This study is retrospective, single-centre, and observational in nature. The study proposal was submitted to and approved by the Research Ethics Committee of the Hong Kong Sanatorium and Hospital Group, and the Human Subjects Ethics Application Review System of the Hong Kong Polytechnic University. Patients with lung tumours and referred for CK treatment with fiducial tracking and those whose FMs will be implanted using VBN were included in this study. Patients who already had FMs in their lungs from previous CK treatment and those whose FMs will be used again for current CK treatment were excluded from this study. A total of 32 patients with lung tumours were referred for FM implantation using VBN before CK treatment had been selected. They were divided into two groups according to the methods of FM implantation. The first group of 18 patients were those who underwent VBN FM implantation and were treated with CK from June 2017 to August 2019. The second group of 14 patients were those who underwent additional pre-procedure planning before the VBN FM implantation from March 2019 to July 2020.

Pre-procedure planning involves two steps. The first step is to predefine 5–6 optimal FM positions around the tumour on the patient’s CT images in Eclipse (Eclipse, Varian Medical Systems, Inc., Palo Alto, CA, USA) radiotherapy treatment planning system (TPS). The second step is to import the CT image set together with the predefined FM positions into a VBN procedure planning system (LungPoint Bronchus Medical, Inc., San Jose, CA, USA) for verification and simulation. FMs were then deployed through VBN implantation like previously, but this time with the guidance of a procedure planning system. For the second group of 14 patients who underwent pre-procedure planning, a set of low-dose CTs was performed and imported into the Eclipse RT planning system for an oncologist to delineate the preliminary target. After that, 5–6 predefined optimal FM positions were also contoured around the preliminary target. The procedures and guidelines of the FM contouring were as follows:

1. Delineate each of the FMs with a 0.5-cm-diameter sphere around the tumour and then label them as FM1 to FM5/6.

2. Keep the distance between each FM and the tumour between 1 and 3 cm. This can be performed by creating a pseudo-structure by expanding the tumour to 1 to 3 cm, depending on the size of the tumour and contouring the FMs on the circumference of the pseudo-structure. Although the co-relationship between the FMs and the tumour is better if the distance between them is shorter, the distance between any two individual FMs may be too short or less than the 18 mm, which is the minimum inter-FM distance required for CK tracking. Generally, for a tumour volume of 1 cc or less, a pseudo-structure should be created by expanding the tumour by 3 cm. For a tumour volume ranging from 1 to 3 cc, the pseudo-structure should be created by expanding the tumour by 2 cm. For a tumour volume of 3 cc or more, the pseudo-structure should be created by expanding the tumour by 1 cm.

3. Delineate the FMs as distal as possible in the small bronchi to increase the chance of fixation.

4. For tumours located in the middle of the lung with at least 2 cm of circumferential lung tissues, delineate 6 FMs and arrange the FMs, as follows ( Figure 1 ):

5. For peripheral lung tumours with at least one side of the tumour not possible for FM implant, delineate 5 FMs and arrange the FMs, as follows ( Figure 2 ):

6. Create a set-up plan with two 450° and 3150° oblique fields to simulate the two X-ray imaging views in CK, and review the FM configuration to obtain the following ( Figure 3 ):

7. After the optimal FM positions in the Eclipse RT planning system are defined, import the low-dose CT again into another system, the LungPoint VBN procedure planning system (2018 version, Bronchus Medical, Inc., San Jose, CA, USA).

8. In the LungPoint system, generate the patient’s bronchial tree automatically with the procedure planning system.

9. Copy all the delineated FMs in the Eclipse RT planning system to the LungPoint procedure planning system manually.

10. Review the FMs’ virtual positions to see if they are at or near the end of the most proximal small bronchi ( Figure 4 ).

11. Press the play button. The system will then simulate the VBN process and guide the virtual bronchoscope to the target positions.

12. Assess and confirm that each of the pre-planned FM positions is reachable at the end of the simulation ( Figure 5 ).

All FMs, measured 0.8 × 5 mm (PointCoil MTCTXPC08) ( Figure 6 ), were then implanted by a respiratory medicine specialist using VBN under the active guidance of the LungPoint procedure planning system.

