Correlation between endobronchial ultrasonography (EBUS) images and histologic findings in normal and tumor-invaded bronchial wall

Correlation between endobronchial ultrasonography (EBUS) images and histologic findings in normal and tumor-invaded bronchial wall

Lung Cancer 35 (2002) 65 – 71 www.elsevier.com/locate/lungcan Correlation between endobronchial ultrasonography (EBUS) images and histologic finding...

323KB Sizes 0 Downloads 59 Views

Lung Cancer 35 (2002) 65 – 71

www.elsevier.com/locate/lungcan

Correlation between endobronchial ultrasonography (EBUS) images and histologic findings in normal and tumor-invaded bronchial wall Masayuki Baba a, Yasuo Sekine a, Makoto Suzuki a, Shigetoshi Yoshida a, Kiyoshi Shibuya a, Toshihiko Iizasa a, Yukio Saitoh a, Edward K. Onuma b, Hidemi Ohwada c, Takehiko Fujisawa a,* a b

Department of Thoracic Surgery, Graduate School of Medicine, Chiba Uni6ersity, 1 -8 -1 Inohana, Chuo-ku, Chiba 260 -8670, Japan Department of Endoscopic Diagnostics and Therapeutics, Chiba Uni6ersity Hospital, 1 -8 -1 Inohana, Chuo-ku, Chiba 260 -8670, Japan c Department of Basic Pathology, Graduate School of Medicine, Chiba Uni6ersity, 1 -81 -Inohana, Chuo-ku, Chiba 260 -8670, Japan Received 3 March 2001; received in revised form 9 July 2001; accepted 11 July 2001

Abstract The aim of this study was to examine the ability of endobronchial ultrasonography (EBUS) to image the bronchial wall structure in order to assess the depth of bronchial tumor invasion. Sixty-one patients who underwent lobectomy, pneumonectomy or forceps biopsy were included in this study. In 21 patients with bronchoscopically visible bronchial malignant tumors, EBUS was performed during bronchoscopy. In the remaining 40 patients, ultrasonography was performed on the resected specimens. The EBUS findings obtained using thin ultrasonic probes (20 MHz radial scanner) were compared with the macroscopic and histologic findings of the corresponding areas in the resected specimens. When the bronchial walls were imaged while immersed in normal saline, six ultrasonically distinct layers were detected in the cartilaginous and membranous portions. A similar wall structure was imaged when EBUS was performed during bronchoscopy using a latex balloon sheath. The image of the lamina propria and submucosa was occasionally compressed and mixed with a balloon echo due to the latex balloon sheath, whereas the cartilage layer was always distinctly imaged. A good correlation was observed between the EBUS-determined cartilage thickness and the actual histologic measurement, as measured with vernier calipers. Malignant tissues were imaged as hypoechoic areas, and tumor invasion of the cartilage layer was clearly detected. In conclusion, using high-resolution (20 MHz) ultrasonic probes, the bronchial wall structure could be imaged as six distinct layers. The cartilage layer was easily identified and could be used as a reference to evaluate the rest of the bronchial wall structure. © 2002 Elsevier Science Ireland Ltd. All rights reserved. Keywords: Ultrasound; Bronchoscope; Bronchial cancer; Staging; Treatment method; Invasion

1. Introduction Ultrasonography has been reported to be an equivalent or more sensitive modality than computed tomography (CT) for detecting cancer of the esophagus [1,2] and stomach [1,3]. Endobronchial ultrasonography (EBUS) can image parabronchial lymph nodes, vessels, and mediastinal tumors, and is expected to become a valuable tool for accurately evaluating the bronchial wall structure. * Corresponding author. Tel.: + 81-43-222-7171x5464; fax: + 8143-226-2172. E-mail address: [email protected] (T. Fujisawa).

