Oxybuprocaine

Background: Electromagnetic navigation in bronchoscopy is a novel method for assisting in the localization of peripheral lung lesions.

Study objective: To assess the usability, accuracy, and safety of electromagnetic navigation during flexible bronchoscopy in a clinical setting.

Design: Prospective evaluation.

Patients: Consecutive patients referred to the bronchoscopy unit for the diagnosis of peripheral infiltrates or solitary pulmonary nodules (SPNs).

Methods: Navigation was performed using an electromagnetic tracking system with a position sensor encapsulated in the tip of a flexible catheter that was pushed through the working channel of the bronchoscope. Real-time, multiplanar reconstruction of a previously acquired CT data set provided three-dimensional views for localization of the catheter. To match the position of the sensor with the CT scan, four anatomic landmarks were used for registration. The sensor position generated in the navigation system was controlled by fluoroscopy, and the corresponding error distances were measured. This was performed with all SPNs and at two different peripheral locations of the right upper lobe (RUL).

Results: Sixteen patients (10 men and 6 women; mean age, 63.7 years) were studied. Navigation prolonged bronchoscopy by 3.9 [+ or -] 1.3 min (mean [+ or -] SD). The navigation system identified all lesions. The position sensor achieved a direct hit in three of five SPNs. Fluoroscopy failed to recognize three SPNs (60%) and three infiltrates (38%). The mean error distances between sensor tip position and fluoroscopically verified RUL reference position were 10.4 mm (lateral position) and 12.5 mm (apical position) respectively. The mean error distances between the sensor tip and two endobronchial registration points at the end of the procedure were 4.2 mm and 5.1 mm, respectively.

Conclusion: Electromagnetic navigation is useful, accurate, and safe in the localization of peripheral lung lesions and may help to improve the yield of diagnostic bronchoscopic procedures.

Key words: bronchoseopy; CT; diagnostic imaging; image-guided; stereotaxy; three-dimensional imaging; virtual bronchoscopy

Abbreviations: RUL = right upper lobe; SCU = system control unit; SPN = solitary pulmonary nodule

**********

Image-directed bronchoscopy plays a major role in evaluating peripheral localized pulmonary lesions since they cannot be directly visualized. Histopathologic, cytologic, or microbiological assessment is essential in order to initiate adequate therapy. The diagnostic yield of studies addressing the success of flexible bronchoscopy in solitary pulmonary nodules (SPNs) is 18 to 62%. (1-3) Lesion size in particular is a major determinant of diagnostic accuracy. (4-7) The diagnostic yield for lesions [less than or equal to] 2 cm does not exceed 30%. (4,7-9) In fact, the yield for a lesion of [less than or equal to] 2 cm located in the outer third of the lung is only 14%. (3) Normally, fluoroscopy is used during bronchoscopy in order to identify lesions and to guide diagnostic instruments. The difficulty is that it is limited by a two-dimensional display format, producing overlapping structures and potentially hindering biopsy. Improved diagnostic accuracy has been suggested using a CT-guided bronchoscopic approach. (10-12) However, drawbacks to this technique such as costs and radiation exposure may restrict efficient application. (13)

A novel method for guiding transbronchial catheters or forceps is electromagnetic navigation. (14) In comparison to fluoroscopy or CT scanning, electromagnetic navigation as a method not only has minimum technical and spatial requirements, it also indicates the position of the catheter in three dimensions without radiation exposure; all that it needs is the availability of a preprocedure CT data set. Moreover, the feasibility and reproducibility of this method have already been demonstrated. (15,16) The purpose of this study was to assess the feasibility and safety of electromagnetic navigation, with fluoroscopy validation of the position of the electromagnetic sensor generated by the tracking system in a clinical setting.

MATERIALS AND METHODS

Patients

Consecutive patients with peripheral lung infiltrates or SPNs undergoing diagnostic bronchoscopy were prospectively evaluated using the electromagnetic navigation system. All patients gave written informed consent. Selection criteria were an SPN, a pulmonary mass, or a localized infiltrate necessitating the use of fluoroscopy to guide biopsy forceps or transbronchial needle aspiration (Table 1). An additional requirement was the availability of a CT data set, which had to be obtained within 7 days prior to bronchoscopy.

Bronchoscopy

Flexible bronchoscopy (BF-1T40; Olympus; Tokyo, Japan) was carried out with topical anesthesia and IV sedation via the nasal route. The patient was placed in a supine position. Topical anesthesia was achieved by administering 10 to 20 mL of 2% oxybuprocaine via nebulizer to the spontaneously breathing patient. Premedication consisted of IV midazolam, 2 to 5 mg. If appropriate, propofol was administered IV in a single dose of 20 to 50 mg. Patient monitoring consisted of continuous finger pulse oximetry and ECG. BP was measured every 5 min.

