central 78/225 35 781708t 11

Peripheral 64/147 44 64/337t 19

Total 142/372 38 142/1,045 14

Bronchial washings

Central 2121675 31 2121708 30

Peripheral 110/334 33 110/337 33

Total 322/1,009 32 32211,045 31

*P�sitivity 1, number (percent) of positive results out of actually

performed diagnostic techniques; and positivity 2, number �per-

cent) of positive results out of all evaluable patients.

tUsing x2 test, differences between central (trachea, and main and

lobar bronchi) and peripheral (segmental and subsegmental bron-

chi) locations were significant.

cific findings, or extrinsic compression; LCCs (and,

naturally, the nonclassified cancers) had an interme-

diate appearance (p<O.OOl). As a direct consequence

of both the aforementioned relationships, central lo-

cations were more frequently associated with tumor-

like morphology, whereas peripheral locations were

commonly associated with compressive aspects

(p<O.001; Fig 7).

The diagnostic yield of both forceps and brush

biopsies and that of bronchial lavages were finally

correlated with bronchial location (Table 1), histotype

(Table 2), and endoscopic morphology (Table 3). When

done (Table 1), the three techniques had a similar

yield independently of the location of the tumor;

however, taking into account the entire population of

the study, forceps biopsies were more effective in

Table 2-Diagnostic Yield ofThree Bronchoscopic RoutineTechniques by Histotypes4

CellType

Bio

Percent

1

psies

Percent

21

Bra

Percen

1

shings

t Percent

2

Was

Percen

it

hings

t Percent

2t

SC 96 94 57 17 44 43

Small-cell 97 92 33 10 25 24

anaplastic

AC 100 51 59 22 64 56

Large-cell 97 95 53 24 38 35

anaplastic

Other 46 39 65 26 43 40

Total 91 85 57 18 43 41

*perceflt 1, percent of positive results out of actually performed

diagnostic techniques; and percent 2, percent of positive results

Out of 782 evaluable patients.

tUsing x� test, differences among four major histotypes were

significant.

 © 1991 American College of Chest Physicians by guest on October 28, 2010chestjournal.chestpubs.orgDownloaded from

CHEST I 99 I 4 I APRIL 1991 813

Table 3-Diagnostk Yield ofThree Bronchoscopic Routine

Techniques by EndOIJrOnCIsiGJ Morphotogy*

Biopsies Brushings Washings

Morphologic Percent Percent Percent Percent Percent PercentCategories 1 2 1 2 1 2

Tumor 85 83 40 10 37 35

Necrosis 76 57 33 12 25 22

Infiltration 80 72 52 23 36 35

Compression 62 37 27 11 20 19

Aspecific 80 31 44 15 42 38

findings

Normal 71 11 17 7 37 36

Total 79 63 38 14 32 31

CAll differences were significant by x’ test. Percent 1 , percent of

positive results out of actually performed diagnostic techniques;

percent 2, percent of positive results out of 1,045 evaluable

patients.

central tumors (67 percent vs 57 percent; p<O.Ol);

and vice versa was the case for brushings (11 percent

vs 19 percent; p<O.OOl). Similar observations are

valid for the yield of forceps biopsies, as a function of

the cellular type (Table 2); accounting for all of the

patients, biopsies were more effective in central tu-

mors, 1#{128},in the SSC and the SC variants (p<(0.001).

On the contrary, brushings and, significanfly, washings

yielded more in peripheral tumors, ie, in ADs. The

LCCs were very well diagnosed by both brush and

forceps biopsies, still acting with an intermediate

behavior. The last, and more important, correlation

crable 3) shows that forceps biopsies were very effec-tive in the tumor-like morphology (83 percent of all

cases) and less effective in infiltrative appearances (72

percent). On the contrary, cytologic studies had their

maximum yield in infiltrative or quasi-infiltrative le-

sions, ie, in the nonspecific findings. In invisible

lesions, only the yield ofwashings was relatively good

(36 percent). Necrotic and compressive lesions were

associated with a quite low yield in any of the three

techniques considered.