For the comparison of the FM implantation duration, the total time used for FM implantation with VBN for both groups was recorded. For the comparison of X-ray exposure to patients and staff, the dose area product (DAP) in dGy*cm², a quantity used in assessing the radiation risk from diagnostic X-ray examinations, was collected from the internal chamber of the manufacturer (Axiom Artis, Siemens, Munich, Germany) for each exposure for both groups. Furthermore, the total number of FMs implanted during the VBN session was collected and compared for each patient in both groups.

Data analysis was performed using Data Analyses, an add-in statistical tool of Microsoft Excel (Version 16.50, Microsoft 365 Subscription). A parametric, independent t-test was used to compare the means of the two groups of the patients. All p-values were two-sided. A p-value of less than 0.05 was statistically significant.

FMs can be implanted into the lung using two methods, transthoracic/percutaneous and transbronchial through the patient’s respiratory tract. Research showed that percutaneous or CT-guided transthoracic insertion of FMs is associated with a high risk of pneumothorax (16, 26–28). In contrast, transbronchial FM implantation is much safer. A review of 5 studies using the transbronchial FM implantation method showed that 4 of them recorded 0% implantation-induced pneumothorax (14, 15, 29, 30), and the remaining one recorded 2.3% (31). Other recent studies showed that the rate of complication is usually associated with guided transbronchial biopsies (TBBs) performed before FM implantation (32, 33). In this study, there was neither bleeding nor pneumothorax in both groups of our patients after FM implantation with or without pre-procedure planning.

Previous FM insertions were complicated due to the absence of pre-procedure planning. The procedure is dependent on the experience and real-time decision making of a group of different health professionals, including an anaesthetist handling the patients’ condition, a respiratory specialist manipulating the bronchoscope and navigating inside patients’ bodies to look for a good position for the FM, a medical physicist, and a radiation therapist from the CK team staying in the control room, giving advice and confirming the location for FM deployment. With pre-procedure planning, we can decide where and how many FMs can be inserted and what will be the overall resulting configuration. We will also be aware of the possibility that one or more directions could not be feasible for FM deployment or that there are no small bronchi for FMs to be firmly anchored to. Thus, the FM insertion procedure becomes simpler.

The results of this study show that with pre-procedure planning, the average time of FM insertion can be reduced from 41.1 to 22.9 min or by 44% (median = 22.5 min). Furthermore, the overall procedure duration is more predictable, where 9 out of 14 patients’ procedure times are within 20 to 25 min and the SD is reduced from 8.7 to 4.7 min. Improving efficiency brings several advantages. First, the time of patients under anaesthesia can be reduced. It is expected that the risk due to anaesthesia will also be reduced, and hopefully, this procedure can be tolerable by more patients. Second, utilization of the endoscopy room can be increased. With a shorter and more predictable duration, resource allocation will be more effective, and more patients can be arranged for endoscopy procedures. Furthermore, the cost of the procedure is expected to be reduced. The endoscopy room is one of the costliest components of a hospital, and a reduction in the procedure time is expected to be followed by a reduction in cost to patients, which results in a higher chance of continuing to enjoy organizational success.

In a more recently published study by Casutt et al. (2021) (33), the researchers evaluated FM implantation procedures using endobronchial insertion under fluoroscopy like ours but without pre-procedure planning. Their results showed that the median time of the procedure was 31.5 min (10-95 min). Our study showed a shorter duration (22.5 vs. 31 min) and a more consistent range of time (15–30 vs. 10–90 min). In addition, the average number of FMs implanted in Casutt’s study was 3.0, while in our study, it was 5.6 with pre-procedure planning. The comparison reinforced that pre-procedure planning can improve FM implantation efficiency.

The main principle of radiation protection is to protect patients from unnecessary radiation and perform medical procedures with as low as reasonably achievable (ALARA) doses. In this study, we use the DAP measured during the examination as a quantitative tool to compare the radiation exposure of patients as well as staff during the FM insertion procedure.

The results of this study show that with pre-procedure planning, the average DAP is reduced from 67.4 to 25.3 dGy*cm² or by 62%. Furthermore, the overall radiation exposure to patient and staff become more consistent and predictable, which is shown by the reduced SD from 56.1 dGy*cm² in Group 1 patients to 13.8 dGy*cm² in Group 2 patients. This is largely due to the reduction in time of fluoroscopy. Because the location of FMs is predefined and the route to the designated location is guided by the VBN system, respiratory specialists can avoid frequent fluoroscopy during the procedure to locate the position of the endoscope and look for ways to the desired location. Furthermore, staff radiation exposure is expected to be reduced too, which also reduces the chance of staff overexposure resulting in manpower shortage.

