Factors Related to Diagnostic Sensitivity Using an Ultrathin Bronchoscope Under CT Guidance*

Transcription

1 Original Research INTERVENTIONAL PULMONOLOGY Factors Related to Diagnostic Sensitivity Using an Ultrathin Bronchoscope Under CT Guidance* Naofumi Shinagawa, MD, PhD; Koichi Yamazaki, MD, PhD; Yuya Onodera, MD, PhD; Hajime Asahina, MD; Eiki Kikuchi, MD; Fumihiro Asano, MD, PhD; Kazuo Miyasaka, MD, PhD; and Masaharu Nishimura, MD, PhD Background: We investigated factors related to the diagnostic sensitivity of CT-guided transbronchial biopsy (TBB) using an ultrathin bronchoscope and virtual bronchoscopy (VB) navigation for small peripheral pulmonary lesions. Method: We have performed this procedure on 83 patients with 85 small peripheral pulmonary lesions (< 20 mm in diameter). We analyzed the relationship between the diagnostic sensitivity and the location of the lesions, the bronchial generation to which an ultrathin bronchoscope was inserted, and the lesion-bronchial and lesion-pulmonary arterial relationships on high-resolution CT. Results: Fifty-six of the 85 lesions (66%) were diagnosed following CT-guided TBB using an ultrathin bronchoscope with VB navigation. The lesions located in the left superior segment of the lower lobe (S 6 ) had a significantly low diagnostic sensitivity compared to other locations (p < 0.01). When an ultrathin bronchoscope could be inserted to the fifth or greater bronchial generation, the yield was above the average diagnostic sensitivity of 66%. Moreover, not only the patients with the presence of a bronchus leading directly to a lesion (CT-bronchus sign), but also the patients with the presence of a pulmonary artery leading to a lesion (CT-artery sign), had high diagnostic sensitivity (p < 0.01). Multivariate analysis revealed that the location of lesion was an independent predictor of diagnostic sensitivity (p < 0.05). Conclusions: The location of the lesion, the bronchial generation to which an ultrathin bronchoscope was inserted, and the presence of a bronchus as well as a pulmonary artery leading to the lesion were valuable for predicting successful CT-guided TBB using an ultrathin bronchoscope with VB navigation. (CHEST 2007; 131: ) Key words: CT artery sign; CT-guided transbronchial biopsy; small peripheral pulmonary lesion; ultrathin bronchoscope; virtual bronchoscopic navigation Abbreviations: FB flexible bronchoscopy; HRCT high-resolution CT; TBB transbronchial biopsy; VB virtual bronchoscopy Recent advances in CT equipment have increased the detection rate of small pulmonary peripheral lesions. For diagnosing these lesions, the transbronchial approach using flexible bronchoscopy (FB) remains one of the most feasible methods. However, *From First Department of Medicine (Drs. Shinagawa, Yamazaki, Asahina, Kikuchi, and Nishimura) and Department of Radiology (Drs. Onodera and Miyasaka), Hokkaido University School of Medicine, Sapporo; and Department of Pulmonary Medicine and Interventional Bronchoscopy (Dr. Asano), Gifu Prefectural General Medical Center, Gifu, Japan. The authors have no conflicts of interest to disclose. the yield of FB is lower for small lesions ( 20 mm in diameter). To overcome this problem, an ultrathin bronchoscope has been recently developed. 1 4 In Manuscript received March 25, 2006; revision accepted September 12, Reproduction of this article is prohibited without written permission from the American College of Chest Physicians ( org/misc/reprints.shtml). Correspondence to: Koichi Yamazaki, MD, PhD, First Department of Medicine, Hokkaido University School of Medicine, North 15, West 7, Kitaku, Sapporo , Japan; kyamazak@med.hokudai.ac.jp DOI: /chest CHEST / 131 / 2/ FEBRUARY,

