Multi-Dimensional Display of Wang’s Lymph Node Map Using Virtual Bronchoscopic Navigation System

Background: Transbronchial needle aspiration (TBNA) is a classical technique for diagnosing mediastinal-hilar lymph node enlargement. However, the diagnostic value of conventional TBNA (cTBNA) is limited in small lymph nodes. Methods: Here, we generated an innovative multi-dimensional virtual lymph node map on top of Wang’s lymph node map using a Lungpoint Virtual Bronchoscopic Navigation System. Results: The virtual bronchoscopic navigation (VBN) system was combined with computed tomography (CT) images to generate extrabronchial, endobronchial, sagittal, coronal as well as horizontal views of the 11 intrathoracic lymph node stations and their adjacent tissues and blood vessels. We displayed the specific puncture site of each lymph node station. The 11 stations were divided into four groups: right mediastinal stations, left mediastinal stations, central mediastinal stations and hilar stations. Conclusion: The VBN system provides a precise view of the intrabronchial landmarks, which may increase the diagnostic accuracy of intrathoracic lymph node adenopathy and assist bronchoscopists with practicing TBNA.


INTRODUCTION
Transbronchial needle aspiration (TBNA) is a classical bronchoscopic technique for diagnosing benign and malignant mediastinal-hilar lymph node enlargement. With technological progress, endobronchial ultrasound-guided TBNA (EBUS-TBNA) provides real-time ultrasound imaging, resulting in a higher diagnostic yield than conventional TBNA (cTBNA) and has been widely practiced in clinical settings worldwide (Herth et al., 2004;Bonifazi et al., 2017;Madan et al., 2017). However, even with ultrasound guidance, the diagnostic yield of TBNA varies, largely due to the skill and experience of the operator (Rodriguez de Castro et al., 1997;Hsu et al., 2004). Therefore, whether the operator can accurately locate the target lymph nodes is of great importance, and requires a comprehensive understanding of the anatomy of mediastinal-hilar lymph nodes.
The International Association for the Study of Lung Cancer (IASLC) lymph node map and Wang's lymph node map are the two most commonly used lymph node maps (Wang, 1994;Rusch et al., 2009). The eighth edition of the IASLC map was introduced in 2009 and most of the lymph node stations are defined according to the positions of blood vessels (El-Sherief et al., 2014). Without ultrasound, it is challenging to identify the blood vessels under bronchoscopy. Wang's map was proposed in 1994 and the targeted lymph nodes are located by airway landmarks under endoscopic view, which is more practical for practitioners (Wang, 1994).
Electromagnetic navigation bronchoscopy (ENB) was first applied in human subjects in 2006 (Schwarz et al., 2006). Compared with traditional bronchoscopy, ENB improved the diagnostic accuracy for peripheral pulmonary lesions (PPLs) and mediastinal lymph nodes (Gildea et al., 2006). However, the EBN process requires expensive equipment and special training for operators, and therefore it can only be applied in qualified medical centers. On the other hand, virtual bronchoscopy can simulate a three-dimensional view of the bronchi and adjacent structures using. The reported diagnostic yield of virtual bronchoscopic navigation (VBN) system for peripheral pulmonary lesions was 70.9%-76.8% for peripheral pulmonary lesions, that of thin bronchoscopy combined with computed tomography (CT) and VBN was 65.4%-81.6% and that of X-ray fluoroscopy plus VBN was 62.5%-78.7% (Asano et al., 2014). Electromagnetic navigation and virtual navigation could also be used as advanced lymph node localisation systems. Previously, we reported the intrabronchial display of hilarmediastinal lymph nodes by the SPiN Thoracic Navigation System (Wu et al., 2018). However, only an endobronchial map of mediastinal-hilar lymph nodes was generated, without demonstration of the extrabronchial anatomical position of each lymph node. Here, we further devised a virtual extrabronchial and endobronchial map together with horizontal, coronal and sagittal views on chest CT for each mediastinal and hilar lymph node using the Lungpoint Virtual Bronchoscopic Navigation System, in accordance with the recommended lymph node puncture site in Wang's map. Therefore, clinicians can gain an overall impression of the three-dimensional (3D) spatial structure of the most frequently punctured lymph nodes, thus improving the accuracy and safety of TBNA.

METHODS
CT images 1 mm thick were acquired and transferred to a workstation where the Lungpoint Virtual Bronchoscopic Navigation System (Broncus Technologies, Inc., Mountain View, CA, United States) automatically generated a 3D model of the airway. We set up 3D markers of the locations of intrathoracic lymph nodes. The extrabronchial anatomical positions of each target lymph node were shown as green spheres. The positions and borders of 11 target lymph nodes were described according to Wang's lymph node map.