The lung tumour moves in all directions, and the amplitude can be up to 12 mm in the cranial–caudal direction (16). According to the CK guideline on FM placement, at least 3 FMs are required for 3D tumour tracking, while 4–6 are recommended for more secure and confident tumour motion tracking along with the CK treatment (34). Meanwhile, researchers also recommended implanting 4–6 FMs to improve marker reliability and tracking accuracy (35).

The results of this study show that the average number of FMs inserted only increased slightly from 4.7 to 5.6 with the application of pre-procedure planning. However, with the new workflow, 10 out of 14 patients can be inserted with 6 FMs, and no patient will be inserted with less than 4 FMs. This largely increases the reliability of the FMs in representing the true tumour motion and the confidence of continuous 3D tracking along with the treatment in considering the possibilities of FM migration.

Theoretically, we could expect better trackability based on the concept that with pre-procedure planning, the ideal locations of the FMs were preliminarily identified. The ideal locations could fulfil both the criteria that it should be at the distal end of small bronchi and located at a predefined distance away from the tumour, making a good overall configuration around the tumour, which means a higher chance of trackability. In reviewing this, we defined the trackability of the target or FMs by the number of FMs implanted to the number of FMs tracked when the CK treatment starts and compare the results of the two groups of patients.

Results showed that the trackability was increased from 41% (minimum 0% to maximum 75%) for patients without pre-procedure planning (Group 1) to 50% (minimum 33% to maximum 100%) for patients with pre-procedure planning (Group 2). However, the difference was insignificant, p = 0.12 (1-tailed). This could be explained by the patient-dependent nature of the probability of FM migration. For patients suffering from frequent coughing due to lung disease, migration or even loss of FMs is common.

FM tracking is the only choice for CK of non-small cell lung cancer (NSCLC) patients with small tumours that are invisible under the planar X-ray view. Real-time fiducial tracking together with Synchrony respiratory motion management results in excellent motion synchronized treatment with sub-millimetre margins to the targets (17). This study introduced a new set of implantation rules and guidelines that could become a helpful reference for other clinical centres using CK in SBRT. We expect that if the suggested pre-planned FM positions method is proved to be more effective, FM implantation procedures will be well-organized, and the results will be more predictable.

There are some assumptions for this study. First, it is not possible to move all FMs in the same vectors and not cause differences in the FM centroid and displacement. Second, tumour shrinkage during the period of CK treatment is negligible. Although one previous study showed that the tumour will shrink during SBRT, the shrinkage rates were not necessarily uniform (18). In the present study, it is assumed that the size of the tumour remains unchanged. Therefore, all the FM displacements that will be measured after the CT images are fused are due to FM displacement.

This study also has some limitations. First, although an ideal FM configuration can be determined, it is nearly impossible to achieve. If the lung tumour is located near the lung circumference of the diaphragm, there will be a no-FM zone such as the chest wall and the diaphragm, which limits the all-around distribution of the FMs, resulting in an unevenly distributed FM arrangement and increasing the distance between the tumour and the FM centroid ( Figure 10 ). Second, it is not well-known how close the VBN combined with ultrathin endoscopy-guided FM implantation can be to the pre-planned FM positions. It is assumed that the discrepancies between the two can be within 1 cm. Third, only small bronchi near the lung circumference are highly FM fixed. Therefore, only about one-third of the lung volumes are possible for FM implantation. Moreover, to some extent, the fixation of FMs was patient dependent. For example, coughing is one cause of FM migration. Patients with a coughing problem during the period of CK treatment may have a higher chance of FM migration and displacement. Finally, some patients have more than one tumour located close to each other. A set of FMs can be used to treat a combination of one or more tumours at the same time, but this will complicate the definition of the tumour centroid and the distance of the FM to the tumour.

The CK FM tracking system is complex. To achieve 6DOF FM tracking, criteria such as minimum distance between FM angles between FM configuration should all be fulfilled. However, there were no technical guidelines on how to implant FMs in fulfilling the CK tracking criteria. Furthermore, FMs cannot easily be fixed in the small bronchi after implantation. Future studies should focus on increasing FM stability and fixation by using different kinds of FMs and how to implant FMs into a designated position accurately. Moreover, future similar studies should consider increasing their sample size.