CT Scans

Preoperative spiral CT scans (Somatom Sensation 16; Siemens; Munich, Germany) with IV contrast were performed in our hospital using a standardized protocol (120 kilovolt peak; 120 mA; pitch, 1.3; collimation, 0.75 mm; and 5-mm reconstruction). The patients kept their arms in standard position over the head. A standardized digital imaging and communications in medicine (DICOM; NEMA; Roslyn, VA) transfer transmitted data to the navigation computer via the local area network.

Navigation System

For navigation within the bronchial tree, a flexible electromagnetic tracking system was used (Aurora; Northern Digital; Waterloo, ON, Canada). This system is based on a magnetic field generator with dimensions of 220 x 220 x 170 mm; a minimized receiver sensor (diameter, 0.8 mm; length, 10 mm), which is encapsulated in the tip of a flexible catheter (diameter, 1.5 mm); a small sensor interface unit (60 x 60 x 25 mm) for data conversion from analog to digital; and the system control unit (SCU) [Fig 1].

Inside the field generator (transmitter), a pulsed, direct-current electromagnetic field is generated with a symmetric setup of nine coils. Within the resulting mapping region measuring 500 x 500 x 500 mm, the sensor position can be detected in five degrees of freedom (X, Y, Z-axes, pitch, yaw). The SCU contains a processor board to calculate the position and rotation of the sensor from the incoming signals and sends it to the host computer. The navigation software (Syngo; Siemens Medical Solutions; Erlangen, Germany) runs on a standard personal computer with a standard graphics controller. The transmitter was located on the left side of the examination table with its height adjusted to the level of the patient's thorax (Fig 2). After loading the CT data set into the registration application, the software performed a multiplanar reconstruction providing axial, sagittal, and coronal views. At each of these views, the sensor tip was displayed as a crosshair, showing its actual position. The catheter with the sensor was pushed through the working channel of the bronchoscope and then advanced to the lung periphery.

[FIGURE 2 OMITTED]

Image Registration Process

To match the position of the sensor with the previously acquired CT scan, four anatomic landmarks were used. The sensor tip was placed on the particular anatomic point, and the corresponding CT position was then marked on the computer. Landmarks were the proximal rim of the manubrium sterni (landmark 1), the distal end of the xiphoid (landmark 2), and two endobronchial anatomic points (main carina [landmark 3] and right upper lobe [RUL] carina [landmark 4]) that were obtained bronchoscopically. According to preclinical investigations, three anatomic landmarks are mathematically sufficient, but using four landmarks improves accuracy.

Navigation and Testing for Accuracy

The sensor was introduced and advanced to the bronchial orifice most likely to lead to the lesion as determined from a review of the preliminary CT scan. When the sensor missed the lesion, it was repositioned until the tip of the sensor reached or penetrated the lesion on the computer display of the navigation system (Fig 3). In SPNs, the position of the sensor was controlled by fluoroscopy in a posteroanterior view. Then the distance from the margin of the nodule to the tip of the sensor was measured. All lesions were sampled for biopsy.

[FIGURE 3 OMITTED]

In a second step, the sensor was inserted into the RUL and, with the aid of the tracking system, was then forwarded until it came to a stop at the most lateral position. The identical procedure was performed with the most apical position of the upper lobe (segment 1). These positions were each considered to be the visceral pleura, representing the inner lining of the thoracic wall. Fluoroscopy then verified whether the sensor tip was actually adjacent to the inner lining of the thoracic wall (Fig 4). On the computer display, the distance between the crosshair, representing the sensor tip, and the visceral pleura was measured and defined as error distance. Subsequently, the sensor tip was again positioned on landmark 3 and landmark 4 (main carina and RUL carina) to determine whether a registration error had developed during the course of the procedure, ie, by body movements.

[FIGURE 4 OMITTED]

RESULTS

Sixteen patients (10 men and 6 women; mean age, 63.7 years; range, 42 to 84 years) undergoing diagnostic flexible bronchoscopy with fluoroscopic guidance were included in the study. On the computer screen of the navigation system, sensor tip and lesions were clearly identified in all cases. No complications occurred during bronchoscopy. The image registration process lasted 4.1 [+ or -] 1.9 min (mean [+ or -] SD). Use of the navigation catheter including fluoroscopic control prolonged bronchoscopy by 3.9 [+ or -] 1.3 min.