RESULTS
We generated an overall view of 11 lymph node stations in Wang' map under 3D reconstruction ( Figure 1). Next, we collected the horizontal, coronal and sagittal views of chest CT, labeled extrabronchial 3D anatomical locations of the 11 nodal stations in terms of Wang's lymph node map and generated a virtual view from an extrabronchial perspective. We classified the 11 stations into four groups: right mediastinal stations (Figure 2 Figure 5). The recommended TBNA puncture sites of particular lymph nodes were also described in detail (Wang, 1994;Wang, 1995;Xia and Wang, 2013) (see Supplementary Table S1).

DISCUSSION
In the present study, we introduced a virtual map of lymph node landmarks under the assistance of a VBN system. Our map displays the intrabronchial and extrabronchial spatial relationships of 11 lymph node stations and their recommended puncture sites corresponding to Wang's map.
The VBN system in combination with CT of the 11 mediastinal and hilar lymph nodes and their surrounding structures from extrabronchial and endobronchial, sagittal, coronal as well as horizontal views provided a full understanding of the thoracic lymph nodes and their intrabronchial and extrabronchial relationships with adjacent tissues and blood vessels. A detailed view of the lymph nodes and their adjacent tissues can reinforce the understanding of the anatomy and assist in training of operators for TBNA.
There are two major differences between Wang's map and the IASLC map (Xia et al., 2015). First, stations five and six in Wang's map are located in front of the proximal proportion of the right and left main bronchi, respectively. Both W5 and W6 stations are defined as mediastinal lymph nodes. Second, W5 covers two IASLC lymph nodal stations: the distal end of   distinct from the trachea and central bronchus. The development of image-guided bronchoscopy techniques significantly enhances the diagnostic accuracy of PPLs. The VBN system was designed to reach and guide aspiration of PPLs with a diagnostic value of 72.0%-73.8% (Wang Memoli et al., 2012;Asano et al., 2014), and the diagnostic value reached 67.4% for lesions <2 cm in diameter (Asano et al., 2014). In combination with CT-guided ultrathin bronchoscopy, the diagnostic value of VBN increases to 65.4%-81.6% (Asano et al., 2014). On the other hand, the VBN system has also been applied in diagnosis of thoracic lymph node adenopathy. The sensitivity and diagnostic accuracy are both significantly increased for mediastinal lymph nodes by VBN (McAdams et al., 1998;Fiorelli et al., 2017), suggesting that the VBN system is also valuable in cases of mediastinal lymph node enlargement.
Aside from the advantages of the VBN system, the bronchoscopist still has to perform cTBNA blindly in clinical practice, as the exact puncture site cannot be confirmed in real-time. In general, the main purpose of generating our multi-dimensional virtual Wang's lymph node map is for training junior interventional pulmonology fellows or bronchoscopists. Compared with EBUS-TBNA, our approach does not require the use of a regular scope for the airway survey. In addition, although EBUS-TBNA allows real-time visualisation of a punctured lymph node, the operator cannot precisely identify the puncture site. In contrast, our system probes the 3D anatomy of each lymph node station with the puncture site. We expect that our multi-dimensional virtual Wang's lymph node map would allow precise targeting of lymph nodes 1-2 cm in diameter. Hence, the two methods are not competitive but complementary for bronchoscopists within the learning curve. The diagnostic yield is satisfactory upon combining virtual endobronchial ultrasound with EBUS-TBNA bronchoscopy when diagnosing mediastinal-hilar lymph node enlargement (Sato et al., 2013), but its necessity is doubtful. In addition, rapid on-site evaluation (ROSE) may be a good alternative to EBUS for combination with the VBN system. In addition, we recommend a probe designed to determine whether the needle is aligned in the puncture site. When it is aligned properly, the inducer could turn from green to red to alert the practitioner of puncture. Such a device may further improve the diagnostic value of VBN-guided TBNA.
There was a significant limitation of this study in that the diagnostic yield of this particular VBN-assisted TBNA has not yet been verified. A clinical trial is required to examine the validity of this VBN plus TBNA system. In addition, the quality of the CT image can also affect the accuracy of the derived 3D images.
Overall, we have proposed a virtual map for intra-and extrabronchial landmarks of hilar and mediastinal lymph nodes based on the Lungpoint VBN system. This map can help in training bronchoscopists as well as enhancing the

DATA AVAILABILITY STATEMENT
The original contributions presented in the study are included in the article/Supplementary Material, further inquiries can be directed to the corresponding authors.

AUTHOR CONTRIBUTIONS
KW, QLiu, HoS, and YX contributed to study conception and design. FL, YY, HuS, QW, XY, QLi, LC, and YX conducted data collection and data analysis. QLiu, HoS, HuS, KW, YY, FL, and YX drafted the manuscript. All authors contributed to revision of the manuscript, and confirmed the final approval of manuscript.