DISCUSSION

This study has diverse results but has one message

ofparticular interest. The number ofbiopsy specimens

needed to obtain an optimal diagnostic yield is an

important consideration in applying fiberoptic bron-

choscopy. Although recommendations exist in the

literature for performing up to six forceps �7

the optimal number of specimens has not been de-

fined. In 1982, Popovich et alh8 found a diagnostic

yield of 92 percent, after six biopsies of each endo-

bronchial lesion and 40 histologic sections of each

specimen. In contradistinction, Shure et al’s could

correctly diagnose by the third biopsy all of the

carcinomas of their series. This was obtained by

examining only subjects with tumor-like endobron-

chial abnormalities and no more than nine tissue

sections. Various explanations were suggested to ac-

count for this difference in results, including the size

and type of �9 Our data indicate that the

difference might be due to another cause, ie, to the

type of endobronchial lesion selected for the study.

The more successful investigators, indeed, had limited

their analysis to those abnormalities which we have

found to be more fruitfully biopsied using the forceps

instrument. Should they have studied patients with

compressive bronchial lesions, results would have

certainly been inferior. The message from our study

is therefore that the yield of a particular technique

must be evaluated in the light of the endobronchial

morphology.

Although probably each bronchoscopist, using his

own experience, can associate particular endobron-

chial findings to the diagnostic technique which has

the best potential for success, we are unaware of the

existence of any previous study specifically directed

to quantify this association. In his classic treatise on

flexible bronchoscopy, Ikeda15 gave some insight into

the relationship existing between bronchoscopic find-ings, bronchial location, and histopathologic classifi-

cation, but not between endoscopic morphology and

diagnostic efficiency. Our study is retrospective and

has at least one important limitation. We did not

perform all three diagnostic procedures concerned on

each patient, but, instead, brush biopsies were done

when forceps biopsies were judged unfeasible. Spo-

radically, also the same forceps biopsies and bronchial

washings were not obtained. This prevented us from

comparing directly, making reference to each morpho-

logic category, the yield of the three techniques.

Nonetheless, we believe that we have demonstrated

the following: (1) cancers presenting endoscopically as

exophytic masses or frankly infiltrating the mucosal

surface are the easiest to diagnose; (2) the first pattern

is best diagnosed by forceps biopsies (85 percent

success), although brushing and washing are also

relatively effective (40 percent and 37 percent, re-

spectively); (3) the second pattern is still well diag-

nosed by tissue biopsies (80 percent), but cytologic

studies and, particularly, brushings are also effective

(this latter has its maximum yield in this type of

lesion); (4) necrosis, bronchial compression, and non-

specific findings have the lowest diagnostic potential

among visible lesions for forceps biopsies (overall 57

percent, 37 percent, and 30 percent of positivity) or

brushings (12 percent, 1 1 percent, and 16 percent) or

washings (22 percent, 19 percent, 38 percent); and (5)

also, invisible lesions, either blindly biopsied or

washed, have a nonmarginal rate of positivity.

There are other findings in the current study which

merit a comment. The overall diagnostic efficiency of

forceps biopsies, even accounting for tumors not

direcfly visible through the FB, is well comprised in

 © 1991 American College of Chest Physicians by guest on October 28, 2010chestjournal.chestpubs.orgDownloaded from