The curvilinear sector scanning ultrasonic probe was reported to be effective for imaging mediastinal lymph nodes or large vessels from the inner wall of bronchi [4]. However, the resolution of the image was not sufficient to accurately determine the bronchial wall structure. The development of thinner miniature probes [5–7], which can be inserted through a biopsy channel of the bronchoscope, has made it possible to perform EBUS during endoscopic examination of the bronchial lumen [6,8 –15]. Recently EBUS has been used for evaluating the depth of tumor invasion into the bronchial wall and the enlargement of metastatic lymph nodes in a few medical research centers. Although the thinner minia-

0169-5002/02/$ - see front matter © 2002 Elsevier Science Ireland Ltd. All rights reserved. PII: S 0 1 6 9 - 5 0 0 2 ( 0 1 ) 0 0 2 8 4 - 7

66

M. Baba et al. / Lung Cancer 35 (2002) 65–71

ture probes provide higher resolution, the accuracy of the correlation between the ultrasonographic images obtained and the actual bronchial wall structure has not been completely clarified. The purpose of this study was to investigate the ability of EBUS to image the tracheobronchial wall structure and to measure the depth of tumor invasion as shown by comparison with the actual macroscopic and histologic findings.

2. Patients and methods Sixty-one patients, including 59 with lung cancer (two metastatic), one with thymic cancer, and one with a unilateral end stage fibrotic lung, who underwent lobectomy, pneumonectomy or forceps biopsy at Chiba University Hospital from January 1996 to October, 1999 were included in this study. The patients consisted of 15 females and 46 males, with a mean age of 65.0 years (range, 34–85). The anatomical nomenclature used in this study was based on the classification of the Japan Lung Cancer Society [16] and a textbook by Zavala [17]. In 21 patients with malignant tumors located in central (main, intermediate, lobar or segmental) or subsegmental bronchial areas, EBUS was performed during bronchoscopic examination in vivo. In the remaining 40 patients who had peripheral malignant tumors or no cancer, EBUS was performed on the resected specimens ex vivo within 2 h after surgery. In this ex vivo study, the specimens were kept at 4 °C prior to ultrasonography in order to retard autolytic changes. At first, the normal bronchial wall structure was examined, and then the ultrasonography probe was used to assess eight intrapulmonary bronchial sites that were found to be invaded by malignant tumors. Informed consent was obtained from all patients prior to bronchoscopy or surgery.

2.1. Ultrasonography during bronchoscopy (in 6i6o examination) In 21 patients, ultrasonography was performed during bronchoscopy under local or general anesthesia using the newly developed thin ultrasonography probe UM-BS20-26R (combined probe and balloon sheath; outer diameter including the sheath, 2.6 mm; 20 MHz radial scanner; rotation rate 6.67/s; Olympus Optical Co. Ltd, Tokyo, Japan) and a BF-XT30 bronchoscope (outer diameter 6.1 mm; channel diameter 3.2 mm; Olympus Optical Co. Ltd, Tokyo, Japan). The advantage of UM-BS20-26R is that a balloon sheath is attached to the EBUS probe and the inflated balloon makes it possible to visualize bronchial images without instilling saline into the bronchus. The histologic diagnoses of the patients were 17 squamous cell carcinomas, two adenocarcinomas and two metastatic lung tumors

(colon and breast primary tumors). Before the EBUS examination, bronchoscopy was performed and the exact location of the site to be examined was determined and recorded. We also examined enhanced computed tomography images obtained with 5 mm thick slices to determine the extent of the tumor. The sites examined by EBUS consisted of four main bronchi, six lobar bronchi, two superior segmental branches, five segmental bronchi, and four subsegmental bronchi. After the operation, cross sections of obtained specimens were made in accordance with the EBUS images. The EBUS images were compared with the gross and histologic views of the same areas in the resected specimens in 20 patients after pulmonary resection. In one patient with microinvasive squamous cell carcinoma, EBUS images were compared only with histologic views of a transbronchial biopsy specimen.