2 addition, Asano et al 5 combined an ultrathin bronchoscope with CT-guided transbronchial biopsy (TBB) and navigation by virtual bronchoscopy (VB) for diagnosing small peripheral pulmonary lesions. We have previously applied this procedure for diagnosing 26 small peripheral pulmonary lesions ( 20 mm in diameter) and reported its safety and high diagnostic sensitivity of 65.4%. 6 In the previous series of experiences, we noticed several problems with this procedure. First, small biopsy forceps (diameter, 1.0 mm) occasionally retrieved insufficient tissue specimens. However, repetitive biopsy in which the forceps were manipulated with moderate pressure from a slightly proximal position improved this outcome. Nevertheless, the diagnostic sensitivity was still approximately 65%, not closer to 100%. One primary reason was that lesions inaccessible even by an ultrathin bronchoscope with VB navigation still existed. Accordingly, we attempted to determine the characteristics of the lesions that could not be reached by forceps with an ultrathin bronchoscope. For that purpose, we analyzed the relationship between the diagnostic sensitivity and the location of the lesions, the bronchial generation to which an ultrathin bronchoscope was inserted, and the lesion-bronchial relationship on high-resolution CT (HRCT). Previously, several investigators 7 10 reported the value of the CT-bronchus sign, seen as the presence of a bronchus leading directly to a lesion, as a factor related to a higher diagnostic sensitivity with an FB. Because an ultrathin bronchoscope can be inserted into more peripheral areas where bronchi are not seen on HRCT, we also analyzed the relationship between lesions and pulmonary arteries on HRCT. Here, we evaluated whether the CT-artery sign, seen as the presence of a pulmonary artery leading directly to a lesion, was valuable for a higher diagnostic sensitivity with a CT-guided TBB using an ultrathin bronchoscope. Subjects Methods and Materials Between June 2001 and April 2005 at Hokkaido University Hospital, 83 patients (41 men and 42 women) with 85 small peripheral pulmonary lesions (mean diameter, 20 mm) underwent CT-guided TBB using an ultrathin bronchoscope with VB navigation. In this study, the examined subjects include 26 lesions already reported in our pilot study. 6 On HRCT, the average diameter of the target lesions was 13.6 mm. The institutional ethics committee approved the study. All patients were given detailed descriptions of the examination and informed that this was a new approach. Informed consent was obtained in all cases. VB VB images were reconstructed from CT data as previously described by Onodera et al. 11 All VB images were reconstructed from helical CT scans and transferred to a workstation (Alatoview; Toshiba, Tokyo, Japan; or Virtual Place Advance; AZE; Tokyo, Japan). 6 CT-Guided TBB CT-guided TBB was performed as previously described 6 using an ultrathin bronchoscope (BF-XP40 or BF-XP260F; Olympus; Tokyo, Japan). Referring to VB navigation, the ultrathin bronchoscope was inserted into the target bronchus as deep as possible under direct vision. The position of the forceps inserted through the bronchoscope was then confirmed and adjusted by real-time multislice CT fluoroscopy. Subsequently, biopsy was repeated until adequate specimens were collected. As long as possible, we also performed brushing cytology and bronchial lavage. CT Signs Images of 67 lesions from 66 patients who had undergone HRCT before CT-guided TBB were retrospectively reviewed. We used a multidetector CT scanner for HRCT with 1.0-mmthick section. The HRCT parameters were as follows: 1.0 mm collimation; power, 135 to 149 kv; 100 ma; rotation time, 1.0 s. Images on HRCT were reviewed by two of three experienced pulmonologists to assess the relationship between the lesions and bronchial or arterial trees without any information on the final diagnosis of the lesions. When the decisions of lesion-bronchial and lesion-arterial relationships by two pulmonologists were not compatible, the third pulmonologist determined them. Statistical Analysis All data were processed using standard statistical methods (StatView, version 