In all pulmonary infiltrates, the navigation system was able to guide the sensor tip to the center of the lesion. In three of these eight patients, the infiltrates were not detected by fluoroscopy. In the five patients with SPNs, the sensor tip could be directed tangent to the nodule in three cases. In the remaining two cases, the sensor tip only approximated the nodule. The average diameter of the nodules was 22 [+ or -] 6 mm, with a maximum distance to the pleura of 35 mm. Three of the nodules were not identified by fluoroscopy. In the other two nodules, the position of the sensor tip in relation to the nodule was controlled by fluoroscopy. Both of these nodules were reached by the sensor tip when guided by the navigation system. Masses were hit in all three cases. The resulting error distances are displayed in Table 1. When the sensor tip was forwarded to the lateral and apical thoracic wall of the RUL, fluoroscopy verified the pleural position of the sensor tip in all cases. Error distances between sensor tip and visceral pleura as measured on the display of the tracking system ranged from 0 to 23 mm (Table 1). Respiratory position variation between inspiration and expiration during normal tidal volume had no measurable impact on the error distances. The mean error distances between the sensor tip and landmark 3 and landmark 4 at the end of the procedure were 4.19 mm and 5.13 mm (Table 1).

The biopsy results in three of five SPNs were positive for carcinoma. The remaining two cases revealed normal lung tissue. All masses were positive for carcinoma. Biopsy results of infiltrations were found to be positive in five cases (granulomatous disease, invasive aspergillosis, pneumonitis associated with radiation, interstitial lung disease, bronchioloalveolar cell carcinoma) In the remaining three cases, histology findings were unspecific.

DISCUSSION

In this prospective case series of 16 patients, electromagnetic navigation for the localization of intrapulmonary lesions provided a number of advantages over conventional fluoroscopy-guided bronchoscopy. Particularly in lesions not visible under fluoroscopy, electromagnetic navigation is capable of identifying nodules and infiltrates and of guiding a catheter to the target in order to obtain diagnostic material. This is specifically relevant for small lesions since the sensitivity of bronchoscopy in small peripheral nodules is comparatively low. (3) Although competing techniques such as CT or endobronchial ultrasound also potentially increase the yield of endobronchial sampling in peripheral lesions, there are other concerns with these methods. The use of CT-guided bronchoscopy is associated with radiation exposure, (14) and endobronchial ultrasound may show limitations when investigating lung parenchyma since this can be hampered by total reflection of the ultrasound signal due to air.

The accuracy of the navigation sensor position when controlled by fluoroscopy is similar to that of two experimental studies in pigs, (14-16) even though methods are not directly comparable. The average error distance measured when the sensor tip reached the lung periphery was < 13 mm. It is difficult to assess how the size of error distances correlates with biopsy results. This has not been evaluated yet. The high proportion of positive histology results suggests that electromagnetic navigation provides a level of accuracy that makes it possible to improve the diagnostic yield of transbronchial biopsies. Due to technical and procedural limitations (patient movement, lung expansion, CT acquisition, and bronchoscopy at different points in time), certain error distances are unavoidable. This may limit the capabilities of the method in very small lesions, but is less important in larger lesions that are still invisible by fluoroscopy. We want to emphasize that the sample size of the current study is too small and the lesions are too inhomogeneous to draw definitive conclusions regarding the diagnostic yield of the method. In addition, there is no control group that would include patients with similar lesions who had undergone biopsy without electromagnetic navigation.

We did not use a secondary reference sensor to maintain the patient's position in case of body movements and cough. However, for the most part this was compensated by sedation and fixation of the patient. The use of a position sensor is likely to improve accuracy even further in future. The position of the patient's arms at the acquisition of the CT scan was different from the position during bronchoscopy. While during the CT scan arms were above the head, they were kept on the body during bronchoscopy. This was intended, since it reflects normal routine in daily clinical practice.

Precision measurements obtained at the end of the procedure at two anatomic registration points again revealed only small errors which are comparable with other investigations. (14,16) One limitation may be the lack of control of lung expansion. CT scans are normally obtained with an inspiratory breath hold. During bronchoscopy, we see tidal breathing. It is questionable whether this has a significant effect on navigation. We observed that upper lobe landmarks and the bronchial landmarks (eg, main carina, upper lobe carina) are not much affected by breathing artifacts, since they do not change their position during inspiration. Lower lobe lesions, however, may change their position during deep inspiration. This may have an impact on the accuracy in very small lesions, but is probably less important in larger lesions still invisible by fluoroscopy. Future developments will have to include steerable biopsy forceps with an integrated navigation sensor and algorithms to compensate for drawbacks induced by lung expansion.