2.2. Ultrasonography of the resected specimens (ex 6i6o examination) A total of 60 sites (52 normal areas and eight areas associated with tumors) in 40 patients were imaged in the resected specimens. The sites examined included four main stem bronchi, three intermediate bronchi, 12 lobar bronchi, five superior and four lingular segmental branches, 10 basal bronchi, 19 segmental bronchi and three subsegmental bronchi. The eight cases with tumor invasion were identified as five squamous cell carcinomas, two small cell lung carcinomas and one large cell carcinoma. In the majority of the cases, the EUM30 processor and the UM-3R ultrasonic probe (20 MHz radial scanner; rotation rate 6.67/s; Olympus Optical Co. Ltd, Tokyo, Japan) were used for the ultrasonic probe system. In the early cases, the TMES2000 ultrasonic Micro-Endoprobe System, which consists of a processor and a PMD-2024 ultrasonic probe (20 MHz radial scanner; rotating rate 15/s; Toshiba Medical Corp., Tokyo, Japan) [7] was used. In both systems, the ultrasonic probe was 2.5 mm in diameter and could pass through a bronchoscopic biopsy channel 2.8 mm or more in diameter. All three ultrasonic probes we used had 20 MHz transducers and the same resolutions. The brightness and contrast of the EBUS images were adjusted to obtain the best quality. To optimize the ability of EBUS to image the bronchial wall, 33 sites were examined while the resected specimens were immersed in normal saline. Thirty sites were examined using a latex balloon sheath MH-246R (outer diameter 3.5 mm, Olympus Optical Co. Ltd, Tokyo, Japan) filled with distilled water, which could be inserted through the biopsy channel of the BF-ST30 bronchoscope (outer diameter 6.1 mm; biopsy channel diameter 3.7 mm; Olympus Optical Co. Ltd, Tokyo, Japan) without saline immersion. Three sites out of 60 examined areas were imaged by using

M. Baba et al. / Lung Cancer 35 (2002) 65–71

67

both methods. The probes were inserted into the bronchial lumen and positioned parallel to the longitudinal axis of the bronchi during imaging. The EBUS images were compared with the gross and histologic views of the same slice. To determine which ultrasonic layers corresponded to the cartilage layer, two 27G needles were inserted along the inner and outer margins of the cartilage layer in five resected specimens. The specimens were immersed in normal saline, or the bronchial lumen of the specimens was obstructed with the latex balloon sheath filled with distilled water, and then ultrasonography was performed. In a similar fashion, 27G needles were inserted along the inner and outer margins of the muscle layer in the membranous portions.

2.3. Thickness of bronchial cartilage measured by EBUS and the correlation with the actual histologic measurement To investigate the accuracy of EBUS, the EBUS-determined thickness of 53 bronchial cartilage layers was compared with the actual cartilage thickness, as measured with vernier calipers, of the same sites in resected specimens. The studied bronchial sites consisted of the main (four sites), intermediate (two sites), and upper lobe (sven sites) and lower lobe (two sites) bronchi, the superior (five sites), lingular (four sites) and basal (nine sites) segmental branches, and segmental (17 sites) and subsegmental (three sites) bronchi. The mean EBUS-determined thickness and the mean actual thickness were compared using Student’s t-test. The correlation between the EBUS-determined thickness and the actual thickness was determined by simple linear regression analysis. StatView Ver. 5.5 software (SAS Institute Inc., Cary, NC) was used for statistical analysis, and P B 0.05 was considered significant.

Fig. 1. EBUS image obtained from a specimen immersed in normal saline (A) with corresponding macroscopic view (B) of the left lingular bronchus in a 66-year-old male. The number of each bronchial layer in (B) corresponds to the same number of the EBUS layers in (A).

microscopic findings (Fig. 1). We also confirmed that the third and fourth layers were cartilage by insertion of 27G needles which produced hyperechoic signals (Fig. 2). When the latex balloon sheath was inflated and attached to the bronchial wall, the second layer, which corresponded to the lamina propria, smooth muscle and extramuscle layers, was sometimes compressed and mixed with the balloon echo. As cardiac contraction affected the EBUS images during bronchoscopy, the