5.0; SAS Institute; Cary, NC). Results were presented as mean SD. Statistical evaluation was performed using 2 2 frequency tables and Pearson correlation coefficient test. Logistic regression analysis was applied for multivariate analysis of factors related with diagnostic sensitivity; p 0.05 was regarded as significant. Results Fifty-six of the 85 lesions (66%) were diagnosed following CT-guided TBB using an ultrathin bronchoscope with VB navigation. These lesions were found to be 37 cases of primary lung cancer (33 adenocarcinomas, 1 squamous cell carcinoma, 1 large cell carcinoma, and 2 small cell carcinomas), 7 cases of metastatic cancer (3 colon cancers, 1 hepatic cell carcinoma, 1 pancreas cancer, 1 renal cell carcinoma, and 1 thyroid cancer), and 12 cases of benign disease (5 inflammatory changes, 2 sarcoidoses, 3 nontuberculosis mycobacterioses, 1 radiation pneumonia, and 1 nocardiosis). The 29 lesions not diagnosed by CT-guided TBB were subsequently diagnosed by surgery or long-term follow-up. No significant differences were observed between diagnosed and undiagnosed lesions regarding patient age, patient sex, and average diameter of the lesions. Regarding the location of the lesions classified by 550 Original Research

3 Table 1 Diagnostic Sensitivity and Bronchial Generation of VB Image Constructed and of FB Inserted in Each Bronchopulmonary Segment* Segment No. Yield, No. (%) Bronchial Generation of VB Image Constructed Bronchial Generation of FB Inserted RUL (65) S (40) S (78) S (67) RML 7 6 (86) S (75) S (100) RLL (74) S (56) S 7 0 No data No data S (100) S (100) S (67) LUL (71) S (90) S (33) S (75) S (100) LLL 14 5 (36) S (0) S (50) S (33) S (75) Total (66) *Data are presented as mean SD unless otherwise indicated. RUL right upper lobe; RML right middle lobe; RLL right lower lobe; LUL left upper lobe; LLL left lower lobe. anatomic lobes and segments, the lesions in the left lower lobe (36%) had a significantly lower yield, compared to other locations (right upper lobe, 65%; right middle lobe, 86%; right lower lobe, 74%; left upper lobe, 71%; and left lower lobe, 36%; p 0.05) [Table 1]. Of interest, none of the lesions in the left superior segment of the lower lobe (S 6 ) were diagnosed by this procedure. Compared to other lobes, this was significantly low diagnostic sensitivity (0% vs 70%; p 0.01). No significant differences were found in the bronchial generation of VB images constructed following FB insertion between each segment. Next, we analyzed the relationship between diagnostic sensitivity and the bronchial generation to which an ultrathin bronchoscope was inserted. The ultrathin bronchoscope could be inserted between the third and tenth generations of bronchi in 85 lesions. When an ultrathin bronchoscope could be inserted to the fifth bronchial generation, the yield was greater than the average diagnostic sensitivity of all lesions (66%) [Fig 1]. However, even in the lesions in which an ultrathin bronchoscope could be inserted to more than the seventh bronchial generation, the yield was still 66%. This result revealed that the relationship between the diagnostic sensitivity and the bronchial generation to which the instrument could be inserted was not a simple association. We next attempted to correlate diagnostic sensitivity and CT signs on HRCT. The relationship between the lesions and the bronchial or arterial trees were classified into five types according to their radiologic appearance on HRCT: (type 1) bronchus (with or without pulmonary artery) leading to the center of lesion; (type 2) bronchus (with or without pulmonary artery) leading to the edge of the lesion; (type 3) pulmonary artery alone leading to the center Figure 1. Diagnostic sensitivity and bronchial generation to which an ultrathin bronchoscope was inserted. CHEST / 131 / 2/ FEBRUARY,

