Thorax 1992;47:565-567

New techniques

Endobronchial sonography: feasibility andpreliminary results

Thomas Hurter, Peter Hanrath

AbstractEndobronchial sonography, a new

ultrasound technique, has beenevaluated for the assessment of normallungs and bronchial carcinomas. Theprocedure was performed with ultra-sound catheters, which were introducedinto central and peripheral bronchithrough the operating channel of fibre-optic bronchoscopes. The bronchial wallis highly echogenic and laminated. Thelung parenchyma appears echo rich andpatchy. Pulmonary arteries can beidentified by the pulsatile changes andfloating echoes within the echo freelumen. Echo poor bronchial carcinomaswere detected in 69 out of 74 patientswith endoscopically visible tumours andin 19 out of 26 patients with peripheralcarcinomas. The correct implantation ofmetallic stents was facilitated by endo-bronchial sonography in nine patients.The sonographic examination carried no

particular risk and caused little dis-comfort.

Transthoracic or transoesophageal sono-

graphic approaches'-' to the lungs and media-stinum are of limited value for the diagnosis ofbronchial carcinoma. This is due to the inter-position of air in lung parenchyma andbronchi between the transducer and thetumour. To overcome these diagnosticlimitations we have assessed whether an

intrabronchial approach using ultrasoundcatheters is useful. In this report the technicaldetails and the sonomorphology of normallungs and bronchial carcinomas are presented.

Medizinische Klinik I,

Rheinisch-West-fiilische TechnischeHochschule Aachen,D-5100 Aachen,GermanyT HurterP HanrathReprint requests to:Dr T Hurter

Accepted 10 April 1992

MethodsEndobronchial sonography with an ultrasoundcatheter was applied to 100 consecutivepatients (77 male, 23 female; mean age 59 (SD35-81) years) during routine fibreoptic bron-choscopy from June 1990 to February 1991. Allpatients had bronchial carcinomas (51squamous cell, 26 small cell, and 14 large cellcarcinomas and nine adenocarcinomas). Thediagnosis was confirmed by forceps biopsy(n = 62), transbronchial needle aspiration(n = 9), brushing (n = 5), or transthoracicneedle biopsy (n = 24). In all cases chestradiographs or computed tomograms had sug-

gested bronchial carcinoma. All patients gave

their written informed consent to the combinedendoscopic-sonographic examination.The endoscopy was performed after a sub-

cutaneous injection of 15 mg hydrocodone and0 5 mg atropine as an antitussive as well as

intravenous sedation with 2-5-5 mgmidazolam. For mucosal anaesthesia 10 ml of4% lignocaine was inhaled from an ultrasonicnebuliser. During inspection up to 10 ml of2%lignocaine was applied. The bronchoscopicexamination started with the inspection of thecentral airways. Tumours were categorisedendoscopically as central (n = 74), when therewere typical signs ofendobronchial (n = 56) or

peribronchial (n = 18) growth, or peripheral,when endoscopic inspection revealed no

abnormalities (n = 26). Necropsy studies wereperformed in eight patients and the lungsinvestigated sonographically in a waterbathwithin 24 hours of death.

Endobronchial sonography was performedwith 4-8 and 6-2 F catheters (Boston Scientific,Watertown, Massachusetts), which are flexibleenough to be passed into all lobar segments viathe operating channel of a fibreoptic broncho-scope (Olympus BF1T20D). These ultrasoundcatheters consist ofa single element transducer,which is placed at the tip of a metal shaft (fig 1,

a

b

c

d

k ---

.:.-l-

o:waai9-N-

J__

Figure 1 Ultrasound catheters for endobronchialsonography. The transducer (a, arrow) rotates within a

polyethylene sheath (b) filled with water. A waterinsufflated latex balloon (c) is used in large bronchi, whilebiopsy specimensfrom the peripheral lesions are

accomplished by an 8 F guiding catheter (d, arrow) forthe 4-8 F ultrasound catheter and theflexible forceps(arrowhead). Top: a match to show the scale.