3. Results

3.1. EBUS findings of normal intrapulmonary bronchi Ultrasonography of normal intrapulmonary bronchi, including segmental and more peripheral bronchi, immersed in normal saline revealed six distinct layers. The first, third and fifth layers were hyperechoic due to interface echo [5,7,18], the second and fourth layers were hypoechoic, and the sixth layer was slightly hyperechoic. The EBUS images of bronchi immersed in normal saline accurately imaged the wall structure of normal intrapulmonary bronchi; i.e. the first and second layers: epithelium, lamina propria, and submucosa, the third and fourth layers: cartilage, and the fifth and sixth layers: adventitia. These layers were confirmed by

Fig. 2. EBUS images of the right upper bronchus in an 85-year-old male using a latex balloon in the resected specimen. The outer (A) and inner (B) margins of the cartilage layer are identified by two 27G needles inserted along the corresponding margins. Arrows indicate the beginning of the ultrasonic tails produced by the needles.

68

M. Baba et al. / Lung Cancer 35 (2002) 65–71

perechoic. In contrast to the cartilaginous portion, the third and fourth layers corresponded to longitudinal muscle layer [19] and bronchial glands instead of cartilage. Each layer was confirmed by microscopic findings and 27G needle insertion (Fig. 3). When the latex balloon sheath was inflated and attached to the bronchial wall without saline immersion, the first and second layers represented the latex balloon echo as well as epithelium, lamina propria, and submucosa. The other layers were the same as those seen by EBUS without a balloon sheath.

3.3. EBUS findings of the bronchial walls associated with malignant tumors Among 60 areas examined in 40 patients who had peripheral malignant tumors or no cancer, EBUS was performed in a total of eight intrapulmonary bronchial sites which were invaded by malignant tu-

Fig. 3. EBUS images of the membranous portion using a latex balloon in the right main stem bronchus of a 72-year-old male. The membranous portion is limited to the region between the two solid arrowheads (A). Thin 27G needles were inserted in the resected specimen along the submucosa and adventitia in order to confirm these layers on the EBUS image. Open arrows indicate the beginning of the ultrasonic tails caused by the needles in the submucosa (A) and adventitia (B). The cartilage layer is indicated by open arrowheads at the edge of the cartilaginous portion (A). The histologic view of the membranous portion shows the longitudinal muscle layer and bronchial glands (open arrowheads), which correspond to the third and fourth EBUS layers (C) (C, low power, H&E). The number of each bronchial layer in C corresponds to the same number of the EBUS layers in A.

images of resected specimens immersed in normal saline (see Figs. 1A, 2 and 3) were clearer than those obtained by in vivo examination (see Fig. 4A and Fig. 5C). However, the EBUS images obtained by in vivo examination were clear enough to distinguish each layer of the bronchi visually.

3.2. EBUS findings of normal extrapulmonary bronchi In the extrapulmonary bronchi, which consist of the main, intermediate, upper lobe and lower lobe bronchi, the EBUS images were basically similar to those of the intrapulmonary bronchi. However, there was an additional finding: a membranous portion lacking cartilage was seen. As in the cartilaginous portion, six ultrasonically distinct layers were imaged in the membranous portion while the specimen was immersed in normal saline. The first, third and fifth layers were hyperechoic, the second and fourth layers were hypoechoic, and the sixth layer was slightly hy-

Fig. 4. The EBUS image (A) and corresponding histologic (B) view of the right B10b, c bronchus in a 58-year-old male with squamous cell carcinoma. The open arrows indicate the cartilage layers (A and B). In the EBUS image (A), the solid arrowheads indicate a hypoechoic area disrupting the cartilage layer, which corresponds to tumor tissue. Tumor invasion beyond the cartilage layer was detected in the EBUS image (A) and is clearly shown in the histologic views (B). Lymphatic tissue (Ly in A and B) is imaged as hypoechoic areas in the EBUS image (A) (B, low power, H&E).