Although we used monoplanar fluoroscopy to measure error distances, which has the potential of incorrect distance assessment in parenchymal abnormalities, we assume that this is not of significant relevance. In our view, this does not account for the most lateral and most apical positions within the upper lobes, since these positions are clearly identifiable also by monoplanar fluoroscopy. In lesions within the lung tissue (eg, SPNs), the image plane may play a role in the calculation of error distances. However, when the navigation system indicates contact of the sensor with the most peripheral spot of the lesion, we assume that sensor and lesion are within the same plane.

Overall, electromagnetic navigation was well tolerated, and proved to be both safe and useful in localizing small or fluoroscopically invisible lung lesions with a sufficient level of accuracy. An important advancement is the three-dimensional display of the sensor tip without radiation exposure. Despite some procedural limitations, there is good evidence that this method has the potential for considerable improvements in the diagnostic possibilities of bronchoscopic procedures. Studies with clinical end points will now have to prove whether this is true for the routine use of electromagnetic navigation.

* From Medizinische Klinik I (Drs. Hautmann, Pinkau, and Peltz), and Workgroup MITI (Mr. Schneider and Dr. Feussner), Klinikum rechts der Isar, Technische Universitat Munich, Germany.

REFERENCES

(1) Torrington KG, Kern JD. The utility of fiberoptic bronchoscopy in the evaluation of the solitary pulmonary nodule. Chest 1993; 104:1021-1024

(2) Fletcher EC, Levin DC. Flexible fiberoptic bronchoscopy and fluoroscopically guided transbronchial biopsy in the management of solitary pulmonary nodules. West J Med 1982; 136:477-483

(3) Baaklini WA, Reinoso MA, Gorin AB, et al. Diagnostic yield of fiberoptic bronchoscopy in evaluating solitary pulmonary nodules. Chest 2000; 117:1049-1054

(4) Wallace JM, Deutsch AL. Flexible fiberoptic bronchoscopy and percutaneous needle lung aspiration for evaluating the solitary pulmonary nodule. Chest 1982; 81:665-671

(5) Ellis JH Jr. Transbronchial lung biopsy via the fiberoptic bronchoscope: experience with 107 consecutive cases and comparison with bronchial brushing. Chest 1975; 68:524-532

(6) Stringfield JT, Markowitz DJ, Bentz RR, et al. The effect of tumor size and location on diagnosis by fiberoptic bronchoscopy. Chest 1977; 72:474-476

(7) Radke JR, Conway WA, Eyler WR, et al. Diagnostic accuracy in peripheral lung lesions: factors predicting success with flexible fiberoptic bronchoscopy. Chest 1979; 76:176-179

(8) Cortese DA, McDougall JC. Biopsy and brushing of peripheral lung cancer with fluoroscopic guidance. Chest 1979; 75:141-145

(9) Shure D, Fedullo PF. Transbronchial needle aspiration of peripheral masses. Am Rev Respir Dis 1983; 128:1090-1092

(10) White CS, Weiner EA, Patel P, et al. Transbronchial needle aspiration: guidance with CT fluoroscopy. Chest 2000; 118: 1630-1638

(11) Goldberg SN, Raptopoulos V, Boiselle PM, et al. Mediastinal lymphadenopathy: diagnostic yield of transbronchial mediastinal lymph node biopsy with CT fluoroscopic guidance-initial experience. Radiology 2000; 216:764-767

(12) Garpestad E, Goldberg S, Herth F, et al. CT fluoroscopy guidance for transbronchial needle aspiration: an experience in 35 patients. Chest 2001; 119:329-332

(13) Wagner U, Walthers EM, Gelmetti W, et al. Computer-tomographically guided fiberbronchoscopic transbronchial biopsy of small pulmonary lesions: a feasibility study. Respiration 1996; 63:181-186

(14) Solomon SB, White P Jr, Acker DE, et al. Real-time bronchoscope tip localization enables three-dimensional CT image guidance for transbronchial needle aspiration in swine. Chest 1998; 114:1405-1410

(15) Solomon SB, White P Jr, Wiener CM, et al. Three-dimensional CT-guided bronchoscopy with a real-time electromagnetic position sensor: a comparison of two image registration methods. Chest 2000; 118:1783-1787

(16) Schwarz Y, Mehta AC, Ernst A, et al. Electromagnetic navigation during flexible bronchoscopy. Respiration 2003; 70:516-522

Manuscript received May 24, 2004; revision accepted January 5, 2005.

Reproduction of this article is prohibited without written permission from the American College of Chest Physicians (www.chestjournal. org/misc/reprints.shtml).

Correspondence to: Hubert Hautmann, MD, Pneumologie, Klinikum rechts der Isar, Ismaninger Str.22, D--81675 Munchen, Germany; e-mail: hautmann@web.de

COPYRIGHT 2005 American College of Chest Physicians
COPYRIGHT 2005 Gale Group