4 Figure 2. CT signs on HRCT. Top left, type 1: Bronchus (arrow) [with or without pulmonary artery] leading to the center of the lesion (triangle); top center, type 2: Bronchus (arrow) [with or without pulmonary artery] leading to the edge of the lesion (triangle); top right, type 3: Pulmonary artery alone (arrow) leading to the center of the lesion (triangle); bottom left, type 4: Pulmonary artery alone (arrow) leading to the edge of the lesion (triangle); bottom right, type 5: Neither bronchus nor pulmonary artery leading to the lesion (triangle). of the lesion; (type 4) pulmonary artery alone leading to the edge of lesion; (type 5) neither bronchus or artery leading to the lesion (Fig 2). In this classification, types 1 and 2 have been called CT-bronchus signs In comparison with this sign, we named CT-artery sign for type 3 and type 4. Table 2 shows the diagnostic sensitivity for each group of patients by type. Not only patients with CT-bronchus signs but also patients with CT-artery signs revealed high yields by TBB using an ultrathin bronchoscope. However, patients with no CT signs showed significantly lower yield compared to the other patients Table 2 Relationship Between Diagnostic Sensitivity and CT Signs CT Signs Lesions, No. Diagnostic Sensitivity, No. (%) Type 1, bronchus to the center of lesion (79) Type 2, bronchus to the edge of lesion (71) Type 3, artery to the center of lesion 10 8 (80) Type 4, artery to the edge of lesion 9 5 (56) Type 5, neither bronchus nor artery 7 1 (14)* to the lesion Total (67) *p 0.01 compared to other groups. (p 0.01). Multivariate analysis revealed that the location of the lesion was an independent predictor of diagnostic sensitivity by CT-guided TBB using an ultrathin bronchoscope and VB navigation (p 0.05) [Table 3]. Discussion In our institute, we have combined CT-guided TBB, an ultrathin bronchoscope, and VB navigation for diagnosing peripheral pulmonary lesions 20 mm in diameter, following a previous report. 5 In spite of increasing experience with this procedure, the diagnostic sensitivity has still remained approxi- Table 3 Multivariate Analysis of Features Associated With Diagnostic Sensitivity Variables Odds Ratio 95% Confidence Interval p Value Location of lesions* Bronchial generation of FB inserted CT signs *Left lower lobe vs other lobes. Types 1, 2, 3, and 4 vs type Original Research

5 mately 65%. We therefore analyzed the factors related with diagnostic sensitivity. First, the lesions in the left S 6 segment had a significantly lower yield than those in other segments. For insertion of an ultrathin bronchoscope to the peripheral area of S 6, multiple manipulation of the ultrathin bronchoscope in upward and downward directions must be made. It is not easy to manipulate a limp ultrathin bronchoscope to various directions with various angles after it has been bent upward and downward. In addition, an FB must be negotiated with a sharper curve for leading the FB to the left main bronchus. However, the present study shows that it was not difficult to manipulate an ultrathin bronchoscope in the apical segment, in which the yield using a conventional FB was low. 12,13 We have also shown that insertion of a bronchoscope to the fifth bronchial generation was sufficient to diagnose peripheral pulmonary lesions 20 mm in diameter. This result provides important information on constructing VB images for diagnosing small lesions. In the present study, the average bronchial generation of VB images was 7.5, which was sufficient for navigation of an ultrathin bronchoscope. The CT-bronchus sign seen as the presence of a bronchus leading directly to a lesion has been shown to be a valuable factor for the diagnostic sensitivity of peripheral pulmonary lesions However, in the present study, the CT-bronchus sign could not be detected on HRCT in 26 of the 67 lesions. On the other hand, an ultrathin bronchoscope can be inserted into more peripheral areas where bronchi are not seen on HRCT. Therefore, we paid attention to the CT-artery sign, seen as the