M. Baba et al. / Lung Cancer 35 (2002) 65–71

mors. As the ultrasound radiation of the 20 MHz transducer reached 1.5– 2 cm in depth, EBUS could visualize adjacent structures including lymph nodes and vessels as well as the total bronchial wall thickness. However, since EBUS produces a radial scanning image, it could evaluate only the depth of tumor invasion based on the imaging of the cartilage layer, but not the width and size of the tumor. The tumor tissue appeared hypoechoic, and invasion of the bronchial wall was clearly detected as a disruption of the cartilage layer (Fig. 4). However, the EBUS images could not differentiate the histologic type. A total of 21 bronchial sites were examined clinically during bronchoscopy. All 21 tumor sites, for which the endoscopic findings consisted of five thickened, nine

69

nodular, and seven polypoid types, were judged as having no invasion to the cartilage layer by EBUS. All of these sites were confirmed histologically to have involvement of the cartilage layer. However, EBUS was not able to image the basement membrane clearly enough to differentiate between carcinoma in situ and submucosal invasion (Fig. 5).

3.4. The EBUS-determined thickness of the bronchial cartilage, and its correlation with the actual histologic measurement The mean thickness of the cartilage in the resected bronchial walls, as measured with vernier calipers, was 1.03 mm (range, 0.62–1.75 mm). The mean thickness of the ultrasonographic measurements of the same cartilage was 1.01 mm (range, 0.63–1.90 mm). A good correlation was observed between the EBUS-determined thickness of the bronchial cartilage and the actual histologic measurement (correlation ratio of 0.773, PB 0.0001) (Fig. 6).

4. Discussion The first report about EBUS [6] using an ultrasonic thin probe with a radial scanning transducer described the bronchial wall as a three-layered structure. We believe that the first hyperechoic layer in that initial report corresponded to our first three layers, the second hypoechoic layer to our fourth hypoechoic layer, and the third hyperechoic layer to our fifth hyperechoic and sixth slightly hyperechoic layers, respectively. Because the EBUS in that initial report attempted to describe not only bronchial wall structure but also adjacent vessels and lymph nodes, lower contrast and higher gain were used in their setup. This may be the reason the hypoechoic submucosal and muscle layers and the hyperechoic elastic layer were not clearly imaged in that study.

Fig. 5. A bronchial carcinoma found in a 67-year-old male. Bronchoscopy revealed necrotic tissue and erythematous thickening of the mucosa (open arrow) in the bifurcation of the left upper division (U) and the lingual bronchus (L) in (A). Histology of the biopsied specimen revealed microinvasive early squamous cell carcinoma (B, low power, H&E). By EBUS examination in vivo (C, which is a magnified view of the tumor-containing area), the bronchial cartilage (open arrows) was shown to be intact and free from tumor invasion, and the tumor tissue was seen as a hypoechoic area (arrowheads) on the inner side of the cartilage layer. The patient was a poor operative candidate due to ventricular arrhythmia and successfully underwent radiotherapy.

Fig. 6. Correlation between the actual thickness of the cartilage and the corresponding thickness measured by EBUS.