Editor: PMA Calverley

565

on Decem

ber 31, 2019 by guest. Protected by copyright.

http://thorax.bmj.com

/T

horax: first published as 10.1136/thx.47.7.565 on 1 July 1992. Dow

nloaded from

Hurter, Hanrath

Figure 2 Macroscopicappearance of a necropsyspecimen showing normallung tissue.PA-pulmonary artery;B-bronchus.

a), rotating within a polyethylene cathetersheath filled with sterile water (fig 1, b). Stenoticcentral bronchi and peripheral parts of thebronchial system were assessed with the ultra-sound catheter alone. Good ultrasonic couplingbetween the mucosa and the ultrasound cath-eter was facilitated by bronchial secretions. Insome cases only the instillation of0 8% sodiumchloride solution, as used in bronchoalveolarlavage, was necessary. The normal sized main,lobar, and segmpental bronchi were assessed inthe last eight patients with the help of a waterfilled latex balloon at the tip ofthe catheter (fig 1,c), which allowed the whole circumference tobe visualised undisturbed by air. Peripheralbiopsy specimens were obtained with a 4 8 Fultrasound catheter simultaneously passed intothe tumour within an 8 F sheath through theworking channel of a larger bronchoscope(Olympus BFXT20). After sonographic local-isation of the carcinoma the sheath was left inplace, which thus allowed a biopsy to beperformed at the correct place after retractionof the ultrasound catheter (fig 1, d). Thistechnique was used in the last six patients withperipheral carcinomas, making a successfulbiopsy without fluoroscopic control possible.The procedure took from one to 20 minutes,

according to the technique used. When theultrasound catheter alone was used sonographywas performed as quickly as forceps biopsy.Central carcinomas were assessed after theendoscopic tumour had been localised. Duringthe initial phase of these investigations allsonographic examinations of peripheral lesionswere controlled by fluoroscopy. In the sono-graphic assessment the healthy contralaterallungs served as controls.The ultrasound frequency of the transducer

is 20 MHz, the sector 3600 perpendicular to thelongitudinal axis of the catheter. The max-imum field of view is 5 cm in diameter underoptimal conditions. The limit of the radialresolution is of the order of 200 ,um. The twodimensional image is analysed in real time on avideo monitor of the imaging console (DRF

1000, Diasonics, Milpitas, California). Forcontrast sonography of the pulmonary arteries10 ml oxypolygelatine is injected into the left orright cubital vein.No serious complications occurred. In some

cases the ultrasound catheter caused minimalbleeding, which stopped spontaneously.Patients sometimes noted more coughingduring sonography, especially after waterinsufflation of the balloon.

ResultsThe bronchial wall is highly echogenic and hasa unilaminar and a trilaminar structure inadjacent parts of the same bronchus.Trilaminar areas are composed ofan inner echodense layer followed by an intermediateecholucent zone of nearly the same diameter.These structures are surrounded by a thin echodense band. Comparative in vitro studiesshowed that the three layered pattern corre-sponds to the regions with cartilage (figs 2 and3), whereas the unilaminar structure is found inthe intercartilagenous areas. Bronchial secre-tions appear echo free except for echo rich airbubbles moving within the lumen. Pulmonaryarteries in close proximity to the bronchial wallcan be identified in vivo by pulsatile changes oftheir diameter, spontaneous echoes within theecho free lumen, or microbubbles passingthrough the vessel after intravenous injectionof an echo contrast medium. The artery wall isthin and echo rich. Both bronchi and pulmon-ary arteries are surrounded by lung paren-chyma, which is characterised by its echo rich,patchy sonographic pattern.

Bronchial carcinomas of all histological typeshave an echo poor texture both in vivo and invitro and thus can be differentiated from thehighly echogenic normal bronchial wall and the

Figure 3 Ultrasound picture of the specimen shown infigure 2. The cup shapedpulmonary artery (PA) has anecho rich, thin layered, #nilamellar wall (arrowheads),in contrast to the bronchial wall, which appearstrilaminated (arrows) with an inner and outer echo richlayer and an echo poor lamina between. Both the waterfilled lumina of the pulmonary artery and bronchi are echofree. C-catheter.

566

on Decem

ber 31, 2019 by guest. Protected by copyright.

http://thorax.bmj.com

/T

horax: first published as 10.1136/thx.47.7.565 on 1 July 1992. Dow

nloaded from

Endobronchial sonography: feasibility and preliminary results

Figure 4 Computedtomogramfrom a 58 yearold patient showing a 2 cmnodule (arrow) in the leftupper lobe.

surrounding lung parenchyma (figs 4 and 5).Endobronchial sonography identifies endo-bronchial and peribronchial tumour growth aswell as infiltration of the bronchial wall. Longi-tudinal and radial tumour extensions can bemeasured by retracting the catheter within thebronchus. Sixty nine out of 74 central carcin-omas, which had been localised endoscopically,were identified by endobronchial sonography.In the remaining five patients peribronchialtumour growth had completely occluded thebronchial lumen, thus making passage andsonographic assessment impossible. Peripheraltumours, which could not be identified bybronchoscopy alone, were seen sonographicallyin 19 out of 26 cases. The remaining sevencarcinomas were missed because the ultra-sound catheter failed to come into close enoughcontact with the lesion. In all patients inves-tigated no echo poor areas of tissue were foundin contralateral lungs when radiograph andendoscopy had shown normal lung structure-that is, there were no false positive results.