814 A Review of 1 .045 Bronchoscoplc Examinations in Carcinoma (Buccheri� Barberis, Deffino)

the range of the reported sensitivity.9”0’8’9 On the

contrary, brushings had a quite low yield, generally

inferior to the reported rates,�” but it must be

remembered that we limited our brush biopsies to

unfavorably selected cases. The overall positivity of

bronchial washings is also low, and this is partially due

to the rigid criteria we have chosen in the classification

of positive vs negative results (indeed, in order to

maintain a high specificity, smears with marked atypia

or highly suspicious for malignancy were pooled as

negative). In any case, bronchial washing was the

method with the lowest yield in other series, as well.8,b0

As far as the cell type composition of our sample is

concerned, we had a very high percentage ofSC. Also

taking into account the diagnostic source, which was

the bronchial way and not the surgical or postmortem

one, this prevalence is completely at variance with

the high incidence ofAC reported in other geographic

areas;2#{176}however, it is well known that both time trends

and variations between countries in histologic sub-

types of lung cancer may be expected on the basis of

differences in exposure to environmental carcino-

gens.2#{176}2’Our figure indeed is quite similar to the 71

percent observed in another area of northern Italy,2’

where identical life habits and environmental factors

may have played the same selective role. The predi-

lection of lung cancer for the upper lobes� and a

central location’s is well recognized and further con-

firmed by us. On the contrary, we did not find a

prevalence of tumors arising from the right lung, as

was generally reported?� As mentioned, Ikeda’5 re-

viewed the relationships between endoscopic findings,

bronchial location, and three histopathologic subtypes.

He found that SCs were often located centrally and

visualized as endobronchial masses (201 cases out of

291). On the contrary, ACs, and less LCCs, were more

often peripheral and showed mostly indirect findings,

such as bronchial obstruction and compression, or no

findings at all. These results are very close to ours.

In conclusion, in addition to having contributed the

descriptive data of a large series, we believe that we

have demonstrated that for a given endoscopic finding,

supposed to be malignant, and for a given bronchial

location, there is a more likely cell type, and a more

fruitful way to achieve the cytohistologic diagnosis. In

practice, it seems to us not so important to perform a

fixed number of both forceps and brush biopsies;

rather, the best policy might be that of modulating

their number in relation to the visible endobronchial

abnormalities; for example, one could choose to make

more forceps biopsies than brush passages in tumor-

like lesions and vice versa in mucosal infiltration,

while increasing the overall number of attempts in

necrotic and compressive bronchial findings.

ACKNOWLEDGMENT: We thank Dr. Luigi Aschero and Dr.Ferruccio Vola, bronchoscopists of our hospital, for their coopera-

tion. Professor Savino Run, Chiefofthe Laboratory of Histopathol-ogy of the Cuneo City Hospital System (USSL-58), is responsiblefor the pathologic data. Mr. James Beauchamp provided Englishediting.

REFERENCES

1 Ikeda S, Yanai N, Ishikawa S. Flexible bronchofiberscope. KeioJ Med 1968; 17:1-16

2 Credle WF Jr, Smiddy JF, Elliott RC. Complications of fiber-

optic bronchoscopy. Am Rev Respir Dis 1974; 109:67-72

3 Sackner MA. Broncofiberscopy. Am Rev Respir Dis 1975; 111:62-

88

4 Joyner LR, Scheinhorn DJ. Transbronchial forceps lung biopsythrough the fiberoptic bronchoscope. Chest 1975; 67:532-35

5 Wang KP, Terry PB, Marsh BR. Bronchoscopic needle aspiration

biopsy ofparatracheal tumors. Am Rev Respir Dis 1978; 118:17-

21

6 Wang KP, Marsh BR, Summer WR, Terry PB. Transbronchial

needle aspiration for diagnosis oflung cancer. Chest 1981; 80:48-

50

7 Linder J, Radio SJ, Robbins BA, Ghafouri MO, Rennard SI.

Bronchoalveolar lavage in the cytologic diagnosis of carcinoma

ofthe lung. Acta Cytol 1987; 31:796-8018 Solomon DA, Solliclay NH, Gracey DR. Cytology in fiberoptic

bronchoscopy: comparison of bronchial brushing, washing and

post-bronchoscopy sputum. Chest 1974; 65:616-19

9 Zavala DC. Diagnostic fiberoptic bronchoscopy: techniques and

results ofbiopsy in 600 patients. Chest 1975; 68:12-19

10 Kvale PA, Bode FR, Kini S. Diagnostic accuracy in lung cancer:comparison of techniques used in association with flexible