70

M. Baba et al. / Lung Cancer 35 (2002) 65–71

Using high frequency 20 MHz probes with improved resolution, and using a high contrast setting, EBUS is able to image the bronchial wall structure very clearly, especially in specimens immersed in saline. It has been reported that seven-layered ultrasonographic images were obtained from examinations of animal or human tracheal models whose wall structure should be similar to that of the extrapulmonary bronchial wall. However, those images were not well analyzed with respect to the concept of interface echo [10,12]. Ultrasound is usually reflected at an interface of tissues, and each ultrasonogenic layer is not necessarily correlated with a different tissue layer. Due to the increased thickness of the first interface echo, it is difficult to accurately measure the thickness of the epithelium by EBUS. The measured thickness of the first hyperechoic layer is much thicker than the actual thickness of the epithelium in various gastrointestinal organs [5,7,18] and in lower respiratory tract airways [13]. Although a five-layered bronchial wall structure imaged by EBUS was recently reported [13– 15], the authors described an additional slightly hyperechoic layer adjacent to the inner hyperechoic layer of the adventitia. The adventitia consists of not only a thin fibrous layer but also loose connective tissue adjacent to cartilage. Unlike cartilage, this loose connective tissue is not homogeneous. In the present study, adventitia was imaged as a combination of hyperechoic and slightly hyperechoic layers. In the membranous portion of the extrapulmonary bronchi, six ultrasonographic layers were also imaged. The longitudinal smooth muscle in the membranous portion was imaged as a hypoechoic structure similar to the cartilage layer in the cartilaginous portion. From the above results, the interface echos of the cartilage in the cartilaginous portion and the longitudinal muscle layers in the membranous portion appear to be a good baseline to assess the depth of tumor invasion in the tracheobronchial wall. EBUS has some limitations. First of all, this new diagnostic modality requires at least 10– 15 manipulations and adjustments of the EBUS probe obtain clear images. An experienced operator, however, can reduce the average examination time to approximately 10 min. Secondly, it is difficult to image the bronchial wall structure using a radial type ultrasonic probe at bronchial bifurcations or in the small bronchi (subsegmental or more peripheral bronchi), where the lumen is not large enough to allow the balloon to inflate and to keep the probe away from the bronchial wall. It has been reported that lymphatic tissue is difficult to distinguish from tumor tissue by EBUS, because both tissues show hypoechoic signals [14]. The submucosal structure is in rare cases not clearly imaged with the latex balloon sheath (especially in vivo), and the bronchial wall structure is in some cases not clearly visualized due to tumor invasion. Also in some cases of carcinoma in situ or submucosal

invasion, it is difficult to image the tumor, because the image of the balloon sheath is thicker than that of the mucosal and submucosal layers, and the tumor image is hidden in the balloon sheath. However, the cartilage layer is easily identified by EBUS, and tumor invasion through the cartilage layer can be detected. The management of central-type bronchial cancer depends on the depth of tumor invasion. Lesions which invade through the cartilage layer should be surgically resected, and external radiotherapy is an alternative in patients who are physiologically poor operative candidates. Lesions which have not invaded the cartilage layer can be treated surgically or with intraluminal radiotherapy [20]. Other options include laser ablation [21] or photodynamic therapy (PDT) [22]. Therefore, accurate assessment of the depth of tumor invasion is important for determining an appropriate therapy. In our study, one patient was diagnosed with microinvasive early squamous cell carcinoma by EBUS, and was successfully treated with radiotherapy.

5. Conclusions EBUS imaged the bronchial wall as six ultrasonographic layers. Using the cartilage layer as a reference, the remainder of the bronchial wall anatomy could be evaluated and the depth of tumor invasion in the bronchial wall could be determined.

Acknowledgements The authors wish to thank Drs Hiromitsu Saisho, Toshio Tsuyuguchi, Akira Iyoda, Tetsuya Toyozaki, and Kenzo Hiroshima for their invaluable comments, and Mrs Kazuko Abe for preparing the histologic specimens. This work was supported in part by Grant-in-aid for Scientific Research (C)(2) c 10671240 of the Japanese Ministry of Education, Science, Sports and Culture.

References [1] Ziegler K, Sanft C, Semsch B, Friedrich M, Gregor M, Riecken EO. Endosonography is superior to computed tomography in staging tumors of the oesophagus and cardia (abstract). Gastroenterology 1988;94(Suppl.):A517. [2] Botet JF, Lightdale CJ, Zauber AG, Gerdes H, Urmacher C, Brennan MF. Preoperative staging of esophageal cancer: comparison of endoscopic US and dynamic CT. Radiology 1991;181:419 – 25. [3] Botet JF, Lightdale CJ, Zauber AG, et al. Preoperative staging of gastric cancer: comparison of endoscopic US and dynamic CT. Radiology 1991;181:426 – 32. [4] Ono R, Suemasu K, Matsunaka T. Bronchoscopic ultrasonography in the diagnosis of lung cancer. Jpn J Clin Oncol 1993;23:34 – 40.