In nine patients with peribronchial tumourgrowth self expandable metallic stents (Wall-stent, Schneider, Zuirich) were implanted fordrainage of abscesses (n = 2), resolution ofpneumonia (n = 2), or the relief of dyspnoea(n = 5). In all ofthese cases sonography was an

adjunct to endoscopy because it allowed thepreinterventional measurement of the lengthand diameter of the stenotic area, which couldnot be done by the fibreoptic bronchoscope.

Figure 5 Endobronchialsonogram from the patientoffigure 4, showing theecho poor bronchialcarcinoma (BC) as well asthe echo rich lungparenchyma (LP) andbronchial wall(arrowheads). Anadenocarcinoma wasdiagnosed byendobronchial brushing.

DiscussionThe sonographic examination of bronchial car-cinoma is still of limited value owing to theinterposition of air within the bronchi and lungparenchyma between the transducer and thetumour. Thus both with the transthoracic andwith the transoesophageal approach carcin-omas are visible only in patients with tumoursofthe mediastinum or pleura.`5 The aim of thisstudy was to test whether this limitation can becircumvented by an endoluminal approach.For this purpose we used an ultrasound cath-eter that was originally designed for vascularexaminations.67 These small sized catheterswere appropriate even in patients with bron-chial stenosis or endoscopically invisibletumours.To judge by our initial experience endobron-

chial sonography can be successfully perfor-med in most patients during routine fibreopticbronchoscopy. The method was well toleratedby the patients because it caused only milddiscomfort. In addition, the technique wasreadily accepted by the investigators as it wassimple and had no complications.Endobronchial sonography allows the echo

poor bronchial carcinoma and the highlyechogenic normal bronchial wall or lung paren-chyma to be differentiated, as shown in vitroand in vivo. Until now we have not had anyexperience with benign solid masses, whichmay contribute to false positive results.Endobronchial sonography definitely cannotdistinguish between different cell types of lungcancer.According to our preliminary results,

endobronchial sonography may be an alter-native to fluoroscopy for localising peripheraltumours. It can measure the length anddiameter of bronchial stenoses that cannot beassessed by inspection or radiography.If computed tomography is not available, itmay help to identify large vessels near tumours.Further studies are necessary to define itsprecise role here.

We wish to thank Dr J Ammon, director of the radiotherapyclinic, -University Clinic Aachen, for providing the chestradiograph.

1 Izumi S, Tamaki S, Natori H, Kira S. Ultrasonically guidedaspiration needle biopsy in disease of the chest. Am RevRespir Dis 1982;125:460-4.

2 Kondo D, Imaizumi M, Abe T, Naruke T, Suemasu K.Endoscopic ultrasound examination for mediastinallymph node metastases of lung cancer. Chest 1990;98:586-93.

3 Pang JA, Tsang V, Hom BL, Metreweli C. Ultrasound-guided tissue-core biopsy of thoracic lesions with Trucutand Surecut needles. Chest 1987;91:823-8.

4 Wernecke K, Vassallo P, Potter R, Luckener HG, Peters PE.Mediastinal tumors: sensitivity of detection with sonogra-phy compared with CT and radiography. Radiology1990;175:137-43.

5 Yang PC, Luh KT, Sheu JC, Kuo SH, Yang SP. Peripheralpulmonary lesions: ultrasonography and ultrasonicallyguided aspiration biopsy. Radiology 1985;155:451-6.

6 Isner JM, Rosenfield K, Losordo DW, Kelly S, Palefski P,Langevin RE, et al. Percutaneous intravascular US asadjunct to catheter-based interventions: preliminaryexperience in patients with peripheral vascular disease.Radiology 1990;175:61-70.

7 PandianNG, Weintraub A, Kreis A, Schwartz SL, KonstamMA, Salem DN. Intracardiac, intravascular, two-dimensional, high-frequency ultrasound imaging ofpulmonary artery and its branches in humans and animals.Circulation 1990;81:2007-12.

567

on Decem

ber 31, 2019 by guest. Protected by copyright.

http://thorax.bmj.com

/T

horax: first published as 10.1136/thx.47.7.565 on 1 July 1992. Dow

nloaded from