presence of a pulmonary artery leading directly to the lesion on HRCT. The pulmonary artery and the bronchus are next to each other in the periphery of the lung. Therefore, using a pulmonary artery as a substitute for the bronchus is considered reasonable. However, this appears to be the first report in which the CT-artery sign was useful for successful TBB using an ultrathin bronchoscope. Therefore, in cases of TBB using an ultrathin bronchoscope, more precise determination of bronchi as well as pulmonary arteries on HRCT is definitely required. This result also supports the strategy of VB construction using the pulmonary artery in place of a bronchus. 11 It has been shown that the lesions with bronchi leading to their edges have a lower yield because the bronchi are compressed by the nodules. 7 However, in the present study, those lesions were diagnosed as having a yield similar to that of lesions with bronchi leading to their centers. This difference appears to have two possible explanations. First, after an ultrathin bronchoscope reaches the edge of the lesions, it can be easily directed to the center of the lesions by bending an ultra-thin bronchoscope to obtain the tissue from the lesions. However, in the case of forceps with a conventional FB, inducing the apparatus to the center of the lesions by bending the FB is difficult. Second, the bronchial wall is so thin in the periphery of the lung that repetitive biopsy could yield the tissue over the bronchial wall. We evaluated three factors related to the yield of small peripheral pulmonary lesions by CT-guided TBB using an ultrathin bronchoscope and VB navigation. However, more complicated factors make it impossible for an ultra-thin bronchoscope to reach the lesions, such as an extremely tortuous bronchial route to the lesion. To foresee such difficulty in bronchoscopy, simulation using VB imaging is thought to be useful. References 1 Rooney CP, Wolf K, McLennan G. Ultrathin bronchoscopy as an adjunct to standard bronchoscopy in the diagnosis of peripheral lung lesions. Respiration 2002; 69: Saka H, Oki M, Kumazawa A, et al. Diagnosis of pulmonary peripheral lesions using an ultrathin bronchoscope. J Jpn Soc Bronchol 2000; 22: Yamamoto S, Ueno K, Imamura F, et al. Usefulness of ultrathin bronchoscopy in diagnosis of lung cancer. Lung Cancer 2004; 46: Oki M, Saka H, Kitagawa C, et al. Visceral pleural perforation in two cases of ultrathin bronchoscopy. Chest 2005; 127: Asano F, Matsuno Y, Matsushita T, et al. Transbronchial diagnosis of a pulmonary peripheral small lesion using an ultrathin bronchoscope with virtual bronchoscopic navigation. J Bronchol 2002; 9: Shinagawa N, Yamazaki K, Onodera Y, et al. CT-guided transbronchial biopsy using an ultrathin bronchoscope with virtual bronchoscopic navigation. Chest 2004; 125: Gaeta M, Barone M, Russi EG, et al. Carcinomatous solitary pulmonary nodules: evaluation of the tumor-bronchi relationship with thin-section CT. Radiology 1993; 187: Naidich DP, Sussman R, Kutcher WL, et al. Solitary pulmonary nodules CT-bronchoscopic correlation. Chest 1988; 93: Gaeta M, Pandolfo I, Volta S, et al. Bronchus sign on CT in peripheral carcinoma of the lung: value in predicting results of transbronchial biopsy. AJR Am J Roentgenol 1991; 157: Bilaceroglu S, Kumcuoglu Z, Alper H, et al. CT bronchus sign-guided bronchoscopic multiple diagnostic procedures in carcinomatous solitary pulmonary nodules and masses. Respiration 1998; 65: Onodera Y, Omatsu T, Takeuchi S, et al. Enhanced virtual bronchoscopy using the pulmonary artery. AJR Am J Roentgenol 2004; 183: Chechani V. Bronchoscopic diagnosis of solitary pulmonary nodules and lung masses in the absence of endobronchial abnormality. Chest 1996; 109: Bandoh S, Fujita J, Tojo Y, et al. Diagnostic accuracy and safety of flexible bronchoscopy with multiplaner reconstruction images and ultrafast Papanicolaou stain. Chest 2003; 124: CHEST / 131 / 2/ FEBRUARY,