M. Baba et al. / Lung Cancer 35 (2002) 65–71 [5] Silverstein FE, Martin RW, Kimmey MB, Jiranek GC, Franklin DW, Proctor A. Experimental evaluation of an endoscopic ultrasound probe: in vitro and in vivo canine studies. Gastroenterology 1989;96:1058 – 62. [6] Hu¨ rter T, Hanrath P. Endobronchial sonography: feasibility and preliminary results. Thorax 1992;47:565 –7. [7] Saisho H, Sai K, Tsuyuguchi T, Yamaguchi T, Matsutani S, Ohto M. A new small probe for ultrasound imaging via conventional endoscope. Gastrointest Endosc 1995;41:141 –5. [8] Goldberg BB, Steiner RM, Liu JB, et al. US-assisted bronchoscopy with use of miniature transducer-containing catheters. Radiology 1994;190:233 – 7. [9] Steiner RM, Liu JB, Goldberg BB, Cohn JR. Interventional pulmonology: the value of ultrasound-guided fiberoptic bronchoscopy. Clin Chest Med 1995;16:519 –34. [10] Lam S, Becker HD. Thoracic endoscopy: future diagnostic procedures. Chest Surg Clin North Am 1996;6:363 – 79. [11] Shannon JJ, Bude RO, Orens JB, et al. Endobronchial ultrasound-guided needle aspiration of mediastinal adenopathy. Am J Respir Crit Care Med 1996;153:1424 – 30. [12] Steiner RM, Becker HD, Liu JB, Goldberg BB. Lungs and mediastinum. In: Liu JB, Goldberg BB, editors. Endoluminal ultrasound: Vascular and nonvascular applications. London: Dunitz, 1998:251 – 73. [13] Kurimoto N, Murayama M, Morita K, Kabayashi A, Uomoto M, Nishizaka T. Clinical applications of endobronchial ultrasonography in lung disease. Endoscopy 1998;30(Suppl. 1):A8 – A12. [14] Kurimoto N, Murayama M, Yoshioka S, Nishisaka T, Inai K, Dohi K. Assessment of usefullness of endobronchial ultrasonog-

[15]

[16]

[17]

[18]

[19]

[20]

[21]

[22]

71

raphy in determination of depth of tracheobronchial invasion. Chest 1999;115:1500 – 6. Tanaka F, Muro K, Yamasaki S, et al. Evaluation of tracheobronchial wall invasion using transbronchial ultrasonography (TBUS). Eur J Cardio-thorac Surg 2000;17:570 – 4. The Japan Lung Cancer Society, Kato H. Anatomy. In: Classification of Lung Cancer (First English Edition). Tokyo: Kanehara, 2000. Zavala DC. Flexible Fiberoptic Bronchoscopy: A training Handbook. Iowa City: University of Iowa, Department of Publications, 1978. Tamada K, Ido K, Ueno N, Kimura K, Ichiyama M, Tomiyama T. Preoperative staging of extrahepatic bile duct cancer with intraductal ultrasonography. Am J Gastroenterol 1995;90:239 – 46. Iizuka K, Dobashi K, Houjou S, Sakai H, Itoh K, Nakazawa T. Evaluation of airway smooth muscle contractions in vitro by high-frequency ultrasonic imaging. Chest 1992;102:1251 –7. Ono R. Indications for bronchoscopic brachytherapy and conditions for irradiation. In: Ono R, editor. Brachytherapy. Tokyo: Nakayama-Shoten, 1995:66 – 78. Fujisawa T, Yamaguchi Y, Saitoh Y, et al. Endoscopic Nd: YAG laser treatment as a curative modality for benign neoplasms of the tracheobronchus. Ann Thorac Cardiovasc Surg 1996;2:411 – 6. Kawaguchi T, Yamamoto S, Naka N, et al. Immunohistochemical analysis of Bcl-2 protein in early squamous cell carcinoma of the bronchus treated with photodynamic therapy. Br J Cancer 2000;82:418 – 23.