fiberoptic bronchoscopy. Chest 1976; 69:752-57

11 Muers MF, Boddington MM, Cole M, Murphy D, Spriggs Al.Cytological sampling at fiberoptic bronchoscopy: comparison of

catheter aspirates and brush biopsies. Thorax 1982; 37:457-61

12 Rudd RM, Gellert AR, Boldy DAR, Studdy PR, Pearson MC,

Geddes DM, et al. Bronchoscopic and percutaneous aspiration

biopsy in the diagnosis ofbronchial carcinoma cell type. Thorax

1982; 37:462-65

13 Matsuda M, Horai ‘F, Nakamura 5, Nishio H, SakumaT, Ikegami

H, et al. Bronchial brushing and bronchial biopsy: comparisonof diagnostic accuracy and cell type reliability in lung cancer.

Thorax 1986; 41:475-78

14 Payne CR, Hadfield JW, Stovin PG, Barker V, Heard BE, Stark

JE. Diagnostic accuracy of cytology and biopsy in primarybronchial carcinoma. J Clin Pathol 1981; 34:773-78

15 Ikeda S. Atlas offlexible bronchofiberscopy. Stuttgart, Germany:

Georg Thieme Verlag, 1974

16 Sobin LH, Yesner R. Histologic typing oflung tumours. Geneva:

World Health Organization, 1981.

17 Kvale PA. Collection and preparation of bronchoscopic speci-

mens. Chest 1978; 73:707-12

18 Popovich J Jr, Kvale PA, Eichenhorn MS. Radke JR, Ohorodnik

JM, Fine G. Diagnostic accuracy of multiple biopsies fromflexible fiberoptic bronchoscopy: a comparison of central versus

peripheral carcinoma. Am Rev Respir Dis 1982; 125:521-23

19 Shure D, Astarita RW. Bronchogenic carcinoma presenting as

an endobronchial mass: optimal number of biopsy specimens

for diagnosis. Chest 1983; 83:865-67

20 Vincent RG, Pickren JW, Lane WW The changing histopathol-

ogy of lung cancer: a review of 1 ,682 cases. Cancer 1977;

39:1647-55

21 Geddes M, Buiatti E, Costantini AS, Santucci M. Cell type

distribution in histological specimens of primary lung cancer.

Tumori 1984; 70:255-260

22 Byers TE, Vena JE, Bzepka TF. Predilection oflung cancer for

the upper lobes: an epidemiologic inquiry. J Natl Cancer Inst

1984; 72:1271-75

 © 1991 American College of Chest Physicians by guest on October 28, 2010chestjournal.chestpubs.orgDownloaded from

DOI 10.1378/chest.99.4.809 1991;99; 809-814Chest

G Buccheri, P Barberis and M S Delfinocarcinoma. A review of 1,045 bronchoscopic examinations.

Diagnostic, morphologic, and histopathologic correlates in bronchogenic

 October 28, 2010This information is current as of

http://chestjournal.chestpubs.org/content/99/4/809Updated Information and services can be found at:

Updated Information & Services

http://chestjournal.chestpubs.org/content/99/4/809#related-urlsThis article has been cited by 2 HighWire-hosted articles:

Cited Bys

http://www.chestpubs.org/site/misc/reprints.xhtmlonline at: Information about reproducing this article in parts (figures, tables) or in its entirety can be foundPermissions & Licensing

http://www.chestpubs.org/site/misc/reprints.xhtmlInformation about ordering reprints can be found online:

Reprints

the right of the online article.Receive free e-mail alerts when new articles cite this article. To sign up, select the "Services" link to

Citation Alerts

slide format. See any online figure for directions. articles can be downloaded for teaching purposes in PowerPointCHESTFigures that appear in Images in PowerPoint format

 © 1991 American College of Chest Physicians by guest on October 28, 2010chestjournal.chestpubs.orgDownloaded from